RELATIONAL AUTONOMY

This page intentionally left blank

RELATIONAL AUTONOMY

Feminist Perspectives on Autonomy,

Agency, and the Social Self

Edited by
Catriona Mackenzie and Natalie Stoljar

New York Oxford

Oxford University Press
2000
 Oxford University Press
Oxford New York
Athens Auckland Bangkok Bogota Buenos Aires C a l c t t a
Cape Town Chennai Dar cs Salaam Delhi Florence Hong Kong Istanbul
Karachi Kuala Lumpur Madrid Melbourne Mexico City Mumbai
Nairobi Paris Sao Paulo Singapore 'Taipci Tokyo Toronro Warsaw
and associated companies in
Berlin Ibadan

Copyright & 2000 by Oxford University Press

Published by Oxford Univcrsity Press

198 Madison Avenue, New York, New York 10016
Oxford is a registered trademark of Oxford University Press.
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording or otherwise,
without the prior permission of Oxford University Press.

Library of Congress Cataloging-in-Puhlication Data

Relational autonomy : feminist perspective on autonomy, agency, and the social self / edited by
Catriona Mackenzie and Natalie Stoljar.
p. cm.
ISBN 0-19-512333-6; ISBN 0-19-512:334-4 (pbk.)
1. Feminist theory 2. Autonomy. 3. Agent (Philosophy) 4. Self (Philosophy)
I. Mackenzie-, Catriona. II. Stoljar, Natalie.
HQ1190.R45 1999
305.42—DC21 99-12792

1357 9 8 6 42
Printed in the United States of America
on acid - free paper
 ACKNOWLEDGMENTS

We wish to express our gratitude to the contributors to this volume for their en-
thusiasm for the project and the generosity with which they responded to edito-
rial comments and suggestions. Diana Meyers also provided encouragement and
invaluable advice at various stages of the project.
Some of the articles in the volume were originally presented at the Conference
on Feminist Perspectives on Autonomy and Agency, held at the Australian National
University in June 1996. We are grateful to Fiona Webster for conference organiza-
tion and to all the participants at the conference, especially to Susan Brison, Lor-
raine Code, Marilyn Friedman, and Diana Meyers, who traveled from North Amer-
ica to present their papers.
For financial assistance in hosting the conference, we are indebted to the Philos-
ophy Program, Research School of Social Sciences, ANU; the Social and Political
Theory Group, Research School of Social Sciences, ANU; the Department of Phi-
losophy, Faculty of Arts, ANU; and the Australia Foundation for Culture and the
Humanities. Financial support for the project as a whole was provided by an Aus-
tralian Research Council Small Grant and a Macquarie University Visiting Scholar
Research Grant.
vi Acknowledgments

We received invaluable research assistance from Fiona Webster in the intial stages
of the project and from Sarah Bachclard in the final stages. Special thanks are due
to Ian Gold, Peter Menzies, and our families.

Sydney, Australia C. M.
May 1999
Canberra, Australia N. S.
May 1999
 CONTENTS

Contributors ix
Introduction: Autonomy Refigured 3
Catriona Mackenzie and Natalie Stoljar

I A U T O N O M Y AND THE S O C I A L

1 Autonomy, Social Disruption, and Women 35

Marilyn Friedman
2 Autonomy and the Social Self 52
Linda Barclay
3 Feeling Crazy: Self-Worth and the Social Character of Responsibility 72
Paul Benson
4 Autonomy and the Feminist Intuition 94
Natalie Stoljar
5 Individuals, Responsibility, and the Philosophical Imagination 112
Genevieve Lloyd
6 Imagining Oneself Otherwise 124
Catriona Mackenzie
viii Contents

7 Intcrscctional Identity and the Authentic Self?:

Opposites Attract! 151
Diana Tietjens Meyers
8 The Perversion of Autonomy and the Subjection of Women:
Discourses of Social Advocacy at Century's End 181
Lorraine Code

II RELATIONAL AUTONOMY IN CONTEXT

9 Choice and Control in Feminist Bioethics 213

Susan Dodds
10 Autonomy and Interdependence: Quandaries in
Genetic Decision Making 236
Anne Donchin
\ 1 Relational Autonomy, Self-Trust, and Health Care for Patients
Who Arc Oppressed 259
Carolyn McLeod and Susan Sherwin
12 Relational Autonomy and Freedom of Expression 280
Susan]. Brison

Index 301
 CONTRIBUTORS

LINDA BARCLAY is a lecturer in ethics at La Trobe University. Her main research is in

applied ethics, including professional ethics and political philosophy.

PAUL BENSON is associate professor of philosophy at the University of Dayton. He

works primarily in the areas of action theory, moral psychology, and social philoso-
phy and is currently writing a book about the social and normative dimensions of
free agency.

SUSAN J. BRISON is associate professor of philosophy at Dartmouth College, where

she also teaches in the women's studies program. She has published articles on fem-
inist ethics, philosophy of law, and social and political philosophy in anthologies
and in journals including Ethics, Legal Theory, and Nomos. Coeditor of Contempo-
rary Perspectives on Constitutional Interpretation (Westview, 1993), she is author of
Speech, Harm and Conflicts of Rights (Princeton University Press, forthcoming
2000) and is currently writing a book on trauma and memory.

LORRAINE CODE is distinguished research professor of philosophy at York Univer-

sity in Toronto. She is the author of Epistemic Responsibility (University Press of
New England, 1987), What Can She Know? Feminist Theory and the Construction
of Knowledge (Cornell, 1991), and Rhetorical Spaces: Essays on (Gendered) Locations

ix
x Contributors

(Routledge, 1995). She is currently developing an ecological model for naturalistic

epistemology and editing a one-volume Encyclopedia of Feminist Theories (Rout-
ledge, 1999).

SUSAN DODDS is a senior lecturer in philosophy and the deputy director of the In-
stitute of Social Change and Critical Inquiry at the University of Wollongong in
Australia. She has published work in political philosophy, bioethics, philosophy of
mind, and philosophy of feminism. With Rebecca Albury, she is currently working
on a critical examination of the debate about policymaking in reproductive tech-
nology, analyzing the demand for regulation in terms of women's status as persons
and as citizens.

ANNE DONCHIN is professor of philosophy at Indiana University, Indianapolis, and

former director of their women's studies program. She teaches and writes princi-
pally in the areas of bioethics and feminist philosophy and has published numer-
ous articles in leading bioethical and feminist journals. She is coedhor (with Laura
Purdy) of Embodying Bioethics: Recent Feminist Advances ( R o w a n & Littlefield,
1999) and is currently completing Procreation, Power, and Personal Autonomy: Fem-
inist Reflections (Temple University Press, forthcoming). She is also a founder and
coordinator of the International Network on Feminist Approaches to Bioethics.

MARILYN FRIEDMAN is professor of philosophy at Washington University in St.

Louis. Her special interests are in ethics, social philosophy, and feminist theory.
She is the author of What Are Friends For? Feminists Perspectives on Personal Rela-
tionships and Moral Theory (Cornell, 1993), the coauthor of Political Correctness:
For and Against (Rowman & Littlefield, 1995), and the coeditor of Feminism and
Community (Temple, 1995) and Mind and Morals: Essays on Ethics and Cognitive
Science (Bradford/MIT, 1996).

GENEVIEVR LLOYD is professor of philosophy at the University of New South Wales

in Sydney. She is the author of The Man of Reason: "Male" and "Female" in Western
Philosophy^ 2nd ed. (Routledge, 1993), Being in Time: Selves and Narrators in Phi-
losophy and Literature (Routledge, 1993), Part of Nature: Self-Knowledge in Spin-
oza's Ethics (Cornell, 1994), and Spinoza and the Ethics (Routledge, 1996).

CATRIONA MACKENZIE is a senior lecturer in philosophy at Macquarie University in

Sydney. She works primarily in ethics, feminist theory, and moral psychology and
has published articles in a variety of journals and edited collections. Her current
research includes a book on the interconnections among emotion, imagination,
embodiment, and autonomy.

CAROLYN MCLEOD is a PhD candidate in the philosophy department at Dalhousie

University in Halifax, Canada. She has been awarded postdoctoral fellowships by
the Center for Human Bioethics, University of Minnesota and the Social Sciences
and Humanities Research Council of Canada. Her dissertation is on the nature of
self-trust and its role in autonomous decision making.
 Contributors xi

DIANA TIETJENS MEYERS is professor of philosophy at the University of Connecti-

cut, Storrs. She is the author of Self, Society, and Personal Choice (Columbia, 1989)
and Subjection and Subjectivity (Routledge, 1994), and she has edited many collec-
tions, most recently Feminists Rethink the S^(Westview, 1997) and Feminist So-
cial Thought: A Reader (Routledge, 1997). She is currently working on two books,
one extending her procedural account of autonomy and the other examining how
cultural figurations of gender confound agency.

SUSAN SHERWIN is professor of philosophy and women's studies at Dalhousie Uni-

versity in Halifax, Canada. She is the author of No Longer Patient: Feminist Ethics
and Health Care (Temple, 1992) and is coordinator of the Feminist Health Care
Ethics Research Network, which wrote The Politics of Women's Health: Exploring
Agency and Autonomy (Temple, 1998). She is a founding member of Feminist Ap-
proaches to the Bioethics Network and is a member of the board of the Interna-
tional Association of Bioethics.

NATALIE STOLJAR received her PhD in philosophy from Princeton University in

1994. She is currently a senior lecturer in philosophy at Monash University, Mel-
bourne. Her main research areas are moral and legal philosophy and feminist
philosophy. She is currently working on a book on legal interpretation.
This page intentionally left blank
RELATIONAL AUTONOMY
This page intentionally left blank
 INTRODUCTION
Autonomy Refigured

Catriona Mackenzie
Natalie Stoljar

Introduction
In the current climate of feminist theory, the notion of individual autonomy may
seem an unlikely topic for a collection of feminist essays. Although the ideal of au-
tonomy once seemed to hold out much promise, in providing both a liberatory goal
and a moral standpoint from which to criticize sex-based oppression, autonomy
is now generally regarded by feminist theorists with suspicion. Crudely stated, the
charge is that the concept of autonomy is inherently masculinist, that it is inextri-
cably bound up with masculine character ideals, with assumptions about selfhood
and agency that are metaphysically, epistemologically, and ethically problematic
from a feminist perspective, and with political traditions that historically have been
hostile to women's interests and freedom. What lies at the heart of these charges is
the conviction that the notion of individual autonomy is fundamentally individu-
alistic and rationalistic.
The aim of this collection is to challenge this conviction. While it is true that
feminist critiques of autonomy have identified serious theoretical and political prob-
lems with some historical and contemporary conceptions of autonomy, the notion
of autonomy is vital to feminist attempts to understand oppression, subjection, and
agency. Moreover, none of the major feminist critiques justifies repudiating the con-
cept altogether. The challenge facing feminist theorists is rather to draw on aspects

3
 4 Introduction

of the feminist critique of autonomy to reconceptualize and refigure the concept of

individual autonomy from a feminist perspective. It is this refigured concept that we
are calling "relational autonomy."1
The term "relational autonomy," as we understand it, does not refer to a single
unified conception of autonomy but is rather an umbrella term, designating a range
of related perspectives. These perspectives are premised on a shared conviction, the
conviction that persons are socially embedded and that agents' identities are formed
within the context of social relationships and shaped by a complex of intersecting
social determinants, such as race, class, gender, and ethnicity. Thus the focus of re-
lational approaches is to analyze the implications of the intersubjective and social
dimensions of selfhood and identity for conceptions of individual autonomy and
moral and political agency.
The purpose of this introduction is to situate relational approaches to auton-
omy in the context of both feminist critiques of autonomy and contemporary
philosophical accounts of autonomy. We outline and assess five major feminist cri-
tiques of autonomy and then map out the conceptual terrain of contemporary philo-
sophical accounts of autonomy, focusing in particular on the vexed relationship
between autonomy and socialization. Finally, we draw on the conclusions of the pre-
vious two sections to provide a more detailed characterization of relational models
of autonomy. Before proceeding to these more detailed discussions, however, it is
worth providing a general overview of the notion of autonomy within contempo-
rary moral and political philosophy.
The concept of individual autonomy has come to occupy a central, if con-
tested, place in moral and political philosophy. For example, questions concerning
the nature and value of individual autonomy and its compatibility with a recogni-
tion of the social embeddedncss of persons are central to current debates in politi-
cal philosophy among liberals, communitarians, and feminists. 2 Whereas the value
of autonomy is contested in political philosophy, this is less the case in applied ethics
and legal philosophy. In these areas, disputants on either side of a range of different
debates often agree on the value, even the primacy, of autonomy but disagree about
how it is best promoted. In bioethics, for example, appeals to patient autonomy fig-
ure prominently in debates about voluntary active euthanasia, in analyses of the
physician-patient relationship, and in debates about the ethics of reproductive tech-
nologies and surrogate motherhood, with the different disputants in these debates
each claiming to have autonomy on their side.3 In legal philosophy, the appeal to in-
dividual autonomy has been used to defend an unrestricted right to freedom of
speech in the context of debates concerning legal restrictions on hate speech and
pornography. It has also been used to contest such a right.4 Finally, recent debates
about the nature of autonomy in moral philosophy and moral psychology intersect
with, and have implications for, an increasingly broad range of philosophical de-
bates. These include the free will determinism debate; theories of personal identity;
theories of practical reasoning and deliberation; conceptions of agency and self-
hood; accounts of character and self-respect; and theories of oppression, embodi-
ment, and subjectivity.
However, despite the importance of the notion of individual autonomy within
contemporary moral and political theory, there is no consensus about what the
 Autonomy Refigured 5

concept means or when it can be legitimately employed. In bioethics, autonomy is

often equated with informed consent. In rational choice theory, autonomy is equated
with voluntary, rational choice. In other contexts, for example, within liberal po-
litical theory, autonomy is considered to be an individual right. For liberals of a lib-
ertarian persuasion, the right to autonomy is construed as negative liberty, a right
of the individual to freedom from undue interference in the exercise of choice
(moral, political, personal, and religious) and in the satisfaction of individual pref-
erences. For Rawlsian liberals, autonomy is understood in Kantian terms as the ca-
pacity for rational self-legislation and is considered to be the defining feature of
persons.
The conceptual thread that links these different uses of the notion of autonomy
is the idea of self-determination or self-government, which is taken to be the defin-
ing characteristic of free moral agents. Notions of autonomy as individual choice or
as a political right flow from, and are derivative of, this defining characteristic. Be-
cause of the core role of the idea of autonomy as self-determination, section 3 of
this introduction focuses on the notion of autonomy as it is understood in moral
psychology and on the debates within that literature concerning the conditions nec-
essary for its development and exercise.
Despite the differences that we have identified among uses of the concept of in-
dividual autonomy, there are nevertheless important overlaps between the different
senses of autonomy and between the different domains in which the concept is em-
ployed. There are clear historical and conceptual links among conceptions of choice,
conceptions of political rights, and conceptions of individuality, autonomy and
selfhood. These links explain why the concept of autonomy is so contested among
liberals, communitarians, and feminists. And they also explain why debates about
autonomy are often fraught with confusion—because the concept of autonomy is
sometimes conflated with one particular conception of autonomy and its attendant
conceptions of choice and rights. The most obvious example is the caricature of in-
dividual autonomy as exemplified by the self-sufficient, rugged male individualist,
rational maximizing chooser of libertarian theory. It is this caricature that is often
the target of feminist critiques of autonomy.5 Given the widespread cultural associ-
ation of individual autonomy with this caricature, it has been, and continues to be,
important for feminists to contest this particular conception of autonomy. How-
ever, it is also imperative for feminists to reclaim and reconceptualize the concept of
individual autonomy and to articulate conceptions of choice and of political rights
that are more adequate from a feminist perspective. To do so, feminist theorists must
draw on both mainstream philosophical theories of autonomy and on feminist cri-
tiques of culturally dominant conceptions of individuality, selfhood, and moral and
political agency.

Five Feminist Critiques of Autonomy

This section briefly outlines five major feminist critiques of the notion of auton-
omy: symbolic, metaphysical, care, postmodernist, and diversity. In what follows, we
outline the major components of each critique, arguing that none of them justifies
rejecting the concept of autonomy altogether.6
6 Introduction

Symbolic Critiques
The symbolic critique of autonomy has been articulated most clearly and force-
fully by Lorraine Code.7 Code's critique is not directed toward any particular the-
ory of autonomy but rather toward the abstraction or character ideal of the "au-
tonomous man." This character ideal, she claims, informs mainstream moral theory
and epistemology, to their detriment, and is at the heart of what she regards as the
"autonomy-obsession" of contemporary Western culture. Central to this character
ideal is the notion of self-sufficient independence, which functions both descrip-
tively and prescriptively to promote a particular conception of human nature and a
particular conception of the telos of human life. The descriptive premise on which
the character ideal is based is the notion that human beings are capable of leading
self-sufficient, isolated, independent lives. From this premise is drawn the prescrip-
tive conclusion that the goal of human life is the realization of self-sufficiency and
individuality. Code explains:

Autonomous man is—and should be—self-sufficient, independent, and self-

reliant, a self-realizing individual who directs his efforts towards maximizing his
personal gains. His independence is under constant threat from other (equally
self-serving) individuals: hence he devises rules to protect himself from intrusion.
Talk of rights, rational self-interest, expedience, and efficiency permeates his
moral, social, and political discourse. In short, there has been a gradual alignment
of autonomy with mdividiialism?

Although this character ideal is an abstraction that theorists recognize is un-

likely ever to be attained, nevertheless, the ideal leads in practice to a number of
problematic views. First, it supports valuing substantive independence over all
other values, in particular over those arising from relations of interdependence,
such as trust, loyalty, friendship, caring, and responsibility. Second, it promotes a
very stripped-down conception of agents as atomistic bearers of rights, a conception
in which the diversity and complexity of agents are pared away and agents are re-
duced to an interchangeable sameness. Third, it suggests that values, social practices,
relationships, and communities that are based on cooperation and interdependence
threaten, or at least compromise, autonomy.
Code is right to be critical of the character ideal of the autonomous man, and
she is correct in pointing out its influence in contemporary Western cultures. How-
ever, the point of Code's critique is not to urge feminists to reject the concept of au-
tonomy altogether. She acknowledges that critical reflection and assessment arc of
vital importance to individual women and to feminist political goals. Her point is
rather that an adequate understanding of critical assessment must begin by reject-
ing the conception of atomistic subjectivity characteristic of "hyperbolized auton-
omy" and replacing it with a relational view of subjectivity, centered on the recogni-
tion that persons are "second persons" who only become persons in relations with
others.9 Thus Code's critique is directed toward a cultural character ideal, rather
than toward contemporary philosophical accounts of autonomy, and it does not
provide reasons for rejecting the attempt to rehabilitate autonomy. The question
that is raised, but not answered, by her critique is how to develop models of critical
 Autonomy Refigured 1

reflection that are consistent with a relational conception of subjectivity. Code's con-
tribution to this collection takes up this question by examining the role that advo-
cacy and testimony can play in furthering the autonomy of oppressed individuals
and groups.

Metaphysical Critiques
The metaphysical critiques of the notion of autonomy are some of the most en-
trenched in the feminist literature. They claim that attributing autonomy to agents
is tantamount to supposing that agents are atomistic, or separate, or radically indi-
vidualistic. Since, as feminists and others have pointed out, agents are socially em-
bedded and seem to be at least partially constituted by the social relations in which
they stand, if attributing autonomy to agents is indeed to presuppose individualism
or atomism, then it seems that the attempt to articulate autonomy rests on a mis-
take.10 There are four possible positions within the metaphysical critiques that we
wish to distinguish, which correspond to four different understandings of the term
"individualism." Individualism can be understood as any of the following claims:
first, that agents are causally isolated from other agents; second, that agents' sense of
themselves is independent of the family and community relationships in which they
participate; third, that agents' essential properties (that is, their natures, or meta-
physical identities) are all intrinsic and are not comprised, even in part, by the social
relations in which they stand; and fourth, that agents are metaphysically separate in-
dividuals. We consider each version of individualism in turn.
Annette Baier argues against individualism in the first sense by proposing that
persons are what she calls "second-persons." This means that the development of per-
sons requires relations of dependency on other persons: "Persons are essentially suc-
cessors, heirs to other persons who formed and cared for them, and their personality
is revealed both in their relations to others and in their response to their own recog-
nized genesis."11 Baier's claim is a causal one about the development of persons, their
personalities and capacities. If she is right, individualism in the first sense is false. Per-
sons are not causally isolated from other persons; indeed, the development of persons
requires relations of dependency with others. We agree with Baier's version of anti-
individualism. However, a commitment to this version of anti-individualism does
not justify rejecting the concept of autonomy. Most theories of autonomy—espe-
cially those we discuss in the next section—are compatible with the notion that per-
sons are relational in Baier's sense.12 Baier's observations recommend that any recon-
ceptualized notion should attend to the status of persons as second persons, not that
autonomy should be ruled out altogether.
The second and third senses of individualism are often run together in a com-
mon feminist critique of autonomy. This critique alleges that theories of autonomy
presuppose "abstract individualism," namely, the thesis that "logically if not empir-
ically, human beings could exist outside a social context."13 However, since persons,
and hence their characteristics and capacities, are constituted, and not simply caused,
by the relations to others in which they stand, theories of autonomy presuppose a
flawed conception of selfhood and should be jettisoned. The claim that agents are
constituted by the social context in which they participate is ambiguous between a
8 Introduction

psychological reading and a metaphysical one. Either it means that an agent's sense of
herself is constituted by her social context, which denies individualism in the second
sense, or it means that social relations are essential components of an agent's iden-
tity, which denies individualism in the third sense.
It is obviously correct to deny individualism in the second sense, yet to do so
does not undermine the notion of autonomy. That is, we can accept that social
relations influence and perhaps constitute agents' senses of themselves and their
capacities, without concluding that capacities such as autonomy are nonexistent.
As relational approaches to autonomy emphasize, reconceptualizations of auton-
omy must acknowledge the effect of social conditions on agents' senses of them-
selves, such as their senses of self-esteem and self-trust. Furthermore, denying in-
dividualism in the third sense, while more controversial,14 does not undermine
the notion of autonomy. Anti-individualism in the third sense entails that social
relations arc essential properties of persons. Yet, as far as we know, there are no
theories of autonomy—with the possible exception of the hyperbolized notion de-
scribed by Code — that take a stand on the question of the kinds of properties that
are metaphysically essential to the existence and persistence of the autonomous
agent. After all, the metaphysical question of the essential nature of persons is sep-
arate from and perhaps prior to the question of the nature of a person's characteris-
tics and capacities, including her autonomy. Thus, even if both the second and third
senses of anti-individualism are true, they provide no reason to reject autonomy.
Let us consider finally the fourth conception of individualism, which claims
that agents are metaphysically separate individuals or entities. As Louise Antony
points out, this sense of individualism is true even if we assume that social relations
are essential properties of these individuals.15 Moreover, the claim that the concept
of individual autonomy presupposes individualism in the fourth sense is trivially
true because the phrase "individual autonomy" implies that agents arc separate en-
tities with a capacity for autonomy. Thus, no theory of individual autonomy could
presuppose anti-individualism in the fourth sense.
It follows that metaphysical critiques of autonomy are mistaken. If we under-
stand "individualism" in any of the first three senses, most theories of autonomy do
not presuppose individualism. If we understand "individualism" in the fourth sense,
it is trivially true that theories of individual autonomy presuppose individualism.
The moral of the discussion, therefore, is that the concept of individual autonomy
should be distinguished from individualistic conceptions of individual autonomy.
The task of this collection, namely, to rehabilitate individual autonomy, does not
entail rehabilitating individualistic autonomy in any of the first three senses. Indeed
taking the feminist critiques seriously is precisely to explore the possibilities for anti-
individualistic conceptions of autonomy.

Care Critiques
According to care critiques, traditional ideals of autonomy give normative primacy
to independence, self-sufficiency, and separation from others, at the expense of a
recognition of the value of relations of dependency and interconnection. Since such
relations have historically been central to women's lives and symbolically associated
 Autonomy Refigured 9

with femininity, it is argued that traditional conceptions of autonomy not only de-
value women's experience and those values arising from it, such as love, loyalty, friend-
ship, and care, but also are defined in opposition to femininity. Traditional concep-
tions are thus masculinist conceptions.
Care critiques overlap with aspects of both symbolic critiques and metaphysi-
cal critiques. They are to be distinguished from symbolic critiques because they do
not characterize the concept of autonomy as inherently tainted by the connotations
of masculinist conceptions of autonomy.16 Rather the latter conceptions are criti-
cized as normatively flawed. Similarly, care critiques differ from metaphysical cri-
tiques because of their emphasis on the value of relations of nurturance and de-
pendency for agents, rather than on the implications of these relations for the
metaphysical nature of the self.
Care critiques appeal to Nancy Chodorow's psychoanalytic account of mascu-
line and feminine psychic development.17 On that account, in gender unequal soci-
eties in which women are the primary caregivers, masculine psychic individuation
and separation are conflated with separation from the mother. Masculine selfhood
is thus defined as other than the feminine mother and psychic autonomy, or the de-
velopment of a strong sense of self, is equated with self-sufficient independence.
Feminine psychic development, by contrast, involves identification with the mother
and so promotes capacities for connection and interdependence, often at the ex-
pense of the development of a strong sense of self.
Virginia Held's account of the notion of the self that is at the heart of the care
critiques, echoes this story:
The self . . . is seen as having both a need for recognition and a need to under-
stand the other, and these needs are seen as compatible. They are created in the
context of mother-child interaction and are satisfied in a mutually empathetic
relationship. . . . Both give and take in a way that not only contributes to the sat-
isfaction of their needs as individuals but also affirms the 'larger relational unit'
they compose. Maintaining this larger relational unit then becomes a goal, and
maturity is seen not in terms of individual autonomy but in terms of competence
in creating and sustaining relations of empathy and mutual intersubjeciivity.18

The target of care critiques, and of Chodorow's critique of masculine selfhood,

is autonomy understood as substantive independence. Although some care critiques
unfortunately conflate autonomy with substantive independence, and therefore re-
ject autonomy on the basis of this critique,19 many are careful to distinguish them,
as does Chodorow herself. Thus other theorists influenced by the care perspective,
such as Jennifer Nedelsky, argue for a reconceptualization of autonomy modeled on
the mother-child relationship.20 Such reconceptualizations give normative primacy
to the relations of care and connection identified by Chodorow and Gilligan and ar-
ticulate autonomy within the context of these relations.
Similarly, Evelyn Fox Keller has developed a notion of dynamic autonomy,
which she contrasts with static autonomy.21 Keller sees autonomy as a kind of com-
petence. The distinction between dynamic and static autonomy is a distinction be-
tween, on the one hand, the kind of competence that promotes an enhanced sense
of self and, on the other, the kind of competence and mastery that are pursued in
10 Introduction

the interests of domination, denied connectedness, and defensive separateness. Dy-

namic autonomy involves both relatedness to, and differentiation from, others. It
promotes a sense of agency in a world of "interacting and interpersonal agents" and
a sense of others "as subjects with whom one shares enough to allow for a recogni-
tion of their independent interests and feelings—in short for a recognition of them
as other subjects."22 Static autonomy, on the other hand arises from seeing others as
a threat to the self, from insecurity about the self, and from fears of dependency and
loss of self-control, leading to patterns of domination and control over others. In its
extreme forms static autonomy gives rise to sadism.23
Care critiques, of the kind developed by Chodorow, Held, Nedelsky, and
Keller thus do not repudiate the concept of autonomy altogether. Rather they rec-
ommend that autonomy be reconceptualized so that it is not defined in opposition
both to femininity and to relations of dependence and connection. The issues
raised by care critiques are therefore to some extent continuous with the project of
articulating a relational conception of autonomy as we understand it. However, in
focusing primarily on intimate dyadic relations, particularly between mother and
child, care critiques provide a very circumscribed reconceptualization of autonomy.
In particular, they fail to address the complex effects of oppression on agents' ca-
pacities for autonomy; and they provide a somewhat limited reconceptualization of
the social dimensions of agency and selfhood. We discuss these issues further in the
last section.

Postmodernist Critiques
What we are calling, rather loosely, postmodernist critiques of autonomy derive
from a number of distinct theoretical perspectives: psychoanalytic theory, Fou-
cauldian theories of power and agency, and feminist theories of sexual difference
and otherness. As far as we are aware, there is no sustained and detailed critique of
contemporary philosophical accounts of autonomy from any of these perspectives.
Rather, critics draw on the so-called "critique of the subject" that emerges from one
or more of these perspectives to criticize the assumptions that they allege to be im-
plicit in the ideal of autonomy. In particular, it is claimed that the ideal assumes that
agents are self-transparent, psychically unified, and able to achieve self-mastery. Crit-
ics who draw on psychoanalysis charge that since Freud, these assumptions have
been shown to be illusory.24 In contrast to the complete self-transparency, seamless
psychic unity, and self-mastery supposedly required by autonomy, the picture of
the psyche that emerges from psychoanalytic theory depicts agents as conflict-
ridden, often self-deluded, fundamentally opaque to themselves, and driven by ar-
chaic drives and desires of which they may not even be aware, let alone able to
master. Critics who draw on Foucauldian theories of power and agency suggest
that theories of autonomy assume a pure Kantian free will, or a true self. This assump-
tion naively ignores the fact that subjects are constituted within and by regimes,
discourses, and micropractices of power. There is no pure, self-determining free
will that somehow escapes the operations of power, nor is there a true self, there to
be discovered through introspective reflection. Agency must be reconceptualized
not as a matter of individual will but as an effect of the complex and shifting con-
 Autonomy Refigured 11

figurations of power.25 Feminist theories of difference and otherness allege that

the notion of autonomy is a historically, socially, and culturally specific ideal that
parades as a universal norm. Not only does this norm suppress internal differenti-
ation within the subject, but also in masking its specificity behind a veneer of uni-
versality, it functions coercively to suppress different others.26
Although they draw on rather divergent theoretical perspectives and criticize
autonomy for rather divergent reasons, there is a unifying theme underlying the
postmodernist critiques. The theme is that the notion of autonomy is a kind of
conceit or illusion of the Enlightenment conception of the subject. 27 Thus it is
charged that defenders of autonomy still cling to the Cartesian idea that con-
sciousness can be transparently self-aware or to the Kantian view of persons as ra-
tional self-legislators, despite the fact that such views have been so decisively chal-
lenged since Nietzsche, Freud, and their heirs. Moreover, the persistence of such
views is not just a harmless anachronistic hangover of the Enlightenment. It is
complicit with structures of domination and subordination, in particular with the
suppression of others—women, colonial subjects, blacks, minority groups—who
are deemed incapable of achieving rational self-mastery.
We cannot do justice here to these different theoretical perspectives, nor have
we space to discuss the complex, and conflicting, accounts of agency and subjec-
tivity developed by each.28 The question is whether the general critique presented
by these perspectives justifies rejecting the notion of autonomy. Our view is that it
does not. The general critique conflates the notion of autonomy with certain con-
ceptions of autonomous agents. Conceptions of agents that require complete self-
transparency and psychic unity are certainly vulnerable to the Freudian challenge.
Kantian conceptions that conceive autonomous agents as free from empirical de-
termination are vulnerable to Foucauldian critiques, among others. Likewise, some
conceptions of autonomous agents have functioned historically to enforce exclu-
sionary norms and ideals of the person. But for reasons that we discuss in the next
two sections, the concept of autonomy need not be based on such assumptions
about agents. It need not require agents to be completely self-transparent and psy-
chically unified; assume that only a pure will, free from all empirical determination,
can be self-determining; or enforce a hegemonic identity. Nevertheless, the post-
modernist critiques are salutary, for they alert us to the need to develop notions of
autonomy based on richer, more psychically complex, and more diverse conceptions
of agents.

Diversity Critiques
Diversity critiques parallel postmodernist critiques in challenging the assumption
that agents are cohesive and unified. Such critiques claim that each individual has a
"multiple identity," which reflects the multiple groups to which the individual be-
longs. For example, Kimberle Crenshaw argues that "the experiences of women of
color are frequently the product of racism and sexism, and these experiences tend
not to be represented within the discourses of either feminism or of antiracism."29
Thus, the identities of individual women are "intersectional" in that they combine
the group affiliations unique to that woman. A similar ambivalence inherent in the
12 Introduction

identities of women of marginalized groups has been described by Maria Lugones as

the experience of living within and traveling between different "worlds."30
The idea of intersectionality may seem incompatible with the presupposi-
tions of theories of autonomy. It implies that because different and sometimes
conflicting group identities intersect in the formation of individual identity, many
individuals do not have a unified or integrated sense of self. As with the post-
modernist arguments we have already discussed, if theories of autonomy are thought
to presuppose a transcendental self, the conception of the self offered by diversity
theorists is incompatible with that offered by theories of autonomy. Even if theo-
ries of autonomy do not presuppose a transcendental self, as we show in the next
section, notions of integration are central to many accounts of autonomy. Hence
intersectionality does seem potentially to undermine conceptions of autonomy
that require integration.
However, while intersectionality provides an important cautionary note for the-
ories of autonomy, it need not undermine the claim that integration is a necessary
condition of autonomy. It is plausible that certain radical kinds of disintegration of
the self undermine the integration condition, for example, the disintegration as a re-
sult of trauma that is described by Susan Brison. But is the fragmentation of the self
implied by intersectionality an analogous kind of disintegration? 31 If not, is the am-
bivalence implied by intersectionality, or indeed other kinds or degrees of ambiva-
lence, perhaps compatible with the integration required for autonomy? Diana Mey-
ers takes up these questions in her contribution to this volume.
In focusing on the particular features of individuals' identities, diversity theo-
rists attempt to explain how and why individuals should not be absorbed into some
group or other. Crenshaw claims that "in the area of rape, intersectionality provides
a way of explaining why women of color must abandon the general argument that
the interests of the community require the suppression of any confrontation around
interracial rape."32 The resistance to seeing individuals as replicating the interests
and identities of the groups to which they belong is congenial to an emphasis on in-
dividual autonomy. Thus, despite the problems posed for certain theories of auton-
omy by the notion of intersectionality, we would argue that the insights of diversity
critiques, especially that agents have particular identities and not group identities
shared with other members of the group, enhance rather than reduce the need for
feminist recharacterizations of individual autonomy.

Conceptions of Autonomy in Moral Psychology

We argued in the previous section that none of the five major feminist critiques jus-
tifies repudiating the notion of autonomy. The critiques, however, do require that
attempts to refigure autonomy pay attention to the complex nature of the au-
tonomous agent and to the differentiated social and historical contexts in which
agents are embedded. Feminists have been particularly concerned, in this regard,
with attending to and analyzing oppressive social contexts and their effects on
agents. It is here that the concerns of feminist critiques of autonomy intersect with
the concerns of mainstream theories of autonomy, to which we turn in this section.
 Autonomy Refigured 13

From both a feminist perspective and that of many mainstream theories, oppressive
socialization often seems inimical to agents' autonomy.
Contemporary accounts of individual autonomy hold that autonomy, or self-
determination, involves, at the very least, the capacity for reflection on one's moti-
vational structure and the capacity to change it in response to reflection. This view is
underpinned by the intuition that there is an important difference between those as-
pects of an agent's motivational structure that she unreflectively finds herself with
and those aspects that, as a result of autonomous reflection, she regards as "her
own." Disagreements among different accounts of autonomy arise in explicating
what is involved in the process of reflection, in explaining how reflection secures au-
tonomy, and in making sense of the notion of "one's own."
The debates within the contemporary literature arise in response to two distinct
sets of issues, which we refer to as the metaphysical problem and the socialization
problem. The metaphysical problem focuses on the implications of determinism for
autonomy and on whether a successful compatibilist response to the problem of de-
terminism must rely on the notion of a "true" or "metaphysical" deep self. The so-
cialization problem focuses on the implications of socialization for autonomy. Since
our primary aim is to highlight issues within the mainstream literature on autonomy
that intersect with feminist concerns and to explain the impetus toward the devel-
opment of relational approaches to autonomy, our discussion in this section focuses
mainly on the socialization problem.
We divide theories of autonomy into procedural and substantive. There are
structural, historical, and competency versions of procedural theories and strong
and weak versions of substantive theories. We argue that procedural theories en-
counter a series of difficulties in attempting to reconcile autonomous agency with
socialization, especially oppressive socialization. In particular, since structural pro-
cedural theories analyze autonomy as a feature of an agent's occurrent mental states,
they cannot do justice to the historical processes of socialization leading up to those
states. Although historical theories address this problem, they nevertheless face trou-
ble in explicating the difference between autonomous and nonautonomous pro-
cesses of critical reflection. In addition, historical theories are to a large extent con-
cerned with the negative effects of socialization on autonomy, rather than the global
implications of socialization for autonomy. Meyers's competency account offers a
fine-grained analysis of these global implications but falters when she attempts to
preserve content neutrality for her theory. These problems with procedural accounts
have led some theorists to the conclusion that only a substantive account of auton-
omy can explain how oppressive socialization impairs autonomy.

Procedural Theories of Autonomy

The vast majority of recent theories are variants of a basic procedural account. On
procedural, or content-neutral, accounts, the content of a person's desires, values, be-
liefs, and emotional attitudes is irrelevant to the issue of whether the person is au-
tonomous with respect to those aspects of her motivational structure and the actions
that flow from them.33 What matters for autonomy is whether the agent has sub-
14 Introduction

jected her motivations and actions to the appropriate kind of critical reflection.
Where structural and historical versions of procedural theories differ from each other
is in their different accounts of the processes of critical reflection necessary and suf-
ficient to secure autonomy.34

Structural Theories. Structural models of autonomy focus on the agent's occurrent

motivational structure and on whether the present desires, beliefs, and values on
which the agent acts have been subject to the appropriate kinds of critical reflection.
Most structural theories arc hierarchical. Hierarchical theories distinguish different,
hierarchically ordered elements of the self and characterize autonomy as requiring
a certain kind of structural organization of these different elements. Of the hierar-
chical accounts, the most influential is that of Harry Frankfurt, which characterizes
autonomy as an accord between an agent's first-order desires and her second-order
volitions.35 Frankfurt identifies an agent's will with an effective first-order desire and
claims that autonomy, or freedom of the will, requires both that the agent exercise
control over her will and that she identify, at the level of her second-order volitions,
with her will. Identification is the outcome of a process of reflection in which the
agent distinguishes those desires that she endorses or regards as "her own" from
those desires that she merely finds herself with and is either indifferent to or regards
as external to herself.
Within the literature, a number of important objections have been raised
against Frankfurt's hierarchical account. We discuss several objections that point to
more general problems facing hierarchical theories. The first objection, articulated
most forcefully by Gary Watson, is the regress objection. In a nutshell, Watson ar-
gues that Frankfurt's distinction between first-order desires and second-order vo-
litions, and the notion of identification, do not provide sufficient conditions for
autonomy. For what ensures that our higher-order identifications are themselves au-
tonomous? Watson claims that since Frankfurt's account makes no qualitative dis-
tinctions among desires, the only resource available within his account to respond to
this question is to introduce yet higher orders of desire. But then the same question
can be asked about our identifications at each of these higher levels.36
The second objection, first articulated by Irving Thalberg, is that hierarchical
theories give ontological priority to certain allegedly "higher" aspects of the self,
which are regarded as the person's "true" or "real" self. But why should we equate a
person's true self with her higher-order desires or her valuational system? Doesn't this
view simply beg the question against, for example, a Freudian account of human
psychology, which depicts us as "conflict-prone systems of libidinal, destructive,
morbid, self-preserving, sociable, conscientious, guilt-ridden, and other 'forces,'
'principles' or 'mini-agencies' "?37 Even if we might prize the rational aspects of our-
selves or our capacities for judgment most highly, this does not provide a justifica-
tion for giving those aspects of ourselves ontological priority.
Marilyn Friedman develops two further sets of objections to hierarchical theo-
ries. First, she claims that the distinction within hierarchical theories between higher-
arid lower-order aspects of the self entails a problematic ontological commitment,
namely, to a "true self," which discloses the agent's real metaphysical identity. If au-
tonomy involves discovering and acting in accordance with this true sell, questions
 Autonomy Refipured 15

arise as to how the true self is formed and whether free agency requires the self to be
undetermined. In other words, to be free, do we have to be metaphysically respon-
sible for ourselves, or self-caused? If so, autonomy would be an impossibility.38
Friedman's second set of objections pinpoints the inadequacy of most hierar-
chical theories in dealing with, or even recognizing, the socialization problem. As
Friedman points out, hierarchical theories either ignore the issue of socialization al-
together or tend to assume that the higher-level self somehow transcends the influ-
ences of socialization to which the rest of the self is subject. But this assumption
reintroduces a socialization version of the regress objection because the critically re-
flecting or higher-order self is just as subject to socialization as what we might call
the "first-order" self. As a result, hierarchical theories cannot explain how agents can
be autonomous with respect to their higher-level motivations, for the question can
always be asked whether the higher-order aspects of the self have themselves been
subject to critical reflection in the right ways. Furthermore, Friedman asks, why as-
sume that a person's higher-level principles and values are indicative of what she re-
ally wants or values? In cases of oppressive socialization, an agent may reflectively
endorse her thoroughly socialized higher-level principles and values, but it may be
her apparently wayward first-order desires that are more indicative of what she really
wants and values.
In response to these objections to hierarchical versions of the structural ap-
proach, Friedman proposes a nonhierarchical variant, which she calls an "integra-
tion" model. Instead of taking a "split-level" or "top-down" approach that favors
higher-level assessment, in the integration model critical reflection is rather under-
stood as a "two-way process of integration within a person's hierarchy of motiva-
tions, intermediate standards and values, and highest principles. Only if a person's
highest principles have been subjected to assessment in accord with her intermediate
standards and her motivations, would it be appropriate to consider them her 'own'
principles."39 Thus lower-level motivations provide the standpoint from which an
agent can judge and critically evaluate her higher-order values and principles. Ac-
cording to Friedman, the autonomy-conferring status of critical reflection is secured
because higher-level principles and values are assessed by reference to lower-level
motivations and vice versa. Autonomy is achieved when the different levels of the
self are integrated.
Friedman's integration model makes an important contribution toward under-
standing the effects of oppressive socialization on autonomy. In particular, it pro-
vides an explanation of the kind of self-alienation that characterizes failures of
autonomy in oppressive social contexts. However, integration cannot provide a suf-
ficient condition for autonomy. For, as Paul Benson points out, "An integration
view detects threats to autonomy only when the total internalization of autonomy-
inhibiting socialization fails to take hold or begins to break down."40 Furthermore,
Friedman's integration account, like other structural theories, takes a time-slice ap-
proach to autonomy and hence overlooks the historical processes of formation of
the agent's beliefs, desires, values, and emotional attitudes.

Historical Models. In response to the problems arising from socialization that

plague structural models, several philosophers, especially Gerald Dworkin and John
16 Introduction

Christman, have introduced a historical dimension to analyses of autonomy. Dworkin

argues that for critical reflection to be autonomy conferring, the processes of criti-
cal reflection must themselves be "procedurally independent."41 Critical reflection
must be understood as a capacity rather than as an occurrent state, a capacity that
enables the agent to reflect on and critically assess the various processes (socializa-
tion, parental or peer influence, etc.) by means of which she came to acquire her de-
sires, beliefs, values, and emotional attitudes. The agent is autonomous with respect
to these processes if she endorses them.
John Christman argues that critical reflection yields autonomy when the pro-
cesses of reflection have not been influenced by "illegitimate external influences."42
Illegitimate external influences are those influences (for example, indoctrination,
manipulation, brainwashing, and oppressive socialization) that interfere with agents'
normal cognitive processes of reflection. Such influences are illegitimate if upon be-
coming aware of them as a result of critical reflection, the agent is led to revise her
identifications. In a later article, Christman proposes a counterfactual analysis of au-
tonomy in which a person is autonomous with respect to her values and higher-
order identifications only if, upon becoming aware of the historical process through
which she acquired them, she does not resist this process, or she would not resist it
were she to become aware of it.43 Both the illegitimate influences and the counter-
factual conditions provide a compatibilist response to the problem of socialization.
As long as our desires and preferences are not the products of illegitimate external
influences, or that we do not or would not resist the way we came by these desires
and preferences, they are autonomous.
However, there are at least two problems with Christman's analysis. First, it
tends to equate autonomy with self-transparency. The more we are aware of the his-
torical processes through which we acquired our desires, values, and so on, the more
autonomous we will be. But if autonomy requires this kind of self-transparency, it is
highly dubious that any of us are ever autonomous, for reasons that Thalberg points
out. Second, although seeming too stringent from one perspective (requiring self-
transparency), it is too weak from another because it allows agents whose processes
of critical reflection are not autonomous to count as autonomous. Paul Benson
points out that a person who has been thoroughly socialized to internalize certain
norms, for example, the norms of feminine appearance, would be unlikely to revise
her higher-order identifications as a result of critical reflection if she were to become
aware that these identifications were the result of oppressive socialization.44 It is not
clear that such a person would modify either her identification with the desire to be
attractive to men or her self-conception in which her sense of self-worth is tied to
her physical appearance. Only an agent whose normal processes of reflection are
already autonomous would be led to modify her identifications upon becoming
aware that they are the result of socialization. The thoroughly socialized woman
does not, or would not, resist these processes. Hence she counts as autonomous on
Christman's view.
Whereas historical approaches thus draw attention to the importance of at-
tending to the historical processes of belief-, desire-, and value-formation, they are
nevertheless insufficient to grapple with the autonomy-impairing effects of oppres-
sive socialization. Moreover, both the historical and structural versions of procedural
 Autonomy Refigured 17

theories that are attentive to the implications of socialization for autonomy never-
theless tend to represent socialization in largely negative terms, as an obstacle, or
threat, to autonomy. So whereas it is admitted that our values, beliefs, and desires,
as well as our characters and life plans, are inevitably shaped by socialization, auton-
omy tends to be represented as the quest to shape an identity for oneself in the face
of, or against, this influence. As a result, both structural and historical theorists have
failed to analyze the differences between the kinds of socialization, or aspects of so-
cialization, that promote autonomy and those that impede or undermine it. They
have also neglected the question of what skills and capacities are necessary for au-
tonomy and what kinds of socialization are necessary to promote these skills and ca-
pacities. Competency theories have to a large extent addressed this problem.

Autonomy Competency. Diana Meyers's theory of autonomy competency45 is mo-

tivated by a concern both to explain the autonomy-impairing effects of oppressive
socialization and to develop a theory that is able to explain how agents who are sub-
ject to oppressive social circumstances may nevertheless be partially autonomous, or
autonomous in certain domains of their lives but not in others. She argues that au-
tonomy is a competency comprising a cluster of different skills and capacities, in
particular skills of self-discovery, self-direction, and self-definition, all of which in-
volve reflection. Autonomy involves the capacity to exercise these skills to achieve an
integrated but dynamic self. Thus the notion of integration plays an important role
in Meyers's theory, as it does in Friedman's structural account.46
Meyers's theory might best be classified as a self-realization approach to auton-
omy.47 Unlike Aristotelian versions of self-realization, however, which are premised
on a normative notion of flourishing, Meyers's approach is procedural. Since indi-
viduals differ so significantly in their talents, capacities, character traits, values, de-
sires, beliefs, and emotional attitudes, she argues, there can be no blueprint for what
constitutes an autonomous life. Rather autonomy can be secured only through the
exercise of autonomy competency, or a coordinated repertoire of skills and capac-
ities that enable each individual to fully realize himself or herself, whatever self-
realization amounts to for each particular individual.
Meyers's account is explicitly relational in that she argues that autonomy compe-
tency can be developed only in the context of social relationships, practices, and in-
stitutions. The social context is important to the development of autonomy compe-
tency for at least three reasons. First, self-realization does not require an agent to be
able to develop or realize all of her potentialities; rather it involves the capacity to de-
velop those potentialities that are central to the agent's authentic self-conception, in
the context of the agent's life plan. Since different social environments encourage or
foster the development of different potentialities in any individual, the agent's so-
cial environment is crucial to the agent's ability to recognize and develop her impor-
tant potentialities. Second, the connection between self-realization and the social
environment explains how an agent's attempts to develop an authentic self may be
thwarted or compromised by that environment. Since agents are more likely to de-
velop or emphasize those aspects of themselves (character traits, potentialities, and
talents) that are socially reinforced and to incorporate these aspects into their self-
concepts, the quest for authenticity may be undermined by conventionality. Third,
18 Introduction

the social context may impair agents' capacities to achieve autonomy in a different
way, namely, because certain kinds of socialization encourage the development of
some of: the skills that make up autonomy competency, at the expense of others.
Patterns of gender socialization in contemporary Western cultures, for example,
tend to encourage in women the skills involved in self-discovery because they en-
courage the development in women of emotional receptivity and perceptiveness.
However, women are less likely to be encouraged to develop skills of self-direction
and self-definition. It is precisely these skills that are more likely to be developed in
men, at the expense of skills of self-discovery.
If autonomy competency comprises a range of skills that may be more or less
developed, exercised, and coordinated, it makes sense to think of autonomy as a mat-
ter of degree. Hence Meyers distinguishes episodic autonomy from programmatic
and narrowly programmatic autonomy. She argues that agents subject to oppressive
socialization may exhibit high degrees of episodic autonomy, that is, the capacity to
decide what one wants in weighing up one's desires or how to act in a particular sit-
uation. They may also exhibit narrowly programmatic autonomy, the capacity to
make autonomous decisions in particular aspects of one's life, for example, choice
of partner. Meyers's view is that women subject to traditional gender socialization
are likely to exercise episodic or narrowly programmatic autonomy. However, their
programmatic autonomy, their capacity to critically and reflectively decide major
life issues such as whether or not to be a mother or whether to dedicate oneself to
the pursuit of a career, is likely to be compromised. Like Rawls, Meyers sees self-
realization as crucial to self-respect. If traditional gender socialization compromises
women's capacities to achieve full autonomy and so damages their self-respect, this
kind of socialization is oppressive.
Meyers's account aims to resolve a conundrum confronting feminist theorists,
which she states in the preface to her book: "If women's professed desires are prod-
ucts of their inferior position, should we give credence to those desires? If so, we
seem to be capitulating to institutionalized injustice by gratifying warped desires. If
not, we seem to be perpetuating injustice by showing disrespect for those individu-
als."48 Her answer to the conundrum is that not all desires should be afforded equal
credence or weight. Autonomous desires, namely, those that express our "authentic
selves," as developed through the exercise of skills of self-discovery, self-definition,
and self-direction, are more worthy of satisfaction than desires that merely reflect
uncritical acceptance of social norms or expectations. The suggestion, then, is that it
is not the content of so-called warped desires that renders them less worthy of satis-
faction but the fact that they have not been acquired or endorsed autonomously.
The question we wish to raise concerning Meyers's theory is whether her content-
neutral approach to self-realization really does provide a satisfactory solution, from
a feminist perspective, to this conundrum. In particular, could warped desires, or de-
sires that arise as a result of oppressive socialization, ever be autonomous? Or does
the content of certain desires and choices show that the deliberative processes in-
volved just could not meet the conditions necessary for autonomy?49 In other words,
does the notion of autonomy competency implicitly rely on a more normative and
substantive view of what is required for women to flourish or achieve full auton-
 Autonomy Refigured 19

omy? Furthermore, does a feminist perspective on oppressive feminine socialization

require more stringent normative conditions on autonomy? It is this second concern
that has led some feminists to argue that the problem of oppressive socialization sug-
gests the need for a more substantive approach to autonomy.

Substantive Theories of Autonomy

Procedural accounts implicitly assume that the content-neutral procedural condi-
tions they identify are both necessary and sufficient for autonomy. It is here that
recent work focusing especially on the articulation of the possibility of autonomy
in oppressive circumstances has revealed a flaw in procedural theories. Such work
shows that whereas some set of procedural conditions may indeed be necessary for
autonomy, it is not sufficient. We call those theories that maintain that procedural
accounts must be supplemented by some nonneutral condition substantive theo-
ries.50 There are two basic categories of substantive accounts: strong substantive and
weak substantive. The former reject the content neutrality of procedural theories by
requiring specific contents of the autonomous preferences of agents. The latter re-
ject content neutrality by suggesting further necessary conditions on autonomy that
operate as constraints on the contents of the desires or preferences capable of being
held by autonomous agents. Both kinds of accounts, in different ways, are respond-
ing to objections to procedural accounts that derive from socialization.

Strong Substantive Theories. Susan Wolf and Paul Benson have both defended
strong substantive accounts of autonomy. (We discuss Benson's more recent weak
substantive theory below.) The central idea of these strong substantive accounts is
that of normative competence: to be autonomous, agents must be competent, or
have the capacity, to identify the difference between right and wrong. Since certain
kinds of socialization, including socialization due to oppression, interfere with this
capacity, agents subject to this kind of socialization are not autonomous. Consider
Wolf's account of Jojo, the son of an evil and sadistic tyrant.51 Jojo is raised to be
evil and sadistic like his father; as an adult, he respects his father's values and emu-
lates his desires. His father's evil and sadistic worldview is thoroughly internalized by
Jojo. Wolf argues that since Jojo identifies in the appropriate ways with his first-
order desires, on hierarchical accounts he would count as autonomous. Yet she ar-
gues that he is not autonomous nor morally responsible for the actions performed
on the basis of his evil values, for "it is unclear whether anyone with a childhood
such as his could have developed into anything but the twisted and perverse sort of
person that he has become."52 Wolf argues that to be free and responsible, one has to
be "sane." Although Jojo acts according to his own values and desires in the way re-
quired by hierarchical theories, he is nevertheless "insane"—and hence neither free
nor morally responsible—because his upbringing has blocked his capacity to distin-
guish right from wrong. In her book Freedom within Reason, Wolf articulates the
failure of autonomy here as a failure of a capacity to track an aspect of the world,
namely, the moral or the right. Since for Wolf the demands of morality are equiva-
lent to the demands of "Reason," to be autonomous, agents must be capable of dis-
20 Introduction

cerning the requirements of Reason. Wolf's theory, therefore, is a new version of the
traditional Kantian account of autonomy, in which autonomy is understood as the
capacity for rational self-legislation.53
In several papers, Paul Benson has advocated a similar strong substantive ac-
count of autonomy.54 As we have seen, he points out that inadequate or inappropri-
ate socialization (as in the case of children), as well as oppressive socialization, can
undermine agents' normative competence. In particular, oppressive socialization can
lead to the acceptance of norms, which once internalized, block agents' capacities
for detecting whether the norms are correct. For example, when women internalize
norms purveyed by the fashion industry, such as those that treat personal value as
being dependent on stereotypical feminine appearance, the typical effect is to un-
dermine their capacity to criticize the norms. Since the stereotype suggesting that
personal appearance is a component of self-worth presupposes a false norm, when
women lose the capacity to criticize the norm effectively, and thereby lose the ca-
pacity to detect that it is incorrect, they lose autonomy with respect to the area of
their lives that they take to be governed by the norm. Because both Wolf and Ben-
son are proposing that a necessary condition of autonomy is the capacity to formu-
late desires and preferences and endorse values, with specific contents, their theories
are strongly substantive.

Weak Substantive Theories. Benson has recently proposed a weaker normative com-
petence theory. Consider the following passage from his article "Free Agency and
Self-Worth":
Imagine a feminist remake of Gaslight in which, as in the original, the female
protagonist falls into a state of helplessness and disorientation as a result of a
profound change in her view of herself. . . . [The husband] . . . is a physician . . .
and regards women who are excitable, who have active imaginations and strong
passions, and who are prone to emotional outbursts in public as suffering from a
serious psychological illness. . . . The protagonist has the suspect traits, her hus-
band makes the standard diagnosis, and the 'hysterical wife' ends up isolated, feel-
ing rather crazy. . . . She arrives at her sense of incompetence and estrangement
from her conduct on the basis of reasons that are accepted by a scientific estab-
lishment which is socially validated and which she trusts. 55

Benson argues that neither structural nor historical procedural accounts explain the
agent's apparent lack of autonomy in such a situation, for the agent satisfies the con-
ditions offered by both accounts. What is lacking in such agents is a sense of wor-
thiness to act, which "involves regarding oneself as being competent to answer for
one's conduct in light of normative demands that, from one's point of view, others
might appropriately apply to one's actions."56 Lacking this sense of self-worth is
quite compatible with agents retaining their "power to put [their] will into effect"
and hence is quite compatible with agents retaining the capacities and undergoing
the processes required by procedural theories. We are calling the type of condition
offered by Benson weakly substantive because although it places constraints on the
desires, preferences, and values that count as autonomous, it abandons the content
specificity of strong substantive theories.57 Similarly, Robin Dillon and Trudy Govicr
 Autonomy Refigured 21

have implicitly endorsed weak substantive accounts in suggesting, respectively, that

self-respect and self-trust are necessary conditions of autonomy.58
Both kinds of substantive accounts articulate convincing responses to apparent
counterexamples to procedural accounts. They also provide a good starting point in
explaining why agents who are operating within oppressive institutions and struc-
tures exhibit failures of autonomy. Broadly, two different kinds of explanation are
being offered by the two kinds of substantive accounts. The first is that oppressive
socialization impedes agents' abilities to discern the false norms accepted and per-
petuated by the oppressive context in which they are operating. The objection that
must be answered by this first kind of substantive account is that it is too stringently
normative and conflates autonomy with moral responsibility.59 The explanation of-
fered by the second kind of substantive account sidesteps this objection by linking
oppression to psychological impairment, for instance, impairment of one's sense of
self-worth, self-trust, or self-respect, which impairment itself undermines auton-
omy. The question that must be answered by this kind of account is whether it suf-
ficiently explains feminist intuitions regarding choices in oppressive social contexts.
Natalie Stoljar's chapter in this volume takes up this question.
Both kinds of accounts treat autonomy as intrinsically relational and introduce
necessary conditions of autonomy that derive from the social relations within which
agents are embedded. Hence they implicitly presuppose a richer account of agency
than do procedural accounts.

Relational Autonomy
The difficulties generated by providing an adequate explanation of impairment of
autonomy in contexts of oppressive socialization, together with feminist critiques of
traditional notions of autonomy, have provided the main impetus toward the devel-
opment of a relational approach. Although the feminist critiques discussed earlier do
not directly address the procedural and substantive theories of autonomy discussed in
the previous section, their effect has been to focus attention on the need for a more
fine-grained and richer account of the autonomous agent. The critiques emphasize
that an analysis of the characteristics and capacities of the self cannot be adequately
undertaken without attention to the rich and complex social and historical contexts
in which agents are embedded; they point to the need to think of autonomy as a
characteristic of agents who are emotional, embodied, desiring, creative, and feeling,
as well as rational, creatures; and they highlight the ways in which agents are both
psychically internally differentiated and socially differentiated from others.
One of the crucial concerns of relational approaches is to investigate the impli-
cations for our understanding of autonomy of this richer conception of agency. For
example, conceptualizing agents as emotional, embodied, desiring, creative, and
feeling, as well as rational, creatures highlights the importance to autonomy of fea-
tures of agents that have received little discussion in the literature, such as memory,
imagination, and emotional dispositions and attitudes. Recognizing that agents are
both psychically internally differentiated and socially differentiated from others calls
for a reconceptualization of certain notions that are central to the literature on au-
tonomy, such as integration, identification, critical reflection, and self-realization.
22 Introduction

Finally, analyzing the way in which socialization and social relationships impede or
enhance an agent's capacities for autonomy has drawn attention to the connections
among an agent's self-conception, her social context, and her capacities for auton-
omy. It is for this reason that many relational approaches investigate the relationship
between autonomy and feelings of self-respect, self-worth, and self-trust.
A second concern of relational approaches is to analyze the specific ways in
which oppressive socialization and oppressive social relationships can impede au-
tonomous agency, at three interrelated levels. The first level is that of the processes of
formation of an agent's desires, beliefs, and emotional attitudes, including beliefs
and attitudes about herself. Relational approaches are particularly concerned with
analyzing the role that social norms and institutions, cultural practices, and social re-
lationships play in shaping the beliefs, desires, and attitudes of agents in oppressive
social contexts. The second level is that of the development of the competencies
and capacities necessary for autonomy, including capacities for self-reflection, self-
direction, and self-knowledge. Diana Meyers's work has been particularly influential
in analyzing the ways in which oppressive social environments can impair agents'
autonomy at this level. The third level is that of an agent's ability to act on autono-
mous desires or to make autonomous choices. Autonomy can be impeded at this
level not just by overt restrictions on agents' freedom but also by social norms, in-
stitutions, practices, and relationships that effectively limit the range of significant
options available to them. 60
These two concerns indicate that agents and their capacities should be con-
ceived relationally. There are two ways of understanding this idea. One regards
agents as intrinsically relational because their identities or self-conceptions are con-
stituted by elements of the social context in which they are embedded. Another re-
gards agents as causally relational because their natures are produced by certain his-
torical and social conditions. The distinction between intrinsically and causally
relational agency points to a corresponding distinction between relational concep-
tions of autonomy. Broadly, relational conceptions can be divided into constitutively
(or intrinsically) relational conceptions and causally relational conceptions.61 Those
approaches focusing on the social constitution of the agent or the social nature of
the capacity of autonomy itself, are constitutive conceptions, whereas those focusing
on the ways in which socialization and social relationships impede or enhance au-
tonomy are causal conceptions. The essays in this volume grapple with both kinds of
relational autonomy. They also grapple with the connected question of whether
procedural theories are sufficient to satisfy the demands of a relational approach to
autonomy or whether a more normative account is necessary. However, there is no
convergence among the contributors toward a single conception of relational au-
tonomy, nor does thete need to be. These questions arc still open as far as relational
approaches are concerned.
The articles are grouped under two main sections: "Autonomy and the Social"
and "Relational Autonomy in Context." The articles in the first group explore a
range of theoretical issues for conceptions of autonomy, agency, and moral respon-
sibility that arise from the social dimensions of selfhood, particularly in oppressive
social contexts. The articles in the second group examine the ways in which a rela-
tional approach to autonomy can contribute to a better understanding of autonomy
 Autonomy Refigured 23

and impaired autonomy in a variety of concrete contexts, for example, in health-

care practice and decision making or in legal debates over freedom of expression.
Marilyn Friedman's and Linda Barclay's articles, which open the first group, ex-
amine the implications of feminist discussions of the social self for procedural ac-
counts of autonomy. If the agent is socially constituted, as many feminists believe,
capacities of the agent like autonomy are also constitutively social and relational.
Whereas Friedman and Barclay are broadly sympathetic to the notion of the social
self, they take issue with some accounts of it and with the conclusions about auton-
omy that some feminists have drawn from these accounts. Friedman's article, "Au-
tonomy, Social Disruption, and Women," addresses the question of whether auton-
omy is antithetical to women's interests because it has the potential to disrupt social
relationships, in particular those that historically have been central to women's lives
and identities. Friedman argues that although popular conceptions of autonomy
may seem to support this view, procedural autonomy is not intrinsically disruptive
of social relationships, nor need it be inconsistent with values that arise from social
relationships, such as love, loyalty, or trust. Although an exercise of procedural au-
tonomy may sometimes disrupt social relationships, this need not always be prob-
lematic because in oppressive social circumstances, achieving autonomy may require
the rejection of oppressive social relationships.
Barclay's article, "Autonomy and the Social Self," asks whether autonomy is
compatible with the social self. Barclay distinguishes three different versions of the
social self thesis: the socially determined self, the motivationally social self, and the
constitutively social self. She argues that whereas a suitably nuanced conception of
procedural autonomy is compatible with plausible versions of the socially deter-
mined and the motivationally social self, it is incompatible with strong versions of
the socially determined and the constitutively social self. However, this incompati-
bility does not undermine the value of autonomy since these strong versions of the
social self are implausible. Moreover, feminists should be wary of an alliance with
communitarianism since the communitarian view that we are so constituted by so-
cial relations and shared values that we are unable to reconsider our attachment to
them is incompatible with a feminist commitment to communities of choice.
The articles by Paul Benson and Natalie Stoljar both investigate the adequacy of
procedural views to explain the impairment of agents' capacities for autonomy and
moral responsibility in oppressive social contexts. In his article, "Feeling Crazy: Self-
Worth and the Social Character of Responsibility," Benson argues that a sense of
self-worth is an additional necessary condition of moral responsibility. The sense of
self-worth is a sense of one's worthiness to account for one's actions to others, and
hence moral responsibility depends conceptually on relations with others. Moreover,
responsibility is relational in two further respects: publicly shareable norms regulate
both the moral appraisal of an agent's actions and the account of her actions that she
provides to others; and an agent's conception of herself as competent to account for
her actions is dependent on her having the status of an eligible participant in the
moral community. Thus Benson's account of moral responsibility, like his account
of autonomy in "Free Agency and Self-Worth," offers a constitutively relational
analysis of the relevant capacity of agents.
Natalie Stoljar's article, "Autonomy and the Feminist Intuition," argues that
24 Introduction

there is a tension between feminist intuitions regarding choices in oppressive con-

texts and a commitment to procedural autonomy. She extracts five conditions from
procedural accounts of autonomy and argues that none fully explains the widely
held feminist intuition that preferences and decisions motivated by oppressive
norms are not autonomous. She argues that only a version of a strong substantive
account of autonomy will do the job of explaining the intuition. Like Benson's,
Stoljar's article explores the nature of the capacity for autonomy rather than the na-
ture of the agent who exercises the capacity.
The articles by Genevieve Lloyd, Catriona Mackenzie, and Diana Meyers re-
turn to the issue of the social dimensions of selfhood. Gcnevieve Lloyd's article,
"Individuals, Responsibility, and the Philosophical Imagination," explores the phe-
nomenon of individuals taking responsibility for actions or events that lie beyond
the scope of their agency, and suggests that understanding this phenomenon re-
quires a reconceptualization of notions of individual identity and moral responsi-
bility. Lloyd considers three different examples of such taking of responsibility: in
friendship, in solidarity, and in historical responsibility for the actions of one's fore-
bears. She argues that all three examples challenge the grip of standard philosophical
conceptions of the individual, in particular conceptions of selves as sharply bor-
dered units and of relations between individuals as contractual and voluntary. Lloyd
argues that we forge our identities through relations of sympathetic identification
with others. Moreover, the narratives of our individual lives, as they unfold in time,
are not discrete but overlap with the narratives of others, including the narratives of
the collective existence and its memories into which we are born. Thus Lloyd draws
attention to an important connection between the social and temporal dimensions
of selfhood.
The following three articles, by Catriona Macken/Je, Diana Meyers, and Lor-
raine Code, take up some of the issues raised by Lloyd. The temporal and social di-
mensions of selfhood that concern Lloyd are central to an analysis of self-definition
developed by Mackenzie in "Imagining Oneself Otherwise." Mackenzie's main con-
cern, however, is to explore the connections among autonomy, selfhood, and imag-
ination. She proposes that agents' abilities to imagine themselves otherwise play an
important but overlooked role in practical deliberation and self-definition. Under-
standing this role can provide an explanation of one way in which oppressive social
environments can impair the autonomy of oppressed agents, namely through a re-
stricted cultural imaginary that curtails agents' abilities to imagine alternate social
and individual possibilities.
Diana Meyers's article, "Intersectional Identity and the Authentic Self: Oppo-
sitcs Attract?" takes up Lloyd's concern with understanding the implications of col-
lective and social group identities for individual identity. Meyers's main concern is
whether an integration account of the authentic self, of the kind she developed in
her earlier work on autonomy, can accommodate selves whose identities are inter-
sectional. She argues that although standard accounts of integration, of the kind
proposed by Frankfurt, are inconsistent with intersectional conceptions of the self,
her own model of dynamic integration is compatible with these conceptions. In-
deed, she argues, acknowledging the intersectionality of one's identity is crucial for
achieving the self-knowledge that is necessary for authenticity. Meyers's article di-
 Autonomy Refigured 25

rectly tackles the postmodernist and diversity critiques of autonomy. In doing so, it
explores how recognition of the internally and socially differentiated identities of
agents calls for a reconceptualization of notions like integration, critical reflection,
and self-realization.
The concern of many of the contributors to this volume is to provide an analy-
sis of the effects of oppressive social contexts on individuals' capacities for auton-
omy and moral responsibility. In "The Perversion of Autonomy and the Subjection
of Women: Discourses of Social Advocacy at Century's End," Lorraine Code turns
away from this concern to provide a detailed analysis of the broader social and epi-
stemic contexts within which autonomy is an issue for oppressed agents. In her
analysis of these contexts she points to a tension or paradox concerning autonomy.
On the one hand, the social imaginary of Western societies is dominated by a hy-
perbolized ideal of autonomy as self-sufficient individualism. This ideal is complicit
with and underpins social structures of oppression and subjection in which those
who do not attain self-sufficient autonomy are marginalized. On the other hand, in
the struggle of the oppressed against their oppression, autonomy is a legitimate as-
piration. The issue that concerns Code is how to develop a model of autonomy that
answers to this aspiration. She sees the problem as a moral and epistemological one:
those who are subject to oppression are less likely to achieve autonomy, partly be-
cause they are marginalized and delegitimized as knowers, even with regard to their
own experiences. Code suggests that a cluster of moral and epistemic practices of
what she calls "advocacy" on behalf of the oppressed may provide a response to this
problem.
The issues raised by Code about the epistemic, social, and political conditions
necessary for autonomy are taken up in very concrete ways by the four articles in the
second section of the book, "Relational Autonomy in Context." Those by Susan
Dodds, Anne Donchin, and Susan Sherwin and Carolyn McLeod are primarily con-
cerned with health-care ethics. In "Choice and Control in Bioethics," Dodds takes
issue with the way in which autonomy is narrowly conceptualized in much bioethi-
cal literature as equivalent to informed consent. She argues that one effect of this
conception is to restrict the exercise of autonomy in health-care contexts to making
choices among a limited set of options. Dodds argues that criticisms of this model
by liberal, radical, and cultural feminists have been important in highlighting some
of its shortcomings. However, she believes that a relational model of autonomy that
draws on and develops some of these feminist criticisms provides the best basis for
an appropriate conception of autonomy in health-care contexts.
Anne Donchin's article, "Autonomy and Interdependence: Quandaries in Ge-
netic Decision Making," uses the example of genetic decision making to bring out
the tensions between interpersonal dependencies and prevailing accounts of auton-
omy in bioethics. These tensions point to the need for a relational approach to au-
tonomy in bioethics more broadly. Donchin makes an important distinction be-
tween weak and strong relational approaches to autonomy. The former acknowledge
that selfhood and capacities for autonomy are developed in the context of social
relationships. But they see social ties as only contingently related to agents' self-
understandings and tend to regard social relationships as entirely voluntary. Focus-
ing on a selection of case studies of individuals and families involved in genetic de-
26 Introduction

cision making, Donchin shows that the complex interplay between individual au-
tonomy and biological and social relationships, particularly nonvoluntary ones, re-
quires the development of a strong relational model. This model recognizes that au-
tonomy is not solely an individual enterprise and that respect for the autonomy of
others requires collaboration, long-term reciprocity, and equitable balancing of
power relationships.
The article by Susan Sherwin and Carolyn JVlcLeod, "Relational Autonomy,
Self-Trust, arid Health Care for Patients Who Are Oppressed," analyzes one dimen-
sion of the autonomy-undermining effects of oppression, namely, its effects on
agents' self-trust. Sherwin and McLeod distinguish three different kinds of self-
trust: trust in one's capacity to choose and decide effectively, trust in one's ability to
act on the decisions one makes, and trust in one's own judgment. All three kinds,
they maintain, are necessary for the exercise of autonomy competency. Further-
more, because self-trust in all three senses is relational, both causally and constitu-
tively, it can be eroded to varying degrees by oppressive social situations. Sherwin
and McLeod illustrate the connections among impaired autonomy, oppression, and
self-distrust by considering a range of different cases in which patients' self-distrust,
arising from oppressive social situations, impairs their abilities to make autonomous
health-care choices.
One of the recurring themes in the articles by Dodds, Donchin, and Sherwin
and McLeod is the importance of a range of significant options to agents' abilities
to exercise autonomous choice. Susan Brison's article, "Relational Autonomy and
Freedom of Expression," is also concerned with the connections among autonomous
choice, agents' relations with others, and the range of options that arc socially and
culturally available to agents. In explaining this connection, Brison draws on and de-
velops Amartya Sen's capability account of human flourishing. She then considers
the implications of the capability account of autonomy for debates about hate
speech. A prominent defense of the right to free speech is that it is grounded in a
right to autonomy. Such a right, it has been argued, precludes restrictions on hate
speech. Brison argues that the capability account of autonomy does not support this
conclusion. Speech does not: always enhance autonomy, but it can in some circum-
stances undermine autonomy by reducing the significant options available to agents,
affecting agents' beliefs about themselves, their preferences and options, and affect-
ing others' beliefs about and attitudes toward agents.
In articulating relational approaches to autonomy, the articles collected in this
volume challenge and enrich contemporary philosophical debates about auton-
omy, while showing why autonomy is an important value for feminists to retain
and defend.

Notes

1. As far as we know, Jennifer Nedelsky was the first to articulate a conception of rela-
tional autonomy from an explicitly feminist perspective. See Nedelsky, "Rcconceiving Au-
tonomy: Sources, Thoughts and Possibilities," Yale Journal, of Law and Feminism 1 (1989):
7 36; "Law, Boundaries and the Bounded Self," Representations 30 (1990): 162-189; "Recon-
 Autonomy Refigured 27

ceiving Rights as Relationship," Review of Constitutional Studies 1 (1993): 1-26; and "Med-
itations on Embodied Autonomy," Graven Images 2 (1995): 159-170. For a recent exposi-
tion of the varieties of possible conceptions of relational autonomy, see Marilyn Friedman,
"Autonomy and Social Relationships: Rethinking the Feminist Critique," in Feminists Re-
think the Self, ed. DianaTietjens Meyers (Boulder, Colo.: Westview, 1997), pp. 55-58.
2. For example, Seyla Benhabib, Situating the Self: Gender, Community and Postmod-
ernism in Contemporary Ethics (New York: Routledge, 1992); Michael Sandel, Liberalism and
the Limits of Justice (Cambridge: Cambridge University Press, 1982); Will Kymlicka, Con-
temporary Political Philosophy (Oxford: Oxford University Press, 1990).
3. For example, for autonomy-based defences of euthanasia, see Ronald Dworkin, Life's
Dominion: An Argument about Abortion and Euthanasia (New York: Vintage, 1994), and Joel
Feinberg, "Voluntary Euthanasia and the Inalienable Right to Life," Philosophy and Public
Affairs 1 (1978). For autonomy-based discussions of surrogacy, see Susan Dodds and Karen
Jones, "Surrogacy and Autonomy," Bioethics 3 (1989): 1-17, and Laura Purdy, "A Response
to Dodds and Jones," Bwethicsl (1989): 40-44.
4. Susan J.Brison, "The Autonomy Defense of Free Speech," Ethics 108 (1998): 312-339.
5. Marilyn Friedman, "Autonomy and Social Relationships."
6. For other discussions of feminist critiques of autonomy, see John Christman, "Fem-
inism and Autonomy," in Nagging Questions: Feminist Ethics in Everyday Life, ed. Dana
Bushnell (Lanham, Md: Rowman & Littlefield, 1995), and Friedman, "Autonomy and Social
Relationships."
7. Lorraine Code, "Second Persons," in What Can She Know? Feminist Theory and the
Construction of Knowledge (Ithaca, N.Y.: Cornell University Press, 1991).
8. Ibid., p. 78.
9. Annette Baier introduces the notion of second persons in "Cartesian Persons," in
Postures of the Mind: Essays on Mind and Morals (Minneapolis: University of Minnesota Press,
1985).
10. This critique parallels certain well-known communitarian critiques of liberalism.
Communitarians have argued that liberalism presupposes a flawed account of human nature,
namely, that agents are (metaphysically) atomistic or unencumbered by the social contexts in
which they are embedded. See note 2 above.
11. Baier, "Cartesian Persons," p. 85.
12. For more discussion, see Friedman, "Autonomy and Social Relationships."
13. Alison Jaggar, Feminist Politics and Human Nature (Totowa, N.J.: Rowman & Al-
lanheld 1983), p. 29.
14. See Natalie Stoljar, "Essence, Identity and the Concept of Woman," Philosophical
Topics 23 (1995): 261-294, and Charlotte Witt, "Anti-Essentialism in Feminist Theory,"
Philosophical Topics23 (1995): 321-344, for discussions of the possibility of relations being
essential to individual identity in a metaphysical sense.
15. Louise Antony, "Is Psychological Individualism a Piece of Ideology?" Hypatia 10
(1995): 157-174, 167. Also see Antony, "Sisters, Please, I'd Rather Do It Myself: A Defense
of Individualism in Feminist Epistemology," Philosophical Topics 23 (1995): 59-94.
16. Although, as Marilyn Friedman points out in her discussion of what we are calling
care critiques, these critiques do often slide between rejecting autonomy and rejecting mas-
culinist conceptions of autonomy. Friedman, "Autonomy and Social Relationships."
17. Nancy Chodo row, The Reproduction of Mothering (Berkeley: University of Califor-
nia Press, 1978) and "Gender, Relation and Difference in Psychoanalytic Perspective" in The
Future of Difference, ed. Hester Eisenstein and Alice Jardine (New Brunswick, N.J.: Rutgers
University Press, 1985). See also Carol Gilligan, In a Different Voice: Psychological Theory and
Women's Development (Cambridge, Mass.: Harvard University Press, 1982).
28 Introduction

18. Virginia Held, Feminist Morality (Chicago: University of Chicago Press, 1993), p. 60.
19. Sarah Lucia Hoagland, Lesbian Ethics: Toward New Value (California: Institute of
Lesbian Studies, 1988), pp. l43ff.
20. Nedelsky, "Rcconceiving Autonomy," p. 12; Held, Feminist Morality, pp. 61-62.
21. Evelyn Fox Keller, Reflections on Gender and Science (New Haven, Conn.: Yale Uni-
versity Press, 1985), chap. 5.
22. Ibid., p. 99.
23. Keller's discussion refers to Jessica Benjamins psychoanalytic work on erotic domi-
nation. See especially "The Bonds of Love: Rational Violence and Erotic Domination," in
Liscnstein and Jardine, The Future of Difference.
24. Jean Grimshaw, "Autonomy and Identity in Feminist Thinking," in Feminist Per-
spectives in Philosophy, ed. Morwcnna Griffiths and Margaret Whitford (Bloomington: Indi-
ana University Press, 1988).
25. Judith Butler, Gender Trouble: Feminism and the Subversion of Identity (New York:
Routledge, 1990); Drucilla Cornell, Transformations: Recollective Imagination and Sexual Dif-
ference (New York: Routledge, 199.3), chap. 2; Jane Flax, "Postmodernism and Gender Rela-
tions in Feminist Theory," Signs 12 (1987): 621-643; Susan Hekman, "Subjects and Agents:
The Question for Feminism," in Provoking Agents: Gender and Agency in Theory and Practice,
ed. J. K. Gardiner (Champagnc-Urbana: University of Illinois Press, 1995); "Reconstructing
the Subject: Feminism, Modernism and Postmodernism," Hypatia 6 (1991): 44-63; and
"Review of Self, Society and Personal Choice" HypatiaG (1991): 222-225; Diana'!'. Meyers,
"Personal Autonomy or the Deconstructed Subject? A Reply to Hekman," Hypatia 7 (1992):
124-132.
26. Benhabib, Situating the Self, chap. 5; Butler, Gender Trouble; Cornell, Transforma-
tions; Code, "Second Persons"; Flax, "Postmodernism and Gender Relations"; Hekman, "Re-
constructing the Subject."
27. Hekman, "Reconstructing the Subject."
28. The question of whether postmodernism is able to provide an adequate account of
agency is the subject of an important debate in contemporary feminist theory. We arc unable
to address this debate here. However, we are broadly sympathetic with Seyla Benhabib's dis-
cussion of this issue in Situating the Self, chap. 7. Benhabib criticizes feminist appropriations
of the rhetoric of the "death of the subject," such as those of Flax ("Postmodernism and
Gender Relations") and Butler (Gender Trouble), on the grounds that they are ultimately in-
coherent and that feminism cannot afford to reject notions such as agency, autonomy, and
selfhood.
29. Kimberle Williams Crcnshaw, "Mapping the Margins: Intersectionality, Identity
Politics, and Violence Against Women of Color," in Critical Race Theory: The Key Writings,
cd. Crenshaw et al. (New York: New Press, 1995), p. 358.
30. Maria Lugoncs, "Playfulness, 'World'-Travelling, and Loving Perception," Hypatia
2 (1987): 3-19.
31. Susan J. Brison, "Outliving Oneself: Trauma, Memory and Personal Identity," in
Feminists Rethink the Self, ed. Diana Tietjens Meyers (Boulder, Colo.: Westview, 1997), 12 —
39.
32. Crenshaw, "Mapping the Margins," p. 377.
33. For discussions of the notion of content neutrality, see John Christman, Auton-
omy and Personal History," Canadian journal of Philosophy 7.\ (1991): 1-24, and Paul Ben-
son, "Free Agency and Self- Worth," Journal of Philosophy 91 (1994): 650-668.
34. We take the distinction between structural and historical theories from Paul Ben-
son, "Free Agency and Self-Worth," pp. 653-654.
 Autonomy Refigured 29

35. Harry Frankfurt, "Freedom of the Will and the Concept of a Person," Journal of
Philosophy 68 (1971): 5-20. See also S. I. Benn, "Individuality, Autonomy and Commu-
nity," in Community as a Social Ideal, ed. Eugene Kamenka (New York: St Martin's Press,
1982), and "Freedom, Autonomy and the Concept of a Person," Proceedings of the Aris-
totelian Society 76 (1976): 109-130; Gerald Dworkin, "Acting Freely," Nous 3 (1970):
367-383, and The Theory and Practice of Autonomy (New York: Cambridge University Press,
1988); Joel Feinberg, "Autonomy," in The Inner Citadel: Essays on Individual Autonomy, ed.
John Christman (New York: Oxford University Press, 1989); Robert Young, "Autonomy and
Socialisation," MW84 (1980): 565-576, and Personal Autonomy: Beyond Negative and Pos-
itive Liberty (London: Groom Helm, 1986).
36. Gary Watson, "Free Agency," Journal of Philosophy 72 (1975): 205-220. In re-
sponse to this objection, Watson develops his own variant of the hierarchical approach by
distinguishing between a person's motivational system and her valuational system. Although
these systems must overlap, the problem of free agency arises because they can sometimes
come apart. Watson identifies the agent's point of view, or the aspects of the self that are truly
"one's own," with the agent's valuational system. An agent acts freely when her motivational
and valuational system are integrated; she acts nonautonomously or unfreely when her ac-
tions conflict with her valuational system. However, this account faces problems of its own in
explaining the special autonomy-conferring authority accorded to a person's values. Watson
does make a distinction between values that are the product of acculturation (acculturated at-
titudes) and values that are judgments, that is, values that have been subjected to processes
of rational critical assessment, claiming that only the latter are autonomy conferring. But he
does not provide an account of the processes of assessment that distinguish values that are
judgments from mere acculturated attitudes, nor does he explain how these processes of as-
sessment confer autonomy on those values that have been scrutinized in this way. For further
discussion, see Harry Frankfurt, "Identification and Wholeheartedness," in Responsibility,
Character and the Emotions, ed. Ferdinand Schoeman (New York: Cambridge University
Press, 1987).
37. Irving Thalberg, "Hierarchical Analyses of Unfree Action," Canadian Journal of
Philosophy^ (1978): 211-225.
38. See Diana T. Meyers, Self, Society, and Personal Choice (New York: Columbia Uni-
versity Press, 1989), for a similar objection. We would argue that this metaphysical objection,
as discussed by Friedman, Meyers, and within the literature more generally, arises from cer-
tain misunderstandings of hierarchical models. First, the idea that hierarchical views involve
an ontological commitment to a metaphysically distinct "true self" there to be discovered
misrepresents the process of critical reflection. On our reading of hierarchical theories, criti-
cal reflection should be understood as a process of self-constitution. As Frankfurt makes clear
(especially in "Identification and Wholeheartedness"), the self is not a static, deep metaphys-
ical entity but is constantly being reconstituted through the processes of reflection, delibera-
tion, and decision, by means of which we define our cares, commitments, and values. Sec-
ond, as Susan Wolf has argued convincingly, the metaphysical problem conflates moral
responsibility for the self with metaphysical responsibility for the self: Wolf, "Sanity and the
Metaphysics of Responsibility," in Responsibility, Character and the Emotions, ed. Ferdinand
Schoeman (New York: Cambridge University Press, 1987), and Freedom within Reason (New
York: Oxford University Press, 1990). Autonomy requires only that we are morally responsi-
ble, or capable of self-revision. It does not matter that our desires may arise from sources
within or external to ourselves that we do not control, nor does it matter that we did not in-
vent the values that guide our lives. What matters is that we are capable of revising ourselves
in the light of critical reflection. Thus, as far as the metaphysical problem is concerned, there
30 Introduction

is no significant difference between Friedman's integration model and hierarchical models.

Both are attempting to provide accounts of self-revision and both see some kind of integra-
tion among the elements of the self as central to autonomy. Where they differ is in their ac-
counts of critical reflection.
39. Marilyn Friedman, "Autonomy and the Split-Level Self," Southern Journal of'Phi-
losophy^ (1986): 19-35, 32.
40. Paul Benson, "Autonomy and Oppressive Socialization," Social Theory and Practice
17 (1991): 385-408, 395.
41. Dworkin, Theory and Practice of Autonomy, chap. 1.
42. John Christman, "Autonomy: A Defense of the Split-level Self," Southern Journal of
Philosophy^ (1987): 281-293.
43. Chisr.man, "Autonomy and Personal History"; sec also "Feminism and Autonomy."
44. Benson, "Autonomy and Oppressive Socialization."
45. Meyers, Self, Society, and Personal Choice, and "Personal Autonomy and the Paradox
of Feminine Socialization" Journal of Philosophy 84 (1987): 619-628.
46. See Meyers's article in this volume for a development of this account of integration
in the context of an intcrsectional approach to identity.
47. For a liberal approach to self-realization, see, for example, John Rawls, A Theory of
justice (Cambridge, Mass.: Harvard University Press, 1971). For an Aristotelian approach,
see, for example, Martha C. Nussbaum, "Human Capabilities, Female Fiuman Beings," in
Women, Culture and Development: A Study of Human Capabilities, ed. Martha C. Nussbaum
and Jonathan Glover (Oxford: Clarendon Press, 1995), and Amartya Sen, Inequality Reex-
amined (Cambridge, Mass.: Harvard University Press, 1992).
48. Meyers, Self, Society, and Personal Choice, p. xi.
49. These questions arise repeatedly in debates in the feminist literature about cosmetic
surgery, surrogacy, and reproductive technologies. On cosmetic surgery, see Kathryn Morgan,
"Women and the Knife: Cosmetic Surgery and the Colonization of Women's Bodies," Hy-
patia 6 (1991): 25-53. On surrogacy, see Dodds and Jones, "Surrogacy and Autonomy," and
Purdy, "A Response to Dodds and Jones." On reproductive technologies, see Margaret Radin,
"Market Inalienability," Harvard Law Review 100 (1987): 1849-1937; Marjorie Schultz,
"Reproductive Technology and Intent-based Parenthood: An Opportunity for Gender Neu-
trality," Wisconsin Law Review 2 (1990): 287-398; Suzanne Uniacke, "In Vitro Fertilization
and the Right to Reproduce," Bioethics 1 (1987): 241-254; Mary Anne Warren, "IVF and
Women's Interests: An Analysis of Feminist Concerns," Bioethics 2 (1988): 37-57.
50. We borrow this terminology from Benson, "Free Agency and Self-Worth."
5 1. Wolf, "Sanity and the Metaphysics of Responsibility."
52. Ibid., p. 54.
53. Thomas E. Hill, Jr., "The Kantian Conception of Autonomy," in Christman, Inner
Citadel. Philip Pettit and Michael Smith offer a different set of arguments for a version of a
Kantian theory of autonomy in "Backgrounding Desire," Philosophical Review 94 (1990):
565-592; "Practical Unreason," Mind 102 (1993): 53-79; and "Freedom in Belief and De-
SIK" Journal of Philosophy: 93 (1996): 429-449.
54. Paul Benson, "Freedom and Value," Journal of Philosophy 84 (1987): 465-486;
"Feminist Second Thoughts About Free Agency," Hypatia 3 (1990): 47-64; and "Autonomy
and Oppressive Socialization."
55. Benson, "Free Agency and Self-Worth," p. 656.
56. Ibid., p. 660.
57. See Benson's discussion of content neutrality and content specificity in "Free
Agency and Self-Worth," pp. 663-665.
 Autonomy Refigured 31

58. Robin Dillon, "Toward a Feminist Conception of Self-Respect," Hypatial (1992):

52-69, and Trudy Govier "Self-trust, Autonomy and Self-Esteem," HypatiaS (1993): 99-120.
59. See Benson's discussion of this issue in "Free Agency and Self-Worth," p. 665.
60. The importance of options to autonomy is emphasized by Joseph Raz, The Moral-
ity of Freedom (Oxford: Clarendon Press, 1986), pp. 205, 373ff.
61. Friedman, "Autonomy and Social Relationships," pp. 57-58.
This page intentionally left blank
 I
AUTONOMY AND
THE SOCIAL
This page intentionally left blank
 1
AUTONOMY, SOCIAL DISRUPTION,
AND WOMEN1

Marilyn Friedman

Of Autonomy and Men

Are women in Western societies alienated by the ideal of autonomy? Many feminist
philosophers have recently suggested that women find autonomy to be a notion in-
hospitable to women, one that represents a masculine-style preoccupation with self-
sufficiency and self-realization at the expense of human connection.
Paul Gauguin's life epitomizes what many feminists take to be the masculine
ideal of autonomy. Gauguin abandoned his family and middle-class life as a banker
in Denmark to travel to Mediterranean France, Tahiti, and Martinique in search of
artistic subjects and inspiration. He deserted his wife and five children, one might
say, to paint pictures in sunny locales. Biographies of Gauguin's life reveal that he
agonized for some time over the decision to leave his family. He once wrote: "One
man's faculties can't cope with two things at once, and I for one can do one thing
only: paint. Everything else leaves me stupefied."2 Gauguin's self-reflective agonies, I
believe, would qualify as autonomous, according to many contemporary definitions.
How has Western culture assessed Gauguin's life and work? Gauguin was can-
onized by Western art history. Of course, he had the moral good luck to have
painted important pictures, something that Bernard Williams might call a "good for
the world."3 Whereas his fame is certainly based on his paintings and not on his fa-
milial desertion, nevertheless, the fact of his having left behind a wife and five chil-

35
36 Autonomy and the Social

dren for sunnier prospects has done nothing to tarnish his stature. If anything, it has
added a romantic allure to his biography.
Narratives of this sort suggest that autonomy in practice is antithetical to
women's interests because it prompts men to desert the social relationships on which
many women depend for the survival and well-being of themselves and their chil-
dren. In the past, because of women's restricted opportunities, the loss of support
suffered by abandoned women has often been worse than the heterosexual relation-
ships on which they depended.
Men are supposed to "stand up like a man" for what they believe or value, in-
cluding the simple assertion of their self-interests. Women arc instead supposed to
"stand by your man." The maxim "stand up like a womanl" has no serious meaning.
It conjures up imagery that is, at best, merely humorous. There is no doubt which
model of behavior as exhibited by which gender receives the highest honors in West-
ern public culture.
Still today, women in general define themselves more readily than men in terms
of personal relationships. In addition, women's moral concerns tend to focus more
intensely than those of men on sustaining and enhancing personal ties.4 Also, pop-
ular culture still presumes that women are more concerned than men to create and
preserve just the sorts of relationships, such as marriage, that autonomy-seeking
men sometimes want to abandon. 5 Feminist analysis has uncovered ways in which
close personal involvement and identification with others have been culturally de-
valued, in tandem with the devaluation of women, by comparison with the public
world of impersonal relationships that men have traditionally monopolized. 6 Fo-
cusing on the importance of the social is one feminist strategy for combating these
traditions of thought and for elevating social esteem for women. Many feminist
philosophers have thus emerged as champions of social relationships and of rela-
tional approaches to diverse philosophical concepts.7
The cultural understanding of autonomy needs to change if the concept is to
be relevant for women. I discuss three such changes: new paradigms of autonomy
that involve female protagonists, redefinitions of autonomy that avoid stercotypi-
cally masculine traits, and redefinitions of autonomy that somehow involve social
relationships or are at least not antithetical to them. Indeed such an account has
been under development for some time in philosophical literature, and my sugges-
tions on these points are not new. Of course, nothing guarantees a priori that we
will find an account of autonomy that synthesizes these elements consistently with
the core notion of self-determination that sets limits to our understanding of au-
tonomy. I am optimistic, however, that a female-friendly account of autonomy can
be, and has in part already been, developed.
At any rate, I mention these points merely to set the stage for my fourth, and
primary, thesis: at the same time that we embrace relational accounts of autonomy,
we should also be cautious about them. Autonomy increases the risk of disruption
in interpersonal relationships. Although this is an empirical and not a conceptual
claim about autonomy, nevertheless, the risk is significant. It makes a difference in
whether the ideal of autonomy is genuinely hospitable to women.
After providing a capsule characterization of autonomy that is typical of ac-
 Autonomy, Social Disruption, and Women 37

counts in the contemporary philosophical literature, I address each of the preceding

points in turn.

Personal Autonomy: A Capsule Account

Autonomy involves choosing and living according to standards or values that are, in
some plausible sense, one's "own." A plausible sense of "ownness" involves at least
two dimensions. First, someone must reflect in an autonomy-conferring manner on
the particular choices she makes and the standards or values by which she will be
guided. Autonomy-conferring reflection, in my view, is not confined to rational re-
flection. Such terms as "critical," "reflection," "consideration," "evaluation," "scru-
tiny," and "choice," as I use them, encompass emotional as well as strictly rational or
narrowly cognitive dimensions of personal processes.8
Second, the reflection itself must be relatively free of those varieties of interfer-
ence that impede the achievement of autonomy. What varieties these are I do not
specify except to say that socialization does not as such impede autonomy, whereas
coercion as such does do so.9 These circular terms ("autonomy-conferring reflection"
and "autonomy-impeding interference") are not meant to articulate the notions in
question but merely to serve as placeholders for a fully fleshed account of the nature
of autonomy, which is not my present concern.
For the most part, my discussion focuses on personal autonomy, something best
defined by reference to moral autonomy. Moral autonomy has to do with what one
regards either as morally required or as morally permissible. It involves choosing and
living according to rules that one considers to be morally binding. Personal auton-
omy involves acting and living according to one's own choices, values, and identity
within the constraints of what one regards as morally permissible.10

Of Autonomy and Women

First, I consider the historic association of autonomy with men. Autonomy, its con-
stituent traits, and the actions and lives that seem to manifest it are publicly es-
teemed much more often in men than in women. As noted earlier, the preponder-
ance of men in narratives of autonomy could easily cast a masculine shadow over
the concept.
Does a concept become irrevocably shaped by the paradigms that initially con-
figure its usage? I believe that it does not and that autonomy can accordingly be
freed of its historically near exclusive association with male biographies and male-
identified traits. Doing so will require systematic rethinking. In part, we need new
paradigms of autonomy that feature female protagonists.11
A particularly feminist appropriation of the concept of autonomy requires nar-
ratives of women who strive in paradigmatically or distinctively female situations
against patriarchal constraints to express and refashion their deepest commitments
and senses of self. Such narratives are already widely available. Susan Brison, for ex-
ample, writes of regaining autonomous control over her life after she was tragically
raped and almost murdered.12 Patricia Hill Collins explains the power and impor-
38 Autonomy and the Social

tance of self-definition to African-American women, who are fighting the dominat-

ing cultural images of themselves as mammies, matriarchs, and welfare mothers. 13
Minnie Bruce Pratt tells of how she struggled to live as a lesbian while at the same
time renouncing the racism and antisemitism that she had derived from her family
and community of origin.14
In addition, there are women's autonomy narratives that are not particularly
about overcoming patriarchal constraints. Sara Ruddick's account of maternal
thinking, for example, draws heavily on stories of women who reflected deeply on
how to care well for their children, an otherwise conventional female task.15 In
short, there is already available a large variety of narratives that exemplify women's
autonomous struggles, both feminist and nonfeminist.
It is, in addition, helpful to remember that autonomy is not always valued in
men. Whole groups of minority men have had their autonomous aspirations crushed
by white Western societies. Moreover, white men do not always tolerate autonomy
from one another. In traditional, patriarchal hierarchies, such as military or corpo-
rate structures, even many white men are routinely punished for being autonomous,
for challenging accepted norms and authoritative dictates, for not being a "team
player."16
Some male philosophers, in addition, criticize the ideal of autonomy in at least
some of its versions. Male communitarians challenge what they take to be the overly
individualistic and ungrounded autonomy of the liberal tradition. 17 Sounding a dif-
ferent note, Loren Lomasky regards autonomy as a source of "massive dislocation"
and "widespread human misery." He criticizes autonomy as a rallying cry of the
"Red Guards" and of proponents of the welfare state who reject the "traditional
ways" of family life.18
Thus, the historical link between autonomy and men is not uniform. It is being
further challenged today by the growing diversity of women's lives. Autonomy is
now available to, and sometimes celebrated in, women, and it is not always cele-
brated in men. The gender paradigms of autonomy are shifting. On the basis of par-
adigms alone, autonomy is no longer straightforwardly male-oriented or alien to
women.

Autonomy and Gender Stereotypes

My second point is that we should seek redefinitions of autonomy that avoid stereo-
typically masculine traits. Autonomy has often been conceptualized in terms of
traits that suggest an antifemale bias. Traditional ideals of autonomy, for example,
have been grounded in reason. Genevieve Lloyd and others have argued that tradi-
tional conceptions of reason have excluded anything deemed "feminine," such as
emotion.19 The exclusion of emotion from the concept of reason, however, is less
prominent a view today than it once was. Some recent accounts of rationality by
both feminists and mainstream philosophers blur the boundary between reason and
emotion and thus promise to undermine this traditional dichotomy.20 In case those
accounts prove to be well grounded, this particular philosophical basis for thinking
that autonomy is an antifemale ideal would have been eliminated.21
Besides connecting autonomy to reason, popular Western culture has also asso-
 Automomy, Social Disruption, and Women 39

elated autonomy with other masculine-defined traits, for example, independence

and outspokenness.22 Traits popularly regarded as feminine, by contrast, have no dis-
tinctive connection to autonomy—social interactiveness, for example.23 Thus
popular gender stereotypes have associated autonomy with men but not with
women; these stereotypes might invidiously infect philosophical thinking about
autonomy.
To be sure, because of gender differences in socialization, autonomy might ac-
tually occur less often in women than in men. As Diana Meyers has documented,
male socialization still promotes autonomy competency more effectively than does
female socialization.24 Overall, men have had far greater opportunities than women
to act and live autonomously.25 Such modes of action and living have in the past
been closed to most women because they required unavailable (to women) resources
such as political power, financial independence, or the freedom to travel unmolested
in public space—to jog safely, for example, through New York City's Central Park.
The more frequent appearance of autonomy in men than in women, com-
bined with the association of stereotypically masculine but not feminine traits with
autonomy, might unwittingly bias philosophical investigations of autonomy. To-
gether with their nonphilosophical peers, philosophers might fail to recognize
manifestations of autonomy by women. Philosophers who try to conceptualize au-
tonomy might do so with autonomous males in mind as paradigm cases. They
might go on to mistake what are merely masculine traits for the traits that make up
autonomy competency as such. Thus contemporary philosophical accounts of au-
tonomy should be scrutinized particularly with a view to eliminating any covert
masculine paradigms that might lie behind them.
In addition to creating a male bias that might influence philosophical reflec-
tions on autonomy, male stereotypes are also easy to exaggerate in ways that could
further distort the conception of autonomy. The male-stereotyped traits of inde-
pendence and self-sufficiency have often been interpreted, both in general culture
and in philosophical traditions, in asocial, atomistic terms that seem to sanction de-
tachment from close personal relationships with others.26 Many feminists have ar-
gued that this illusory goal of atomistic self-sufficiency has indeed structured male
development and male perspectives in those cultures that require men to repudiate
the feminine to consolidate their own masculine gender identity.27 Some feminists
worry that the very concept of autonomy has been irremediably contaminated by
this atomistic approach, which neglects the social relationships that are vital for de-
veloping the character traits required for mature autonomy competency.28 Much of
that socialization consists of women's traditional child-care labor.
Philosophical accounts might err in this regard more by omission than by com-
mission. Some contemporary accounts, for example, fail to mention how the human
capacity for autonomy develops in the course of socialization.29 By neglecting to
mention the role of socialization in the development of mature autonomy com-
petency, traditional accounts of autonomy ignore one crucial way in which au-
tonomous persons are ultimately dependent persons after all, and in particular,
dependent on women's nurturing. This philosophical omission does nothing to
undermine the conceited cultural illusion of the "self-made man" as a paradigm of
autonomy.
40 Autonomy and the Social

To be sure, no respectable philosopher today would explicitly deny that a social

upbringing and ongoing personal interaction are necessary to become autonomous.
These conditions impart the self-concept and resources for critical reflection that au-
tonomy requires. Also, no respectable philosopher would deny that women's labors
still make up the lion's share of child care, especially in the crucial years of early
formative socialization. Careful philosophical thought on these issues should correct
the pop-cultural view of some men as impossibly "self-made," a view that denies
women their proper share of credit for nurturing or supporting the autonomy com-
petency found in those men. The point is that philosophers must actively take pains
to weed out inappropriate male paradigms that might contaminate their own or a
wider cultural understanding of key philosophical notions.
In virtue of disregarding the fundamentally social nature of autonomy and au-
tonomous persons, the myth of the self-made man rests on a mistake. The fact that
mistaken conceptions of autonomy are male-biased, however, does not show that
autonomy properly understood is male-biased or antifemale.

Social Reconceptualizations of Autonomy

My third point is that we need an account of autonomy that brings out its relational
character. Fortunately, a relational approach to autonomy has been emerging for
some time. Two developments are relevant to this issue: a procedural conception of
autonomy and a relational or intersubjective conception of autonomy.30
According to a procedural account, personal autonomy is realized by the right
sort of reflective self-understanding or internal coherence along with an absence of
undue coercion or manipulation by others. Autonomy, in this sort of view, is not a
matter of living substantively in any particular way.31 My own capsule account of
autonomy in the previous section is a procedural account. Although this sort of ac-
count can be debated,32 it is nevettheless common in philosophy today. In a proce-
dural conception, avoiding or abandoning close personal relationships is in no sense
required by autonomy. Nor is it for any reason inherently a better way for any indi-
vidual to strive for autonomy.
Although the language of autonomy in popular culture might still suggest aso-
cial atomistic images of the self-made man, academic philosophers now seldom
share this view. The atomistic "self-made" conception of autonomy is a substantive
conception of a particular sort of life or mode of behaving that someone must
choose in order to realize autonomy. Such an ideal falls outside the bounds of pro-
cedural accounts of autonomy. It is not a proper part of them.
In addition to focusing mainly on procedural matters, many contemporary
philosophers of autonomy have also tended to gravitate toward relational or inter-
subjective accounts of autonomy. This is true of both feminist and mainstream
philosophers. At present, both construe autonomy in social, relational, interper-
sonal, or intersubjective terms.33
According to the relational approach, persons are fundamentally social beings
who develop the competency for autonomy through social interaction with other
persons. These developments take place in a context of values, meanings, and modes
of self-reflection that cannot exist except as constituted by social practices. Also, ac-
 Automomy, Social Disruption, and Women 41

cording to some theorists, autonomy is itself the capacity for a distinctive form of
social and, in particular, dialogical engagement.34
Autonomy is no longer thought to require someone to be a social atom, that is,
radically socially unencumbered, defined merely by the capacity to choose, or to be
able to exercise reason prior to any of her contingent ends or social engagements.35
It is now well recognized that our reflective capacities and our very identities are al-
ways partly constituted by communal traditions and norms that we cannot put en-
tirely into question without at the same time voiding our very capacities to reflect.
We are each reared in a social context of some sort, typically although not al-
ways that of a family, itself located in wider social networks such as community and
nation. Nearly all of us remain, throughout our lives, involved in social relationships
and communities, at least some of which partly define our identities and ground our
highest values. These relationships and communities are fostered and sustained by
varied sorts of ties that we share with others, such as languages, activities, practices,
projects, traditions, histories, goals, views, values, and mutual attractions—not to
mention common enemies and shared injustices and disasters.
Someone who becomes more autonomous concerning some tradition, author-
ity, view, or value in her life does not stop depending on other persons or relation-
ships, nor does she evade her own necessarily social history of personal develop-
ment. Her initial detached questioning does not arise in a social vacuum but is likely
to be prompted by commitments reflecting still other relationships that for the pres-
ent time remain unquestioned and perhaps heteronomous. A shift in social relation-
ships or commitments is not equivalent to, nor need it betoken, wholesale social
detachment.
Autonomy does not require self-creation or the creation of law ex nihilo, a lim-
itation that we need not join Richard Rorty in lamenting.36 Becoming more proce-
durally autonomous concerning particular standards, norms, or dictates involves re-
flecting on them in a language that one did not create—according to further norms
and standards that one has almost surely taken over from others—in light of what is
most central to that product of social development that is oneself.37 Also, autonomy
is always a matter of degree, of more or less. Reflective consideration still counts as
a gain in autonomy even if done in the light of other standards and relationships
not simultaneously subjected to the same scrutiny.

How Autonomy Disrupts Personal Relationships

Feminists have sought a relational account of autonomy to render it relevant to
women. Philosophers in general have sought such an account to make good on a
widely shared intuition that autonomy is not antithetical to other social values and
virtues that concern us all, such as love, friendship, care, loyalty, and devotion.38
Many philosophers seem to expect that most of what we want or value in interper-
sonal relationships will prove to be consistent with the ideal of autonomy, once we
develop an appropriately social conception of it.
That conviction, however, may be unfounded. It underestimates, I believe, the
disruption that autonomy can promote in close personal relationships and in com-
munities. Although autonomy is not inherently antithetical to social relationships,
42 Autonomy and the Social

nevertheless, in practice autonomy often contingently disrupts particular social bonds.

How it sometimes interferes with social relationships and what this implies about its
value for women make up the fourth and primary theme of this article.
Human relationships and communities, as noted, are held together by a variety
of ties that persons share, including languages, practices, traditions, histories, goals,
views, and values. Any of these elements in someone's life can become the focus of
her critical scrutiny. Whenever someone questions or evaluates any tie or commit-
ment that binds her to others, the possibility arises that she may find that bond un-
warranted and begin to reject it. Rejecting values that tie someone to others may
lead her to try to change the relationships in question or simply to detach herself
from them. Someone might also reflect on the very nature of her relationships to
particular others and come to believe that those ties are neglecting or smothering
important dimensions of herself. To liberate those aspects of herself, she might have
to distance herself from the problematic relationships.
Most personal and communal relationships are multifaceted and based on more
than one sort of tie. Kinship, for example, keeps many people in contact with rela-
tives whose values would otherwise repel them. Childhood friends who travel dis-
parate paths in life may retain shared memories that keep them ever fondly in touch
with each other. A shared ethnic identity may link economically diverse people in
the pursuit of collective political ends or cultural self-affirmation. Thus, friends, rel-
atives, or other associates who diverge over important values may still remain related
to each other in virtue of other shared ties.
The resilience of social relationships is, of course, not always a blessing. Rela-
tionships in which one partner exploits or abuses the other can also, and regrettably,
last for years.39 Sometimes, however, a person becomes so disenchanted with her re-
lationships or their underlying values that they become, to her, unbearable. At that
level of discontent, and assuming viable alternatives, she may begin to withdraw
from the relationships. In so doing, she displays just the sort of relational discon-
nection that can stem from a person's autonomous reflections on (and growing dis-
satisfaction with) her prior commitments.
Alternatively, someone's increasing autonomy might result in the breakup of a
relationship not because she rejects it but rather because other parties to the rela-
tionship reject her. They might despise the changes in her behavior that they are wit-
nessing. Some parents, for example, disown children who rebel too strongly against
deeply held parental values. Peer groups often ostracize their members for disregard-
ing important norms that prevail in their own subcultures.
Strictly speaking, to say that autonomy unqualified (sometimes) disrupts social
relationships is misleading. The mere capacity for autonomy is not intrinsically so-
cially disruptive. What disrupts a social relationship in a particular case is the actual
exercise of the capacity. More strictly still, the differences that arise between people
as a result of one party's autonomous rejection of values or commitments that the
other party still holds may lead one party to draw away from or reject the other.
Thus, to borrow a rhetorical turn of phrase from the U.S. lobbying group the Na-
tional Rifle Association (!), it is not autonomy (as a dispositional capacity) that dis-
rupts social relationships; it is people who disrupt social relationships.
The exercise of autonomy, it should be emphasized, is not sufficient as such to dis-
 Automomy, Social Disruption, and Women 43

rupt particular relationships. The connection between autonomy and social disrup-
tion is merely contingent. Someone's autonomous reflections increase the chances of
disruption in her social relationships but do not make it a necessary consequence.
In certain sorts of circumstances, autonomy may not even make social disruption
more likely than it already is. Someone's reflective consideration might lead her to
appreciate in a new light the worth of her relationships or the people to whom she
is socially attached and to enrich her commitment to them. In such cases, autonomy
would strengthen rather than weaken relational ties. Even if someone began to dis-
agree with significant others about important matters, their relationship might still
not suffer. People use many interpersonal strategies to keep differing commitments
from disrupting social harmony—"never discuss religion or politics," for example.40
Thus, someone's autonomy is not a sufficient condition for the disruption of
her social relationships. Nor is it a necessary condition. A person might end a rela-
tionship because of new commitments that she has reached heteronomously. Peer
pressure, for example, can promote knee-jerk rebelliousness that disrupts personal
relationships as much as the greatest soul searching and critical self-reflection. Some-
one's attitudes can also change as the result of traumatic experiences over which she
had no control. These changes may occasion deep rifts in her relationships with
close others.41 Someone's increasing autonomy is thus neither a sufficient nor a nec-
essary cause of disruption in social relationships.
Nevertheless, the contingently possible connection between autonomy and so-
cial disruption is of noteworthy importance. When a culture places great value on
autonomy, members of that culture are thereby encouraged to question their prior
allegiances and the standards that impinge on them. Autonomy as a cultural ideal
creates a supportive climate for personal scrutiny of traditions, standards, and au-
thoritative commands.42 Public discourse in such a culture will tend to promote
open dialogue and debate over values and traditions. Autonomy-idealizing societies
may protect such discourse, and the normative critiques it can foster, with a legal
right to a substantial degree of freedom of expression.
Thus, other things being equal, in a culture that prizes autonomy, all traditions,
authorities, norms, views, and values become more vulnerable to rejection by at least
some members of the society than they would be in a society that devalued auton-
omy. No commitment in such a culture remains entirely immune to critical scrutiny,
whether the commitment concerns religion, sex, family, government, economy, art,
education, race, ethnicity, gender, or anything else.
Once such scrutiny takes place, the likelihood increases that people who are so-
cially linked to each other will begin to diverge over views or values they previously
shared, including the value of their social ties. Once people begin to diverge over
important matters, they are more likely than they were before to disagree and quar-
rel with each other or to lose mutual interest and drift apart. In this way (other
things being equal), an autonomy-idealizing culture increases the risk of (though it
certainly does not guarantee) ruptures in social relationships.
To be sure, cultures that idealize autonomy do not always extend this ideal to all
social groups. Sometimes certain sorts of people, white men, for example, receive
the lion's share of the social protections and rewards for being autonomous. Also,
even an autonomy-idealizing culture may shield certain norms or values from criti-
44 Autonomy and the Social

cal scrutiny. In such a society, values that protect dominant social groups, those priv-
ileged to enjoy the value of autonomy, might not get as much critical attention as
they deserve. Whereas limitations on rampant autonomy might be necessary to pre-
vent wholesale social breakdown, they can also create bastions of unquestioned au-
tonomous privilege. In such a culture, autonomy might well be a restricted, domes-
ticated, socially nonthreatening luxury.
Nevertheless, as long as autonomy is culturally valued even for only some
groups and for only certain issues, its very cultural availability opens up the possi-
bility for wide social transformation. Even if idealized for only a privileged few, it
can always fall into the "wrong" hands. New groups might coincidentally acquire
autonomy competency in virtue of social changes, such as the spread of literacy and
formal education. They might then go on to contest norms and values previously
left unscrutinized. This possibility has been historically crucial for women and other
subordinated groups. The ideal of autonomy is thus always a potential catalyst for
social disruption in interpersonal relationships.
Notice also that the rupture that autonomy can promote in any one particular
social relationship does not necessarily amount to an overall decline in the socie-
tal quantity of relationality, to put the point inelegantly. Typically, when someone
questions some prior commitment, such as a religious commitment, which ce-
mented certain relationships in her life, she is probably doing so in company with
other skeptics whose reflections prompt and reinforce her own rising doubts.
When she turns away from her prior religious community, she is likely to be turn-
ing toward a different community, perhaps a religiously neutral secular commu-
nity or a new religious group. Those with whom someone shares her new com-
mitments may have given her a vocabulary or perspective for reflecting on her
central concerns. Without any empirical backing for this claim, I nevertheless es-
timate that in most cases in which autonomous reflection does lead people to re-
ject the commitments that bound them to particular others, they are at the same
time taking up new commitments that link them through newly shared convic-
tion to different particular others. This is one important reason for thinking of au-
tonomy as social in character.
Although people in an autonomy-valuing society might have as many inter-
personal relationships as those in a society that devalued autonomy, it is reason-
able to speculate that the nature of people's relationships would differ in the two
cases. Where people are permitted with relative ease to leave relationships that
have become dissatisfying to them, we should expect attachments to be less stable,
to shift and change with greater frequency, than in societies in which personal au-
tonomy (or relational mobility) is discouraged. The types and qualities of rela-
tionships in an autonomy-promoting culture would also probably differ from
those of an autonomy-discouraging culture. Relationships into which people are
bom and in which they are first socialized—those of family, church, neighbor-
hood, friendship, and local communities—would probably be disrupted first, if
any, by widespread individually autonomous reflections on basic values and com-
mitments. In a culture that values autonomy, it is likely that more people than
otherwise would gravitate toward voluntary relationships formed in adult life
around shared values and attitudes. 43
 Automomy, Social Disruption, and Women 45

Women and the Social Disruptions of Autonomy

What difference should it make to our theory of autonomy that autonomy, however
social its nature or origin, might promote the disruption of social relationships?
More precisely in the present context, what difference does this possibility make to
women*. If autonomy is sometimes socially disruptive, does that make it inimical to
the relational orientation that many feminists celebrate in women and display in
their own moral concerns?
Some people exhibit what I call autonomophobia, or fear of autonomy. What
they usually fear is not their own autonomy; it is the autonomy of others that scares
them. Their concern is that autonomous people will disrupt or desert valuable,
shared relationships. Which relationships are thought to be valuable will be specified
differently from distinct critical perspectives. Feminists, and many women in gen-
eral, often worry about the relationships on which many women depend for mate-
rial and emotional sustenance.
Whether or not any particular woman benefits or suffers in virtue of the exer-
cise of autonomy depends on how she is positioned in relation to it. When a woman
is connected to someone else whose autonomous pursuits disrupt their relationship,
the immediate effect on her is likely to be simply a loss—of whatever benefits she
derived from their relationship. Autonomophobia is thus a legitimate concern. It
arises from the ways in which our lives are intertwined with those of other persons.
When others who are close to us reflect on their own deepest commitments, they
might well find grounds for challenging or abandoning the relationships and com-
munities that we share with them. We might find ourselves helpless as a result. Social
relationships and communities are collective projects. They function best when sus-
tained jointly by people with important values or norms in common. In a culture
that idealizes autonomy, each individual faces the insecurity of investing herself in
relationships and communities that the other participants might, on critical reflec-
tion, come to reject.
Historically, the disruption of personal relationships has had a different impact
on men than on women. Because women were usually limited to dependence on
men for financial support, whereas men had no comparable limitation, women
doubtless suffered more than they benefited from the cultural idealization of auton-
omy. Men have been historically better situated than women to forsake personal re-
lationships that came to seem dissatisfying to them. Unlike most women, many men
have had the material and cultural resources with which to support themselves, as
well as greater opportunities to seek more satisfying relationships elsewhere. Men
were able to abandon their responsibilities to women and children to pursue forms
of personal fulfillment unavailable to women.
Men who, like Gauguin, produced good enough subsequent works have been
celebrated for autonomous pursuits that involved neglecting or abandoning rela-
tionships that supported women and children. Dependent women and children
have suffered greatly from these male desertions. Women's own autonomous living,
by contrast, brought them much more censure and hardship than praise.44 Since
women tend to be more financially dependent on men than men are on women
46 Autonomy and the Social

even today, autonomophobia is understandably still more often a female than a male
concern. Thus, men's autonomy would have done women little direct good and
could have imposed serious harm.
On the other hand, many social relationships constrain and oppress women,
indeed the very women who work to sustain them. Apart from whether or not
women want to devote their lives to maintaining close personal ties, gender norms
have required it of them. Women have been expected to make the preservation of
certain interpersonal relationships such as those of family their highest concern re-
gardless of the costs to themselves. Women who have had important commitments
other than those of taking care of family members were nevertheless supposed to
subordinate such commitments to the task of caring for loved ones. Many men, by
contrast, have been free to choose or affirm their highest commitments from
among a wide panorama of alternatives. Indeed, men are sometimes lauded for just
the sort of single-minded pursuit of an ideal that imposes sacrifices on all the people
close to them.
Traditionally, the majority of women derived their primary adult identities
from their marriages and families. For at least some groups of women, however, so-
cial and economic opportunities have broadened in the late twentieth century. Be-
cause of expanding financial opportunities in the West, many women no longer
need to accommodate themselves uncritically to traditional marriages or other rela-
tional ties to sustain themselves. As many feminists have well recognized, there is no
reason to defend social relationships without qualification. There is nothing intrin-
sic to each and every social relationship that merits female or feminist allegiance.45
The traditional relational work of women has included sublime joy and fulfillment
but also abuse, exploitation, and subordination. There are some, perhaps many, re-
lationships that women, too, should want to end.
Thus, the disruption of social relationships that can follow someone's grow-
ing autonomy is not itself inherently alien to women, nor is it a dimension of the
ideal of autonomy that women today should automatically reject. What should
matter to any particular woman in any given case is the worth of the relationship
in question and how its disruption would bear on herself and on innocent others.
The old question, "can this marriage be saved?" should be revised to, "can this
marriage be saved from oppressiveness?" Some relationships should be preserved
and others should be abolished. Even relationships that should be preserved can
always be improved. Sometimes what disrupts social relationships is good for par-
ticular women. Since the socially disruptive potential of autonomy can at least
sometimes be good for women, it does not constitute a reason for women to re-
pudiate the ideal of autonomy.
indeed, reflecting on one's relationships or the norms or values that underlie
them might be the only way someone can determine for herself the moral quality of
those relationships. A woman who does not reflect on her relationships, communi-
ties, norms, or values is incapable of recognizing for herself where they go wrong or
of aiming on her own to improve them. Her well-being depends on those who con-
trol her life and on their wisdom and benevolence—regrettably, not the most reli-
able of human traits. Autonomy is thus crucial for women in patriarchal conditions,
in part because of its potential to disrupt social bonds. That autonomy is sometimes
 Automomy, Social Disruption, and Women 47

antithetical to social relationships is oftimes a good for women. With all due respect
to Audrey Lorde, the "master's tools" can "dismantle the master's house."46
Thus, although women still have occasion to fear men's autonomy, it seems that
many women have good reason to welcome our own. When a woman is the one
who is exercising autonomy, even if its exercise disrupts relationships in her life the
value of her gain in autonomous living might well make the costs to her worth her
while. She may plausibly fear what increasingly autonomous others might do to the
relationships between herself and them, but it would not make sense for her to re-
ject autonomy for herself. A woman might choose not to exercise autonomy under
certain conditions. She might, for example, devote herself loyally to an ideal that
she can only serve by working with a group of persons who sometimes take specific
actions she does not understand or endorse. She can hardly want to give up, how-
ever, the very option of so devoting herself. To reflect on the standards or values ac-
cording to which one will behave or live one's life, as one does when resolving to
dedicate oneself to a particular ideal, is already to exercise a degree of autonomy. It
would be self-defeating, at the same time, to reject autonomy altogether as a value
for oneself.
Once women admit that autonomy might be a value for us, it would be difficult
to deny its value for persons in general. The capacity for autonomy seems instru-
mentally valuable as a means for resisting oppression and intrinsically valuable as
part of the fullest humanly possible development of moral personality. In these re-
spects it seems valuable for anyone. The problem arises with the need for reciprocity.
We cannot esteem autonomy in women while deprecating it in men. Yet men's au-
tonomy and the social disruption it can promote does sometimes threaten women's
well-being. I have argued that when women have access to means for their own ma-
terial support, this risk is lessened.47 Women can then gain at least as much from a
generalized cultural idealization of autonomy as they risk by it.
There are, as well, certain mitigating possibilities that reduce, even though they
do not eliminate, the likelihood that autonomy might cause social disruption. Au-
tonomy does not necessarily lead someone to reject her prior commitments. Some-
one else's increasing autonomy might instead enhance her appreciation of her close
relationships. Even if she comes to regard a relationship as seriously flawed, she might
work to improve it rather than abandoning it.
These possibilities suggest that alongside autonomy as a cultural ideal, we should
also idealize the values and responsibilities that make relationships and communities
worthwhile.48 We can emphasize, for example, the ways in which close relationships
are vital sources of care for the most vulnerable members of our society.49 We should
articulate these values in public dialogues in which all can participate, including those
who might become autonomously skeptical about those social ties.
This balanced pursuit of the values of community along with the ideal of au-
tonomy is a partial response to the concern that the empirical social disruptiveness
of autonomy lessens the value of autonomy for women. There is no way, however,
to alleviate this concern fully. The possibility of social disruptiveness is one risk that
must be faced by persons and cultures who would idealize personal autonomy.50 I
have argued that social disruptiveness is, at least, a mixed curse, one that harbors the
potential for good, as well as bad, consequences.
48 Autonomy and the Social

In addition, the mantra of "family values" that is invoked uncritically in so much

public debate in the United States should remind us as feminists of the hazards of
allowing any relationships, including those we most cherish, to be entirely insulated
from the critical reflection of all their participants. Even care for the most vulnerable
can usually be improved. It is a form of respect toward those with whom we most
want affiliation to want them to find forms of commitment to us that reflect their
most cherished values.

Notes

1. I am grateful to Natalie Stoljar and Catriotia Mackenzie for helpful editorial sugges-
tions on an earlier draft of this article.
2. Yann Ic Pichon, Gauguin: Life, Art, Inspiration (New York: Abrams, 1987), p. 26.
3. Bernard Williams's discussion of moral luck deploys the hypothetical biography of
an artist whose life resembles that of the historic Gauguin; see Williams, "Moral Luck," in
Moral Luck (Cambridge: Cambridge University Press, 1981), p. 37. See my discussion of
Williams on the Gauguin-like example in my What Are Friends For?: Feminist Perspectives on
Personal Relationships and Moral Theory (Ithaca, N.Y.: Cornell University Press, 1993), pp.
163-170.
4. See the germinal work on this topic by Carol Gilligan, In a Different Voice (Cam-
bridge, Mass.: Harvard University Press, 1982), and Nel Noddings, Caring: A Feminine Ap-
proach to Ethics and Moral Education (Berkeley: University of California Press, 1984).
5. See Susan Faludi's discussion of the popularization of research results alleging that
women's chances of marrying fall precipitously after age forty, in Backlash: The Undeclared
War against American Women (New York: Crown Publishers, 1991), chap. 1. Faludi argues
persuasively that the conclusions were misrepresented in the mass media. My point is a dif-
ferent one: these research results would not have received popular attention if it hadn't been
for the presumption that people, including women, would want to know about them.
6. Sec Gilligan, In a Different Voice, chap. 1.
7. See, for example, Helen I.ongino, Science as Social Knowledge (Princeton, N.J.:
Princeton University Press, 1990).
8. For an account of emotion as a source of autonomy, see Bennett W. Helm, "Free-
dom of the Heart," Pacific Philosophical Quarterly 77 (1996): 71-87. Harry Frankfurt dis-
cusses caring and love as sources of autonomy but uses those terms to refer to states of will
rather than of emotion; see, for example, The Importance of What We Care About (Cam-
bridge: Cambridge University Press, 1988), and "Autonomy, Necessity, and Love," in Ver-
nunftbegriffe in der Moderne: Stuttgarter Hegel-Kongress 1993, ed. Hans Friedrich Fulda and
Rolf-Peter Horstmann (Stuttgart: Klctt-Cotra, 1994), pp. 433-447.
9. Of course, socialization might itself be coercive. See John Christman's approach to
this problem in "Autonomy and Personal History," Canadian Journal of Philosophy 20
(1991): 1-24.
10. Diana T. Meyers, Self, Society, and Personal Choice (New York: Columbia University
Press, 1989), pp. 13-17.
11. Morwenna Griffiths explores the importance of narratives in the cultural under-
standing of autonomy in Feminisms and the Self: The Web of Identity (London: Routledge,
1995).
12. Susan Brison, "Surviving Sexual Violence? Journal of Social Philosophy 24, no. 1
(Spring 1993): 5-22.
 Automomy, Social Disruption, and Women 49

13. Patricia Hill Collins, Black Feminist Thought: Knowledge, Consciousness, and the Pol-
itics of Empowerment (New York: Routledge, 1991), especially chaps. 4 and 5.
14. Minnie Bruce Pratt, "Identity: Skin Blood Heart," in Yours in Struggle: Three Fem-
inist Perspectives on Anti-Semitism and Racism, ed. Elly Bulkin, Minnie Bruce Pratt, and Bar-
bara Smith (Brooklyn, N.Y.: Long Haul Press, 1984), pp. 9-63.
15. Sara Ruddick, Maternal Thinking: Toward a Politics of Peace (New York: Ballantine,
1989).
16. The chancellor at a university near my own, for example, was recently fired by his
university's governing board. The faculty members who supported him thought the problem
was, as one of them put it, that the chancellor was "too autonomous, too independent." Fac-
ulty supporters described the governing board as wanting a "team player" instead. See Susan
C. Thomson and Kim Bell, "Mizzou Chancellor Wants Buyout," St. Louis Post-Dispatch, June
14, 1996, C7. Note that the figure of a "team player" is a historically masculine metaphor for
a cooperative social agent. The differences between women's and men's paradigm images of
social cooperation deserve some study.
17. Michael J. Sandel, Liberalism and the Limits of justice (Cambridge: Cambridge Uni-
versity Press, 1982), and Alasdair Maclntyre, After Virtue, 2nd ed. (Notre Dame, Ind.: Uni-
versity of Notre Dame Press, 1984).
18. Loren E. Lomasky, Persons, Rights, and the Moral Community (New York: Oxford
University Press, 1987), p. 249. In the now common procedural account of autonomy, a
view that I share (see the section on social reconcepualizations), no particular choices are in-
trinsic to autonomy. An autonomous person might embrace traditional relationships, reject
traditional relationships, welcome the Red Guards, or abhor the Red Guards. What matters
is how she arrived at her political views and whether those views reflect her own considered
convictions. Lomasky construes autonomy as a failing only of those who make political
choices he rejects, but this is just as mistaken as assuming that autonomy is a virtue only of
those who make what one considers to be the right political choices.
19. Genevieve Lloyd, The Man of Reason: "Male" and "Female" in Western Philosophy
(Minneapolis: University of Minnesota Press, 1984). See also Lorraine Code, Rhetorical
Spaces: Essays on Gendered Locations (New York: Routledge, 1995), especially chap. 10, "Cri-
tiques of Pure Reason."
20. Feminist sources include Code, Rhetorical Spaces, and Alison M. Jaggar, "Love and
Knowledge: Emotion in Feminist Epistemology," in Gender/Body/Knowledge, ed. Alison M.
Jaggar and Susan R. Bordo (New Brunswick, N.J.: Rutgers University Press, 1989), pp.
145-171. Nonfeminist sources include Allan Gibbard, Wise Choices, Apt Feelings: A Theory
of Normative Judgment (Cambridge, Mass.: Harvard University Press, 1990).
21. A different approach would be to argue that the stereotypic association of women
with emotion was always groundless and that women are as able as men to exercise a narrowly
cognitive mode of reason. See Louise M. Antony and Charlotte Witt, eds., A Mind of One's
Own: Feminist Essays on Reason and Objectivity (Boulder, Colo.: Westview Press, 1993), espe-
cially the essays by Margaret Atherton and Louise Antony. See the discussion of these essays
by Code, Rhetorical Spaces, pp. 217-223.
22. See Susan Golombok and Robyn Fivush, Gender Development (Cambridge: Cam-
bridge University Press, 1994), pp. 7-8.
23. Ibid., p. 18. A social or relational account of autonomy, such as that presented
here, is one that construes social relationships as necessary for autonomy but not sufficient
for it. There is nothing about social interconnection as such that entails, causes, or suggests
autonomy.
24. Meyers, Self, Society and Personal Choice, especially part 3.
25. This point is, of course, not universal throughout Western cultures. Men of op-
50 Autonomy and the Social

pressed groups, such as racial minorities, may not have had significantly greater opportunities
than the women of their own groups to act and live autonomously.
26. Many feminists have charged the traditional philosophical ideal of autonomy with
excessive individualism; see, for example, Lorraine Code, "Second Persons," in What Can She
Know?: Feminist Theory and the Construction of Knowledge (Ithaca, N.Y.: Cornell University
Press, 1991), pp. 76-79.
27. See, for example, Evelyn Fox Keller, Reflections on Gender and Science (New Haven,
Conn.: Yale University Press, 1985); Nancy Chodorow, The Reproduction of Mothering: Psy-
choanalysis and the Sociology of Gender (Berkeley: University of California Press, I 978); and
Jessica Benjamin, The Bonds of Love: Psychoanalysis, Feminism, and the Problem of Domina-
tion (New York: Pantheon, 1988).
28. Annette Baier, "Cartesian Persons," m Postures of the Mind: Essays on Mind and
Morals (Minneapolis: University of Minnesota Press, 1985). See also the discussion of this
notion in Code, "Second Persons."
29. One prominent philosopher who neglects socialization, and, indeed, social rela-
tionships generally, in his account of autonomy is Frankfurt, Importance of What We Care
About,
30. See the discussion of both of these points by John Christman, "Feminism and Au-
tonomy," in "Nagging" Questions: Feminist Ethics in Everyday Life, ed. Dana E. Bushnell
(Lanham, Md.: Rowman & Littlefield, 1995), pp. 17-39.
31. Gerald Dworkin provides one example of a procedural account of autonomy; see
his The Theory and Practice of Autonomy (Cambridge: Cambridge University Press, 1988),
pp. 18,21-33.
32. Sec Marina Oshana, "Personal Autonomy and Society," Journal of Social Philosophy
29, no. 1 (Spring 1998): 81-102.
33. Feminist theorists who have developed this view include Keller, Reflections on Gender
and Science, Jennifer Nedelsky, "Reconceiving Autonomy: Sources, Thoughts, and Possibili-
ties," Yale Journal of Law and Feminism 1, no. 1 (Spring 1989): 7-36; Meyers, Self, Society,
and Personal Choice; and Code, "Second Persons." Mainstream theorists who have developed
this view include Joseph Raz, The Morality of Freedom (Oxford: Clarendon Press, 1986);
Dworkin, Theory and Practice of Autonomy, and Joel Feinberg, "Autonomy," in The Inner
Citadel, ed. John Christman (New York: Oxford University Press, 1989), pp. 27-53. For a
discussion of the convergence of these two groups around a social conception of autonomy,
see Christman, "Feminism and Autonomy," and my "Autonomy and Social Relationships: Re-
thinking the Feminist Critique," in Feminists Rethink the Self, cd. Diana T. Meyers (Boulder,
Colo.: Westvicw Press, 1997), pp. 40-61. Some mainstream philosophers deny that the tra-
ditional notion of autonomy, even in its rigorous Kantian formulation, ever really excluded or
ignored the importance of interpersonal relationships; see J. B. Schneewind, "The Use of Au-
tonomy in Ethical Theory," in Reconstructing Individualism: Autonomy, Individtiality, and the
Self in Western Theory, ed. Thomas C. Heller, Morton Sosna, and David E. Wellberry (Stan-
ford, Cal.: Stanford University Press, 1986), and Thomas E. Hill, Jr., "The Importance of Au-
tonomy," in Autonomy and Self-Respect (Cambridge: Cambridge University Press, 1991).
34. See, for example, Jurgen Elabermas, Moral Consciousness and Communicative Action,
trans. Christian Eenhardt and Shierry Weber Nicholsen (Cambridge, Mass.: MIF Press,
1990), and Joel Anderson, "A Social Conception of Personal Autonomy: Volitional Identity,
Strong Evaluation, and Intersubjective Accountability," Ph.D. dissertation, Northwestern
University, Evanston, 111., 1996.
35. See, for example, Sandel, Liberalism and the Limits of Justice, p. 19.
36. Richard Rorty, Contingency, Irony, and Solidarity (Cambridge: Cambridge Univer-
sity Press, 1989).
 Automomy, Social Disruption, and Women 51

37. Gerald Dworkin notes the impossibility of creating our own moral principles. Such
a requirement "denies our history. . . . We . . . are deeply influenced by parents, siblings,
peers, culture, class, climate, schools, accident, genes, and the accumulated history of the
species. It makes no more sense to suppose we invent the moral law for ourselves than to sup-
pose that we invent the language we speak for ourselves" (The Theory and Practice of Auton-
omy, p. 36).
38. Ibid., p. 21.
39. If abusive relationships persist for long periods of time, it is usually because the
abused partner has, or thinks she has, no other viable options or because she sacrifices her
own well-being to that of her abuser. For a survey of why long-time battered women finally
seek court orders of protection against abusive male partners, see Karla Fischer and Mary
Rose, "When 'Enough is Enough': Battered Women's Decision Making around Court Orders
of Protection," Crime andDelinquency41, no. 4 (October 1995): 414-429.
40. I do not endorse this maxim; I merely cite it as an example of the strategies that
people use to keep disagreements from disrupting social relationships.
41. See Brison's discussion in "Surviving Sexual Violence," of the difficulties that arose
in her relationships with family, friends, and others after she was violently raped.
42. As Dworkin notes, "Those who practice in their daily life a critical reflection on
their own value structure will tend to be suspicious of modes of thought that rely on the un-
critical acceptance of authority, tradition, and custom" (The Theory and Practice of Autonomy,
p. 29),
43. For further discussion of this theme, see my What Are Friends For:1 chap. 9.
44. See Alison MacKinnon, Love and Freedom: Professional Women and the Reshaping of
Personal Life (Cambridge: Cambridge University Press, 1997), on the hurdles faced by
women in Australia at the end of the nineteenth and beginning of the twentieth century who
sought higher education and careers outside the home.
45. There are many feminist discussions of problems that women face in social rela-
tionships; see, for example, Susan Moller Okin, Justice, Gender, and the Family (New York:
Basic Books, 1989).
46. See Audrey Lorde, "The Master's Tools Will Never Dismantle the Master's House,"
Sister Outsider (Freedom, Cal.: Crossing Press, 1984), pp. 110-113.
47. On this topic, see the essays in Martha Nussbaum and Jonathan Glover, eds.,
Women, Culture and Development: A Study of Human Capabilities (Oxford: Clarendon Press,
1995).
48. See the essays in Feminism and Community, eds. Penny Weiss and Marilyn Fried-
man (Philadelphia: Temple University Press, 1995).
49. See, for example, Robert E. Goodin, Protecting the Vulnerable: A Reanalysis of Our
Social Responsibilities (Chicago: University of Chicago Press, 1985); Neera Kapur Badhwar,
ed., Friendship: A Philosophical Reader (Ithaca, N.Y.: Cornell University Press, 1993); Joan C.
Tronto, Moral Boundaries: A Political Argument for an Ethic of Care (New York: Routledge,
1994).
50. Should we devalue autonomy for individuals, perhaps recasting it as an ideal for
groups only? The notion of group autonomy is extremely important, especially for oppressed
groups; see, for example, Laurence Mordekhai Thomas, Vessels of Evil: American Slavery and
the Holocaust (Philadelphia: Temple University Press, 1993), pp. 182-189. Group autonomy,
however, does not necessarily help individuals when they face oppressive conditions in isola-
tion. It complements but does not replace individual autonomy. In addition, group auton-
omy promotes its own risk of social disruption in the relationships between groups. The pos-
sible advantages, as well as the possible costs, of autonomy's socially disruptive potential
simply reappear at a more encompassing level of social integration.
 2

AUTONOMY AND THE SOCIAL SELF1

Linda Barclay

Widespread dissatisfaction with the ideal of individual autonomy is a prominent

feature of feminist and communitarian perspectives on moral and political theory.
Whether attention is focused on moral impartialism, Kantian ethics, modern po-
litical theory, or in particular, contemporary liberalism, a frequently voiced concern
is that individual autonomy is overemphasized and that an "atomistic" or "abstract
individualistic" conception of the self is presupposed. Both feminists and commu-
nitarians have claimed that numerous moral and political theories promote a vision
of the autonomous self as essentially independent and self-sufficient, a vision that
denies the inescapable connectedness of selves and the fact that their immersion in
networks of relationships forms their desires, aspirations, indeed their very identi-
ties. In other words, what is denied is that the self is essentially social.
Lorraine Code, for example, claims that the prevailing notion of autonomy
posits human beings as "self-sufficient, independent, and self-reliant," a partial and
ultimately distorted view that blithely ignores the fact that we are also interdepend-
ent and cooperating.2 For similar reasons Jennifer Nedelsky argues that "feminism
requires a new conception of autonomy," one that is able to "recognize the inher-
ently social nature of human beings."3 Sarah Lucia Hoagland, somewhat more pes-
simistically, claims that "autonomy . . . is a thoroughly noxious concept" that "en-
courages us to believe that connecting and engaging with others limits us ... and
undermines our sense of self."4

52
 Autonomy and the Social Self 53

Similar objections can be traced through communitarian critiques of liberalism.

Alasdair Maclntyre repudiates the liberal assumption that "individuals . . . are pri-
mary and society secondary" and that individual interests are assumed to be "prior
to, and independent of, the construction of any social bonds between them." Such
a picture ignores the fact that "individuals inherit a particular space within an inter-
locking set of social relationships."5 Similarly, Michael Sandel argues that liberals as-
sume we each choose our ends, an assumption belied by the fact that the self does
not so much choose its ends as "discover" them as constituents of its identity, ends
that are derived from the communities of which it is a part.6
In short, both communitarians and feminists have championed social concep-
tions of the self, which they argue are denied, or at least underplayed, by the promi-
nence afforded to individual autonomy in a range of theories. Their apparently
shared concerns seem in fact to be "strikingly similar,"7 leading a number of femi-
nists to question whether there is scope for an explicit alliance between feminism
and communitarianism.8
In fact, a number of quite distinct issues is often concealed by the very general
suggestion that the concept of autonomy presupposes an asocial conception of the
self. In what follows I set out three distinct senses in which feminists and commu-
nitarians have defined the social self and assess the accusation that the notion of au-
tonomy is incompatible with each. The three senses in which the self is said to be so-
cial are deterministic, motivational, and constitutive.9
I develop these three distinct accounts of the social self in the next three sec-
tions. I attempt to demonstrate that a suitably nuanced conception of autonomy is
compatible with plausible versions of the socially determined and motivationally so-
cial self, but it is incompatible with particularly strong versions of the socially deter-
mined and the constitutively social self. However, I also argue that these strong ver-
sions of the social self are not in fact plausible and thus should not threaten our
commitment to a suitable conception of autonomy. In the final section I consider
whether there are grounds for a communitarian-feminist alliance against the ideal of
autonomy. I argue that despite very superficial similarities, feminist and communi-
tarian claims about autonomy and the social self are quite distinct and, in the main,
incompatible. It is therefore doubtful that feminists can wholeheartedly embrace
communitarianism as an ally.

The Socially Determined Self

Among the numerous philosophers writing on autonomy, broad consensus on
some kind of procedural notion has been emerging.10 Generally, autonomy is said
to consist of a capacity, or the exercise of certain competencies,11 that enables one
to reflect on one's aims, aspirations, and motivations and choose one's ends and pur-
poses through such a reflective process. One of the most frequently voiced criticisms
of autonomy is that it is in tension with the fact that the self's very identity is deter-
mined by the communities of which one is a part, that is, by the pervasive influence
of parents, peers, and culture. A procedural notion of autonomy—which envisions
a person critically reflecting on her desires and aspirations—could be thought to
presuppose that the self can somehow simply transcend the influence of all of these
54 Autonomy and the Social

factors and make oneself anew, to become a fully "self-made (wo) man." But this de-
nies the obvious reality that none of us is self-made in this fashion for we are all, in-
escapably, "a product of our environment." This is the claim that the self is socially
determined. The claim is broad and can incorporate four quite distinct concerns.

The Global Influence of Social Determinism

The first and strongest way of expressing concern about the relation between au-
tonomy and the fact that the identity of the self is socially determined is to deny
that individual autonomy is possible. For every value, plan, or project endorsed by
the self, there is a set of social influences that can account for why one makes the
"choices" one does. Autonomy is an illusion. Even if we go through the motions of
reflecting on our motivational structure to critically reconsider our ends and pur-
poses, the "choices" we make are themselves just as much a product of the social en-
vironment as our prereflective ends and purposes are.
This claim about the impossibility of autonomy bears a strong resemblance to
the familiar free will versus determinism debate. From this perspective, the prob-
lem is to explain how autonomy is possible, given the truism that for any action
we perform or any intention that we form there is a set of antecedent causes that
explains why we do what we do or why we think what we think. In other words,
there are social (and biological) facts that explain why we do what we do or why
we think as we do; the sources of the self's ends and purposes are sources external
to oneself. How is autonomy possible, given the indisputable facts of determin-
ism? It seems ineluctably in conflict with the fact that the self is determined, all
the way down.
I would argue that the claim that social determinism undercuts the possibility
of autonomy—whether that claim is expressed in terms of the metaphysical threat
of determinism, or of the pervasiveness of socialization—is based on a misunder-
standing of the notion of autonomy. Determinism is only a global threat to the pos-
sibility of autonomy on the assumption that agency is only genuinely autonomous if
it is uncaused, or determined by no reasons whatsoever. However, that one is au-
tonomous does not mean that one's choices are uninfluenced or uncaused, for it is
doubtful that such a notion is even coherent. Autonomous agency does not imply
that one mysteriously escapes altogether from social influence but: rather that one is
able to fashion a certain response to it.
Such responses to the problem of social determinism are now widely known.
For example, Harry Frankfurt argues that persons are unique in having the ability
to critically reflect on and choose which of their desires they identify with and which
they reject. 12 Diana Meyers's detailed descriptions of autonomous agents abound
with examples of people whose preferences, ends, and projects can be thought to be
authentically their own.13 Authenticity does not refer to the hidden inner self, which
is finally revealed after the effects of socialization have been stripped away. Rather,
autonomous people exercise a complex array of what she calls autonomy competen-
cies, which include diverse self-reading and self-actualizing skills. The authentic self
continually emerges throughout this ongoing process of exercising autonomy com-
 Autonomy and the Social Self 55

petency. It is not hidden and then found: it is constructed throughout the process of
exercising one's autonomy. Numerous other theorists of autonomy have developed
accounts of how autonomous people can negotiate the effects of socialization.
Robert Young, like Meyers, emphasizes the importance of self-knowledge and gain-
ing understanding of how the influence of one's culture, family, and so on has
shaped one's aims, aspirations, and values.14 Stanley Benn emphasizes the impor-
tance of resolving conflicts within the self to achieve a degree of coherence, some-
thing also stressed by Marilyn Friedman.15
On these various accounts, the difference between an autonomous person and
a person who fails to be autonomous is not the difference between a person who
mysteriously escapes the forces of socialization and one who does not. Both the au-
tonomous and the nonautonomous are conditioned by the forces of society. The
difference is that the autonomous person is not a passive receptacle of these forces
but reflectively engages with them to participate in shaping a life for herself. Thus
the strong sense in which autonomy is said to be incompatible with the socially de-
termined self—the claim that because we are socially determined, autonomy is an
illusion—is to be rejected because it is committed to an unrealistic conception of
autonomy. Once we understand that autonomy does not imply the simple shedding
of social influence but the ability to fashion a certain response to it, then social de-
terminism does not entail that autonomy is an illusion.

Degrees of Social Determinism

A second, weaker concern raised about the relationship between autonomy and so-
cial determinism relates to the extent that a person must critically reflect on her ends
to achieve autonomy. Although we may fashion more or less autonomous responses
to social influence, our aspirations and aims will always bear the markings of those
social relationships most pervasive in our lives—to family, culture, and so on. No-
body ever wholly manages to critically reconsider all of the socially determined val-
ues, aims, and aspirations that indelibly shape her identity, no matter how often she
exercises her autonomy competency. Ultimately, to what extent one believes this un-
dermines autonomy will depend on the degree to which one believes autonomy
competency should be exercised before a threshold has been attained.
Attempts to specify this degree run the risk of either exaggerating or under-
playing the extent to which such critical reflection is both possible and desirable.
Will Kymlicka, for example, quite rightly criticizes some communitarians for their
presumption that the liberal notion of self-determination presupposes a pure ra-
tional will, devoid of any social traits.16 As Kymlicka argues, the claim that an au-
tonomous or self-determining person can question her social roles and values does
not mean she questions them all at once, that she rejects them all and then starts
afresh on the basis of nothing but the will. Someone who is nothing but a free, ra-
tional will would have no reason to choose one way of life over another. Rather, in
exercising our autonomy we hold many of our ends and purposes in place while crit-
ically reflecting on some other particular commitment. Yet, Kymlicka is himself
guilty of excess in this enthusiastic claim:
56 Autonomy and the Social

No particular task is set for us by society, no particular practice has authority that
is beyond individual judgment and possible rejection. We can and should acquire
our tasks through freely made personal judgments about the cultural structure,
the matrix of understandings and alternatives passed down to us by previous gen-
erations, which offers us possibilities we can either affirm or reject. Nothing is "set
for us"; nothing is authoritative before our judgment of its value.17

If this is meant to be a claim about the psychological capacity of each and every
individual, it disregards the force and pervasiveness of social determinism and is
clearly false. To be fair to Kymlicka, his view that every possible end is capable of
evaluation and possible rejection may nor be a psychological claim about each indi-
vidual self. 1 he claim may be a normative one, which I discuss in more detail in the
third section, "The Constitutively Social Self."

Negative Forms of Social Determinism

In addition to the global concern about social determinism and the weaker ques-
tion about the exact degree of social determinism that is compatible with auton-
omy, there is a third concern, which focuses on the incompatibility between au-
tonomy and certain forms of social determinism. In particular, the emphasis shifts
somewhat from the impact of social determinism on the content of our ends, from
the question of whether our ends are really anything more than socially deter-
mined, to the slightly different issue of the impact of social determinism on the
capacities we need to be able to exercise autonomous agency. There are certain
forms of socialization that uncontroversially militate against the development of
the kinds of competencies that are required of a person exercising procedural au-
tonomy.
Feminists have been particularly concerned with the question of socialization
and the pervasive reality of gender subordination. Diana Meyers documents the
way in which gender socialization renders women less likely to develop autonomy
competency at the level that men often achieve.18 Despite the pervasive and often
detrimental influence of gender socialization, feminists have argued that it is fre-
quently ignored by theorists purportedly concerned with individual autonomy. For
example, one of the central feminist criticisms of contemporary liberalism is that
its exclusive emphasis on formal equality and individual rights as a means to pro-
tect individual autonomy consistently fails to deliver substantial equality and indi-
vidual freedom, precisely because of its failure to address those social forces, such
as gender socialization, that radically delimit the actual choices available to some
individuals.19
We cannot presuppose that we each acquire autonomy competency as a natural
consequence of maturation. For even if social determinism per se does not threaten
autonomy, certain forms certainly do, either in virtue of what skills they fail to en-
courage or in virtue of what is actively discouraged and penalized. It follows that a
commitment to protect the autonomous individual in social and political life re-
quires scrutiny of the various social forces that affect the development of autonomy
competency and a preparedness to alter them.
 Autonomy and the Social Self 57

The Positive Influence of Social Determinism

While it may well be true that certain forms of social determinism militate against
the development of autonomy competency, it is equally true that the very precon-
dition of our being able to develop and sustain our capacity for autonomy is attrib-
utable to our developing and remaining embedded within a network of social rela-
tionships. The capacity and aspiration for autonomy is not something we are born
with but something we develop only in society. The fact that any of us has the ca-
pacity for autonomous agency is a debt that we each owe to others.
Although this point seems almost truistic, it is a truism often ignored. Feminists
argue that numerous theorists ignore that fact that we each inescapably experience
long periods of total dependency through which we develop capacities like auton-
omy. For example, the often invoked metaphor of the "state of nature" presupposes
that in the beginning man is essentially alone, as though, says Hobbes, we each
spring out of the earth and are suddenly fully mature, "like mushrooms." Seyla Ben-
habib argues that "this vision of men as mushrooms is an ultimate picture of auton-
omy."20 What is denied in the mushroom metaphor is that it is primarily women's
labors, especially in the early years, that contribute most significantly to the devel-
opment of the skills required for autonomous agency.21 Similarly, Susan Moller
Okin demonstrates how more contemporary champions of individual autonomy,
such as John Rawls and Robert Nozick, presuppose women's labor and care for the
family and at the same time deny, even if implicitly, that they are politically rele-
vant.22
It is not only in virtue of our dependency on others in the family that we ac-
quire a capacity for autonomy. In an interesting contrast with feminist critiques,
Charles Taylor denies that the acquisition of an autonomous identity is acquired
only, or even primarily, in the family.23 He argues that it is only in the context of a
whole civilization, where the value of autonomy is implicit within a whole range of
common practices—the arts, the legal system, intellectual debates, and politics—
that we can each acquire the aspiration and the ability to be autonomous. Of course
most feminists would not deny the broader relevance of the whole society, yet it is
true that their emphasis is often on the family, in particular, women's labor, whereas
communitarians have tended more to emphasize traditions, nations, and society.
The important point, however, is that both feminists and communitarians argue
that our capacity for autonomy is acquired in contexts where we are dependent on
others.
As well as dependency, the importance of interdependency should not be over-
looked. Even after a capacity for autonomy is acquired through a long period of de-
pendency, our ongoing success as an autonomous agent is affected by our ability to
share our ideas, our aspirations, and our beliefs in conversation with others. It is un-
likely that any vision or aspiration is sustained in isolation from others. As Taylor ar-
gues, "A human being can always be original, can step beyond the limits of the
thought and vision of contemporaries, can even be quite misunderstood by them.
But the drive to original vision will be hampered, will ultimately be lost in inner
confusion, unless it can be placed in some way in relation to the language and vision
of others."24 Our ongoing public debates about moral and political issues, our more
58 Autonomy and the Social

personal conversations with particular others, and even our ability at times to imag-
ine conversations with fictional or historic interlocutors enable us to effectively ex-
ercise and maintain autonomous agency. The crucial point is that in many respects
our dependency is ongoing. We do not merely acquire autonomy competency in
childhood and then become fully independent. Although the degree and nature of
a person's dependency certainly shift, that dependency never vanishes.
What emerges from this discussion is the need to acknowledge that our auton-
omy competency is a debt we owe to others. As Annette Baier argues, we arc each
"second persons." A person—a being capable of language, of mathematics and cul-
tural skills, and, we might add, autonomy—
is best seen as one who was long enough dependent upon other persons to acquire
the essential arts of personhood. Persons essentially are second persons, who grow
up with other persons. . . . The fact that a person has a life history, and that a
people collectively have a history, depends upon the humbler fact that each per-
son has a childhood in which a cultural heritage is transmitted, ready for adoles-
cent rejection and adult discriminating selection and contribution. Persons come
after and before other persons.25
I would only add that autonomous persons not only come aftet and before other
persons, but to flourish they must live concurrently with other persons as well.
In this section I have addressed the rnultifaceted claim that the self's ends, aspi-
rations, and capacities arc socially determined. I began by considering whether this
fact renders autonomy nothing more than an illusion. I argued that once we rid the
concept of autonomy of any association with the incoherent idea that there is a core
"inner self" untainted by social influence, there is no conceptual incompatibility
between autonomy and the socially determined self. This implausible global claim
in fact diverts our attention from more interesting and complex concerns about
what forms of social influence are detrimental to autonomy and what forms con-
tribute to its development. I acknowledged that certain types of social determinism
may well be at odds with the development and flourishing of autonomy, as femi-
nists in particular have emphasized. But finally, 1 emphasized that it is only in vir-
tue of the fact that we are exposed to social influence in multifarious ways that we
are ever capable of being autonomous at all, a point that further exposes the inco-
herence of the idea that truly autonomous people must escape the forces of social
determinism.

The Motivationally Social Self

Most of the issues raised in discussions of the socially determined self are concerned
with a set of descriptive facts and whether those descriptive facts about the self are
compatible with autonomy. In this section and the next I consider conceptions of
the social self that incorporate descriptive claims but also question whether the au-
tonomous self is to be valued and promoted. Proponents of these conceptions claim
that a social account of the self provides a more valuable conception of agency and
selfhood.
Some feminists have suggested that the concept of individual autonomy pre-
 Autonomy and the Social Self 59

supposes that the self is disconnected from enduring attachments to others, avoids
intimacy, and is essentially an egoistic or self-interested maximizer. Such a self is mo-
tivated above all by its own narrowly conceived selr-interest and eschews the interests
and comfort of others. For example, Lorraine Code argues, "A cluster of derivative
assumptions now attaches to ideals of autonomy. Autonomous man is—and should
be—self-sufficient, independent, and self-reliant, a self-realizing individual who di-
rects his efforts toward maximizing his personal gains."26 The concern here is not
with the fact that the self is socially determined but with the idea that human beings
are presumed to have a certain individualistic nature, a nature presupposed, either
implicitly or explicitly, by the notion of autonomy. This individualistic notion of
the self incorporates both a descriptive and a normative claim. Descriptively it as-
serts that we are by nature self-interested. Normatively, it argues that these are valu-
able character traits. Feminists deny both the descriptive and the normative claim
and defend instead the view that the self is motivationally social.
Numerous philosophers have presupposed that selves are primarily self-interested
and that the moral or political problem is to secure the legitimate means for social
cooperation, a cooperation that more efficaciously enables each individual to pur-
sue her legitimate self-interest without interfering with the similarly self-interested
pursuits of others. A number of feminists have claimed that such a picture of the
self is implicitly presupposed in a range of contractarian approaches to moral and
political philosophy, which begin with the ideal of self-interested individuals reach-
ing mutually advantageous agreement with one another.27 Some take John Rawls
to be a contemporary paradigm of this approach, graphically illustrated by the de-
vice of the original position, in which "mutually disinterested" individuals seek to
secure as many primary goods for themselves as possible. Similarly, approaches
that emphasize individual rights at the core of moral and political life presuppose
that the isolated individual needs above all to be protected from encroachment by
others.28
It is hardly surprising that feminists have focused on the individualistic concep-
tion of the self, as it is starkly at odds with many women's experience, as well as the
norms of femininity. Traditionally, women's lives have been devoted to the care of
others, and if anything the problem has been to find a space for the expression and
pursuit of one's own interests. The relationship of parent to child, the traditional
conception of what it is to be a wife, or caring for a frail and aging parent can hardly
be characterized as a form of cooperation for the efficacious pursuit of self-interest.
Feminist theory has contributed to the development of alternative theory, which fo-
cuses moral concern more on the qualities and activities appropriate to care of oth-
ers than on legitimizing individual rights as the means for protecting individuals
from one another. They have argued that as we are in fact selves characterized as
much by our capacity for care and concern for others as by our self-interest, we need
moral and political theories that are shaped according to this fact.29
Although it is undoubtedly the case that certain moral and political theories
presuppose an individualistic conception of the self, it is not plausible to suggest
that the concept of autonomy itself presupposes such a conception of the self. As
Marilyn Friedman has argued, this position conflates the concept of autonomy with
that of substantive independence.30 According to most procedural accounts of au-
60 Autonomy and the Social

tonomy, a person's choices must be procedurally independent, rather than substan-

tivcly independent. There is no requirement that one make any particular substan-
tive choices—such as to live as independently of others as possible or to eschew
commitments. It is consistent with a procedural account that an autonomous person
may be motivated above all by a sense of solidarity or attachment to various people,
causes, or communities.
On this account, a person's individual autonomy is not threatened by a deep at-
tachment to other people and thus there is no reason to suspect that there is any in-
compatibility between autonomy and the motivationally social self. The only reason
for questioning the autonomy of someone deeply committed to social and altruis-
tic ends would be if we believed that she had never questioned or thought about her
attachment to those ends, that is, if she had somehow failed to exhibit procedural
autonomy with respect to them. But as Friedman points out, doubts about some-
one's autonomy in this respect could equally apply to a man who unthinkingly ac-
cepts the masculine stereotype of the cold, distant figure. A man who truly exhibits
procedural independence in reflecting on his motivational structure is after all just as
likely to "be the one at home changing his baby's diapers [as] the one riding off into
the sunset on his Harley."31 None of this is to deny that the abstract individualism
feminists have rejected does not in fact pervade a great deal of moral and political
theory. My argument is that this kind of individualism is not a part of the core ideal
of autonomy.
Nevertheless, although autonomy is not threatened by enduring attachments
and the motivation to further the interests of others, the converse does not neces-
sarily hold: the exercise of individual autonomy may threaten certain relationships
to particular others. In fact, given the recent feminist emphasis on the importance of
relationships of attachment, particularly those between mothers and their children,
and the insistence of some feminists that the value of those relationships is not de-
rived from their being objects of choice,32 it might be that feminists should be sus-
picious of even a procedural account of autonomy on evaluative grounds. It could
be thought that a procedural notion of autonomy valorizes critical reflection and
choice over and above the integrity and longevity of relationships of attachment and
the accompanying motivation to further the interests of others.
There are two distinct although related concerns here. First, it might be argued
that it is regrettable that a person may exercise her autonomy and choose to sever
some particular attachments and thus to cease being motivated to further another's
interests. A person may autonomously decide to end a relationship to a particular
other or to a commitment, cause, or community. Although this is true, I would sug-
gest that the flip side of this particular concern is the possibility of autonomously
deciding to sever an oppressive, exploitative, or unsatisfactory relationship. As Fried-
man suggests, we cannot defend the value of relationships of attachment without
qualification, and although the endurance of some relationships is jeopardized by
the exercise of autonomy, sometimes this is a good thing.33 Even when it is not a
good thing that a particular relationship is ended by an autonomous person, we
should not disparage the worth of a certain capacity because it is sometimes exer-
cised in a bad way.34
A second concern is that a procedural account of autonomy seems to encour-
 Autonomy and the Social Self 61

aee
o
constant ratiocination and critical reflection on one's attachments to others in
the light of the acquisition of new ends, desires, or values. It is quite reasonable to
suppose that the integrity ana quality or our most valuable attachments and loyalties
will not survive constant scrutiny and assessment. This concern falsely presupposes
that an autonomous agent must constantly reassess her various attachments and com-
mitments each time she acquires new ends or aspirations. I see no reason to burden
the idea of individual autonomy with this implausible account of how often and to
what degree it must be exercised. It may be ideal that people critically reflect on their
decision to become parents or to marry a particular person, but it should not neces-
sarily be a requirement of autonomy that one be able or prepared to critically reflect
on these ends after they have been chosen. A person who is at every moment re-
assessing her attachment to the care and concern of particular others seems more
anomic than autonomous. Part of making an autonomous decision to commit one-
self to a particular person or project precisely involves a decision to close off certain
possibilities in the future, to make a decision or choice to no longer consider other
options as ongoing possibilities. Perhaps at moments of crisis or when explicit ten-
sions and incompatibilities arise, one may have to reconsider a relationship that one
had chosen in the past, but I see no reason to believe that this would, or should, in-
evitably collapse into constant ratiocination.
In this section I have considered the relationship between autonomy and the
view that people are motivationally social, that is, motivated by care and concern for
others. I have argued that a procedural notion of autonomy is compatible with the
view that the self is motivationally social and does not presuppose that we are by na-
ture essentially self-interested. In addition, I have argued that the exercise of a per-
son's autonomy may nonetheless lead her to sever some particular relationships and
thus no longer have as one of her ends a commitment to the needs and interests of
a particular other. Provided that we do not suppose that an autonomous person
must constantly assess her commitmenrs to others, the fact that autonomous agency
may sometimes lead to the rejection of certain commitments cannot always be clas-
sified as a regrettable outcome. Thus our commitment to the value of the motiva-
tionally social self is not incompatible with a commirment to the ideal of autonomy.

The Constitutively Social Self

The idea of the constitutively social self, which has been developed by a number of
communitarian critics of liberalism, stresses that the identity of the self is constituted
by various social attachments. The claim that the self is constitutively social incor-
porates the claim that the self is socially determined, for it certainly suggests that
the self's aims and aspirations are determined by rhe communities of which one is
a part. However, the constitutively social self is also meant to capture the idea that
the content of the self's ends are social, in the sense that they represent not just my
ends but also our shared ends. Compare the following two examples, only the sec-
ond of which captures the idea of the constitutively social self. A person's aspiration
to be a doctor may be determined by the strong influence of his father's preferences,
in which case the person's aspiration is socially determined. Another person's deep
commitment to the value of Catholicism is also socially determined for he was
62 Autonomy and the Social

raised within a Catholic community, where he acquired his unwavering faith. But the
commitment is one he shares with the community of Catholics and is not, there-
fore, just his own end but a socially shared end. The idea of the constitutively social
self stresses that many of our ends are not only socially determined or derived but
socially shared in this way as well.
As with feminist arguments about the motivationally social self, communitar-
ian concerns about the constitutively social self go beyond descriptive claims and
incorporate various normative arguments as well, particularly about the value of
autonomous agency and the purpose of politics. In this section I consider the com-
munitarian claim that the ideal of autonomy denies that we are constitutively social
and presents a less valuable conception of agency and selfhood than a normative
commitment to the constitutively social self allows.
Probably the most sustained critique of autonomy on the grounds of incom-
patibility with the constitutively social self is Michael Sanciel's Hegelian-inspired cri-
tique of Rawls's theory of justice.35 Sandel acknowledges the point made in the sec-
ond section that the concept of individual autonomy is fully compatible with the
notion that the self's ends may include a concern for others and an abiding interest
in promoting their welfare. In other words, he acknowledges that the autonomous
self can be motivationally social. Sandel argues, however, that there is a further sense
in which the self is bound to community that is incompatible with some of the
more exalted claims made in the name of individual autonomy. According to this
further sense,
to say that the members of a society arc bound by a sense of community is not
simply to say that a great many of them profess communitarian sentiments and
pursue communitarian aims, but rather that they conceive their identity—the
subject and not just the object of their feelings and aspirations — as defined to
some extent by the community of which they are a part. For them, community
describes not just what they have as fellow citizens but also what they are, not a
relationship they choose . . . but an attachment they discover, not merely an at-
tribute but a constituent of their identity.36
On this view, to speak of the social self is not merely to suggest we experience
a spirit of benevolence or cooperation with others or that we choose attachments of
solidarity, love, or commitment. Rather the claim is that certain ends and values con-
stitute our identity and that these ends and values are fundamentally social in nature.
Sandel argues that it is by reference to these shared ends that we each determine our
good. When the self" engages in the sort of self-reflection that theorists of autonomy
describe, when one reflects on what one wants or what one should do, what one dis-
covers are "shared values partly constitutive of a common identity or form of life."37
It is these shared values, partly constitutive of a common form of life—"whether a
family or tribe or city or class or nation of people"—that constitute the self's iden-
tity. Famously, Sandel argues that the fundamental question for the constitutively
social self is not "What ends shall I choose?" but "Who am I?" and the latter ques-
tion is answered by looking inward and discovering one's shared constitutive ends.
Sandel argues that because a conception of autonomous agency presupposes that the
self chooses rather than discovers one's ends, it is incompatible with a conception of
 Autonomy and the Social Self 63

the self as constitutively social. Thus, at this descriptive level Sandel suggests an in-
compatibility between the concept of autonomous agency and the constitutively so-
cial self.
But there is normative incompatibility as well. Sandel claims that agency
which proceed^ via the discovery of constitutive ends is more valuable, deeper, than
autonomous agency. Sandel finds a shallow notion of autonomous agency in Rawls's
account of how we each come by our ends and purposes. According to Rawls, a
person comes by her life plan by exercising deliberative rationality, a process that
involves working out what one wants and how much these things are wanted and
then choosing the plan that satisfies as many of these wants as possible. What
Sandel deplores about this conception of agency are the conceptions of choice and
reflection that underpin it. Within deliberative rationality there is no reflection on
the worth or value of desires, only reflection on their relative intensities. For agency
to be worth our respect, it proceeds not simply by closer inner inspection of one's
desires and inclinations but also on the basis of the values that constitute one's
identity. But, Sandel claims, it is only inasmuch as we have constitutive social
traits, shared values constitutive of a shared way of life, that we have languages of
worth and qualitative hierarchies with which to evaluate our desires and prefer-
ences. If our agency is to consist of more than merely sorting out which of our de-
sires is strongest, "we must be capable of a deeper introspection than a 'direct self-
knowledge' of our immediate wants and desires allows."38 Such deep introspection
reveals values that express a common identity or way of life.
I will discuss Sandel's descriptive and normative claims in turn. In considering
his descriptive claim that autonomy is incompatible with the constitutively social
self, let us concede from the beginning that critical reflection on one's ends often
does proceed with explicit reference to one's values, and thus autonomous agency is
not (always) just a matter of working out which of one's desires is most intense. But
why must we assume that insofar as our values are relied on to structure our critical
reflection, they themselves are not subject to such critical reflection? Why must we
assume that they are simply the "discovered" values of tribe, nation, and family? To
be sure, not everything—values, desires, aspiration, aims—can be up for critical as-
sessment at the same time. Something (most things) are usually held in place. Thus
our values, insofar as they are relied on in reflection, are not themselves at that time
subject to critical questioning. But it does not follow that at other times these val-
ues themselves are not subject to autonomous consideration. We can agree that val-
ues are socially influenced and often socially shared by parents, peers, and culture.
We can follow Gerald Dworkin, who suggests, "It makes no more sense to suppose
we invent the moral law for ourselves than to suppose that we invent the language
we speak for ourselves."39 But it doesn't follow that such values are necessarily con-
stitutive of our identity in the sense of not themselves being subject to revision and
rejection.
Sandel's critique of autonomous agency ultimately hinges on an ambiguity re-
garding what is meant by the descriptive claim that one is constituted by shared
ends. There are two possible interpretations of this claim. The strong version says
that the self is so constituted by her social ends that she is unable to reconsider or re-
ject them. This is very implausible. Most theorists of autonomy believe that the ex-
64 Autonomy and the Social

ercise of individual autonomy may well be directed at such socially derived and shared
values, which are then vulnerable to reshaping or even rejection. It is worth reiterat-
ing here that this is not an implausible psychological claim that any individual can,
at any time, reflect on and reject such attachments. Perhaps many of us, with respect
to certain ends, cannot. But conversely, it is equally implausible to imply that we can
conceptually hive off that part of individual identity defined by socially derived and
shared values as necessarily invulnerable to critical scrutiny and possible rejection. In
certain circumstances and through the deployment of certain competencies, some
individuals can and do reshape their attachment to particular values, including those
that define a common way of life. As Marilyn Friedman insists, we know "that atti-
tudes and behavior sometimes are independent. We know that there are social crit-
ics and social deviants."40
There is a second, weaker interpretation of the claim that we are constituted by
our shared ends that simply claims that even the most autonomous person's identity
is always mediated to some extent by community. Our starting point will always af-
fect where we end up, so that even if some particular shared values are rejected by
the individual, she will continue nonetheless to bear as part of her identity some
markings of her original communities of family, nation, and tribe. Sandel himself
sometimes offers a version of this claim: "As a self-interpreting being, I am able to
reflect on my history and in this sense to distance myself from it, but the distance is
always precarious and provisional, the point of reflection never finally secured out-
side the history itself."41 But this claim is compatible with a suitable account of au-
tonomous agency, as I argued in the first section. Nobody makes radical choices
from an empty starting point. The key question remains: are socially shared values
sometimes rejected by individuals, even if, of course, no individual can reshape her-
self de novo? The only plausible answer is yes, which immediately undercuts the
strong version of the claim that the self is constitutively social. It turns out that
many of us are not, after all, doomed to live out the roles and embrace the particu-
lar ends that are given to us by family or nation. This is fully compatible with most
procedural accounts of autonomy.
The difficulties for Sandel's view do not end here, for his normative claims are
also vulnerable to objection. Sandel claims that agency which proceeds via the dis-
covery of constitutive ends is more valuable, deeper, than agency that proceeds on
the basis of working out which of one's desires is stronger or more intense, that is,
agency that he caricatures as autonomous. Let us suppose for a moment that reflec-
tion on the intensity of one's desires really is all that autonomous agency amounts
to. It is still not obvious that such agency would be less worthwhile than agency in
which we discover those shared values that define us each in the constitutive sense.
Why exactly does this second kind of agency exhibit greater depth than the kind of
agency that Sandel so deplores in Rawls's account? It is unclear how this could be so
if the qualitative or evaluative distinctions the self draws on are nothing more than
the discovered ends of tribe, family, and nation.
Consider the following example.42 A particular woman has as one of her most
abiding values the belief that a woman's place is in the home, particularly if one is a
mother. This woman's constitutive ends are shared ends in just the sense that Sandel
champions, for she lives in a close-knit community defined by its deep commitment
 Autonomy and the Social Self 65

to gender roles and the centrality of motherhood in women's lives. These values rep-
resent their shared way of life. In reality, this particular woman is persistently dis-
satisfied with her life in the home and fantasizes often about a different life, one
centered around projects other than full-time parenting. Despite this, she never
questions her values, for they are constitutive of her identity and it is by reference
to them that she is able to critically assess her desires and preferences. Thus, she is
inclined to feel disgusted with herself for her selfish desires.
Similarly, a single man quite consistently finds his sexual desires oriented to-
ward other men and often indulges in a fantasy in which he develops an ongoing re-
lationship with another man. He reflects on these desires and decides that they are
wrong, indeed absolutely disgusting. The values that he employs to reflect critically
on his desires are values both derived from and shared with his community, a com-
munity that staunchly embraces the heterosexual family unit as constitutive of its
shared way of life. He persistently represses his mere desires and ends up, unsurpris-
ingly, a miserable person.
In each of these cases, although a person acts on the basis of her or his consti-
tutive values as opposed to mere desires, it is far from clear that either exhibits the
kind of agency that deserves respect. Each case represents an example of the kind of
agency that follows from being constitutively social and deploying these discovered,
shared values in reflection. Yet it is difficult to appreciate why the kind of agency
represented in each case is "deeper" or deserves greater respect than that which is pri-
marily devoted to a closer inspection of one's desires and inclinations. Values may
well play a significant role in autonomous agency, but to play an important role,
they have to be more than purely discovered, constitutive social traits. Values them-
selves must to some extent be subject to the self's autonomous scrutiny. This
suggestion is supported by the fact that what seems wrong in both of our not-so-
imaginary examples is that they portray stark examples of a person laboring under
socially imposed and uncritically accepted values in a way that thwarts individual
autonomy. Each has been unduly influenced by the values of others.
I have argued that neither Sandel's descriptive nor his normative criticisms of
autonomous agency based on the ideal of the constitutively social self are plausible.
The only plausible descriptive account of the constitutively social self is a weaker
version fully compatible with the idea of autonomy. The normative value attributed
to agency based on the discovery of shared ends and values is also highly debatable.
Thus, whereas it may well be true that particular versions of the constitutively social
self are incompatible with the idea of autonomy, it is these versions of the consti-
tutively social self that we should repudiate, not the idea of autonomy.

The Value of Autonomy: Rejecting a Communitarian

and Feminist Alliance
So far I have distinguished three different senses in which the self is social in order
to consider more specifically the claim that autonomy is incompatible with social
conceptions of the self. Although social conceptions of the self have been promoted
by both feminists and communitarians, I have attempted to highlight the differ-
ences between them. I now want to argue more explicitly that the concerns that
66 Autonomy and the Social

feminists and communitarians have voiced about autonomy are only superficially
similar and conceal a deep division over the value of autonomy. Some communitar-
ians believe we should repudiate the value of autonomy, whereas, I argue, feminists
should not reject its value. To see how feminism and communitarianism pull in dif-
ferent directions over the value of autonomy, it is important to understand the very
different motivations each has for focusing on the fact that selves are social and the
distinct conclusions each draws from such views.
One of Sandel's more explicit normative aims, an aim shared by many commu-
nitarians, is to see our shared, constitutive values embodied within political institu-
tions. Communitarians reject liberal neutrality, the principle that the state itself is to
play no role in deciding which ways of life are better than others. Many liberals
argue that state neutrality protects individual autonomy and enables individuals to
decide for themselves which values and ways of life they will endorse and which they
will reject.43 Communitarians criticize the liberal assumption that the state's highest
duty must always be to protect individual autonomy. A good state also acts to pro-
tect shared values and a shared way of life. Indeed, Sandel explicitly suggests that
when political institutions embody shared values, we can "know a good in common
that we cannot know alone."44 In other words, embodying our common, constitu-
tive values within our shared institutions represents a more powerful vision of what
politics can be than the favored liberal alternative, which begins with the assumption
that individuals must be protected in seeking their own individual conceptions of
the good.
Where individual autonomy thrives within a political community, very few
shared values remain invulnerable to critical rejection. Underpinning some commu-
nitarian objections to the valorization of autonomy is precisely the fact that individ-
ual autonomy renders shared values vulnerable to rejection. Individual autonomy,
therefore, should be subordinated to the alternative task of enabling shared values to
nourish. Politics, on this view, goes better when individual autonomy is subordi-
nated to the aim of building a flourishing political community based on shared val-
ues. On this view, the incompatibility between autonomy and the social self is a nor-
mative incompatibility, one whose resolution calls for a repudiation of the value of
autonomy.
Should feminists join communitarians in repudiating the value of autonomy in
favor of the common good? Of course, balancing the value of autonomy with the
value of a community of shared ends is difficult, and I cannot attempt an exhaustive
discussion of this issue here. My rejection of Sandel's normative defense of agency
based solely on the discovery of shared ends in the previous section indicates that
feminists should certainly be cautious about repudiating the value of autonomous
agency. I want to conclude this chapter by building on this concern in order to
suggest that feminists should respond with a cautious "no" to a proposed feminist-
communitarian alliance.
It is clear from the first and third sections that communitarians and feminists
share some version of the descriptive claim that selves are socially determined. But
for feminists, the importance of this insight is that it opens up the possibility of
changing our identities and challenging existing social structures. Precisely because
our identities—our aims, aspirations, and capacities — are socially determined, we
 Autonomy and the Social Self 67

can repudiate the historically entrenched view that women (and others) have a cer-
tain fixed and immutable nature, a nature that suits them for specific roles and dis-
qualifies them for others. Similarly, if women have lacked the skills and capacities, as
well as the opportunity, to do certain things, such as exercise autonomy, this is not
a natural fact about women but a consequence of social determination. The truth
that selves are socially determined carries with it a certain liberating potential, a de-
nial that social roles need be fixed and a repudiation of the claim that selves have an
immutable nature that determines their roles. As Benhabib and Cornell argue, fem-
inists begin "with the situated self but view the renegotiation of our psychosexual
identities, and their autonomous reconstitution by individuals as essential to women's
and human liberation."45
In a striking contrast, some communitarians seem to draw precisely the oppo-
site conclusion. The fact that we are each born into certain social relations within
which our identities are shaped is a feature of ourselves not to be questioned but
embraced. The fact that our identity is socially determined does not expose the con-
tingency of that identity for the communitarian but entrenches the view that it is
normatively vacuous to attempt to challenge or reshape the boundaries of our social
roles and values. Of course, communitarians do not think that these roles and iden-
tities are given by nature, but they do seem to believe that they are, or should be,
more immutable than any feminist could be satisfied with. Alasdair Maclntyre
writes:
It is through his or her membership in a variety of social groups that the individ-
ual identifies himself or herself and is identified by others. I am brother, cousin
and grandson, member of this household, that village, this tribe. These are not
characteristics that belong to human beings accidentally, to be stripped away to
discover "the real me." They are part of my substance, defining partially at least
and sometimes wholly my obligations and my duties. Individuals inherit a partic-
ular space within an interlocking set of social relationships; lacking that space,
they are nobody, or at best a stranger or an outcast.46

There are versions of Maclntyre's claim that we could agree with, one example being
that we cannot be autonomous in the absence of social relationships. But Maclntyre
explicitly insists on the necessity of inherited relationships, which he believes we can-
not do without. Feminists have argued that not only can we do without the partic-
ular roles and identities we inherit but also that sometimes, at least, we are better off
without them. As Penny Weiss argues, "Communitarians are concerned with the
loss of 'traditional boundaries,' while feminists are concerned with the costs of those
boundaries, especially for women."47
For feminists, the liberating potential opened up by a recognition that our
identities are socially determined is particularly crucial, given that those identities
are often indisputably oppressive. It has been noted by a number of feminists that
communitarian valorization of "traditions," "community," "tribe" and "nation" fre-
quently ignore their sexist and racist practices and the multifarious ways in which
they contribute to gender oppression, the subjugation of certain ethnic groups, the
exclusion of gays, and so on.48 It is precisely the kinds of communities celebrated
by communitarians that women have often rejected because of their exploitative or
68 Autonomy and the Social

oppressive nature. Feminism as a shared movement has been largely defined by the
various ways it has resisted and challenged traditional communities.
These considerations enable us to express caution toward the normative attach-
ment to a community of shared ends that communitarians have championed. Com-
munitarians like Sandel argue that our common good, our shared values, should be
embodied in the social and institutional arrangements of society and that the value
of a common good of this nature takes precedence over the value of individual au-
tonomy. But we cannot both be alert to the potentially oppressive features of social
relationships and assign unconditional worth to them. Indeed, as I argued in the sec-
ond section, this same problem arises in the context of the motivationally social self.
Similarly, in the previous section, I rejected SanciePs normative defense of agency
that is based solely on the discovery of shared values precisely because such agency
consigns many individuals to oppressive social roles. We need to consider which par-
ticular attachments to nourish, which particular shared values should be part of our
common good. The answer to these questions cannot fail to be enormously complex
once we squarely face up to the fact that the flourishing of autonomy means the loss
of traditional, hermetically sealed communities characterized by shared values and
the lack of dissent. There are losses here undoubtedly; but for women and many
others, there have also been incomparable gains.
Some feminists, like Friedman, have wondered why communitarians valorize
unchosen communities like family, tribe, and nation and ignore "communities of
choice" like friendships. 4y The answer is that to acknowledge that some of our en-
during and most valuable social relationships are communities of choice, as feminists
have done, and that some of our most cherished values grow out of a rejection of
other inherited ones is to concede implicitly both the possibility and the value of au-
tonomous agency. To value communities of choice is to reject the strong descriptive
claim that we are so constituted by social relations and shared values that we are un-
able to reconsider our attachment to them. But it is also to reject the value of agency
that is based solely on the discovery of shared ends and the value of thwarting the in-
dividual's ability to decide for herself which communities she will belong to in the
name of the common good. Precisely because feminists are committed to communi-
ties of choice, I think it is doubtful that we could transform "the communitarian vi-
sion of self and community into a more congenial ally for feminist theory."50
To consider which particular attachments we should reshape, which to reject,
which to choose, and which to promote, we need autonomy. We do not simply want
to endorse what we inherit and later discover. This is neither a descriptively accurate
assessment of what our agency is capable of nor a liberating vision of what kind of
agency deserves our evaluative endorsement. Nor is it a liberating vision of political
community. We need and should cherish the capacity to decide and choose. This
will always be a shared task, one that we do in concert and conversation with oth-
ers. It is also something that any of us can do in virtue of the fact that we experience
long periods of dependency and interdependency. It will be a task motivated by
deep, sometimes irreconcilable differences but also by equally unshakeable shared
commitments and concern for others. The truth of these claims shows that we are
quite compatibly both autonomous agents and deeply social selves. It also shows
that some versions of communitarianism are perilous allies for feminists.
 Autonomy and the Social Self 69

Notes

1. I would like to thank Catriona Mackenzie and Natalie Stoljar for their helpful and
thorough feedback on an earlier draft of this article.
2. Lorraine Code, What Can She Know? Feminist Theory and the Construction of Knowl-
edge (Ithaca, N.Y.: Cornell University Press, 1991), p. 77.
3. Jennifer Nedelsky, "Reconceiving Autonomy: Sources, Thoughts and Possibilities,"
Yale Journal of Law and Feminism 1 (1989): 7-36.
4. Sarah Lucia Hoagland, Lesbian Ethics, Toward New Value (Palo Alto, Cal: Institute of
Lesbian Studies, 1988), pp. 144-145.
5. Alasdair Maclntyre, After Virtue. A Study in Moral Theory (Notre Dame, Ind.: Uni-
versity of Notre Dame Press, 1981), 32, pp. 232-233.
6. Michael Sandel, Liberalism and the Limits of justice (Cambridge: Cambridge Univer-
sity Press, 1982).
7. Marilyn Friedman, "Feminism and Modern Friendship: Dislocating the Commu-
nity," in What Are Friends For? Feminist Perspectives on Personal Relationships and Moral The-
ory (Ithaca, N.Y.: Cornell University Press, 1993), p. 232.
8. See for example, ibid.; Penny A. Weiss, "Feminism and Communitarianism: Com-
paring Critiques of Liberalism," in Feminism and Community, ed. Penny A. Weiss and Mari-
lyn Friedman (Philadelphia: Temple University Press, 1995); Elizabeth Frazer and Nicola
Lacey, The Politics of Community. A Feminist Critique of the Liberal-Communitarian Debate
(London: Harvester Wheatsheaf, 1993). The attempt to assess whether or not there are
grounds for a feminist-communitarian alliance has been a decidedly one-sided affair, with
communitarians overwhelmingly displaying no interest in feminist theory.
9. Of course these three senses of the social self overlap, so I am not therefore suggest-
ing that the deterministic, motivational, and constitutive senses of the social self are entirely
unconnected. Nonetheless they are distinct claims, which for purposes of clarity and effective
critique I have disentangled.
10. See, for example, Gerald Dworkin, The Theory and Practice of Autonomy (Cam-
bridge: Cambridge University Press, 1988); Diana Tietjens Meyers, Self, Society, and Personal
Choice (New York: Columbia University Press, 1989); Robert Young, Personal Autonomy: Be-
yond Negative and Positive Liberty (London: Croom Helm, 1986).
11. Meyers, in Self, Society, and Personal Choice, refers to autonomy as a set of key com-
petencies—the skills of self-reading, self-definition, and self-direction.
12. Harry Frankfurt, "The Freedom of the Will and the Concept of a Person," Journal
of Philosophy 68 (1971): 5-20.
13. Meyers, Self, Society, and Personal Choice.
14. Young, Personal Autonomy.
15. S. I. Benn, "Freedom, Autonomy, and the Concept of a Person," Proceedings of the
Aristotelian Society','6 (1975/1976): 109-130; Friedman, "The Social Self and Partiality De-
bates," in What Are Friends For?
16. Will Kymlicka, Contemporary Political Philosophy. An Introduction (Oxford: Oxford
University Press, 1990), pp. 207-216.
17. Ibid., pp. 210-211.
18. Meyers, Self, Society, and Personal Choice, part 3.
19. See, for example, Frazer and Lacey, Politics of Community, Seyla Benhabib and Dru-
cilla Cornell, eds., Feminism as Critique: Essays on the Politics of Gender in Late-Capitalist So-
cieties (Cambridge: Polity Press, 1987), introduction.
20. Seyla Benhabib, "The Generalized and the Concrete Other," in Situating the Self: Gen-
der, Community and Postmodernism in Contemporary Ethics (New York: Roudedge, 1992), p. 156.
70 Autonomy and the Social

21. See Marilyn Friedman, "Autonomy, Social Disruption, and Women," in this volume.
22. Susan Moller Okin, Justice, Gender, and the Family (New York: Basic Books, 1989),
chap. 4; Robert Nozick, Anarchy, State and Utopia (New York: Basic Books, 1974).
23. Charles Taylor, "Atomism," in Philosophy and the Human Sciences: Philosophical Pa-
pers, vol. 2 (Cambridge: Cambridge University Press, 1985).
24. Charles Taylor, Sources of the Self. The Making of Modern Identity (Cambridge:
Cambridge University Press, 1989), p. 37. See also Friedman, "Friendship, Choice and Change,"
in What Are Friends For?fvi a discussion of the importance of friendships in sustaining the
possibility of individual autonomy.
25. Annette Baier, "Cartesian Persons," in Postures of the Mind: Essays on Mind and
Morals (Minneapolis: University of Minnesota Press, 1985), pp. 84-85.
26. Code, What Can She Know? ?• 77.
27. 1 homas Hobbes is, of course, one of the better-known theorists who presupposes
such a picture of persons, but there are contemporary exponents as well, such as David Gau-
thier, Morals by Agreement (Oxford: Oxford University Press, 1986).
28. See, for example, Nedelsky, "Reconceiving Autonomy," p. 12; Virginia Held, "Non-
Contractual Society: A Feminist View," Canadian Journal of Philosophy, Supplementary 13
(1987): 111 -137; Code, What Can She Know?p. 77. I should note here that I actually think
this is a mistaken way to understand the role played by the notion of mutual disinterest in
Rawls's theory.
29. See in particular Carol Gilligan's In A Different Voice: Psychological Theory and
Women's Development (Cambridge, Mass.: Harvard University Press, 1982). Some of the lit-
erature inspired by Gilligan includes Eva Fedcr Kittay and DianaTietjens Meyers, eds., Women
and Moral Theory (Totowa, N.J.: Rowman & Littlefield, 1987); Virginia Held, ed., Justice
and Care: Essential Readings in Feminist Ethics (Boulder, CO: Westview Press, 1995).
30. Marilyn Friedman, "Autonomy and Social Relationships," in Feminists Rethink the
Self, ed. Diana Tietjens Meyers (Boulder, Colo.: Westview Press, 1996).
31. Ibid., p. 53.
32. See Annette C. Baier, "The Need for More Than justice," in Held, Justice and Care,
Held, "Non-Contractual Society."
33. See Friedman, "Autonomy, Social Disruption and Women," for an extensive dis-
cussion of this point.
34. This seems quite obvious once we consider that an autonomous person can make
evil or malicious choices.
35. Sandel, Liberalism and the Limits of Justice. My aim is not to assess Sandel's critique
of Rawls but to consider his claims about autonomous agency more generally.
36. Ibid., p. 150.
37. Ibid., p. 167.
38. Ibid., p. 172.
39. Dworkin, Theory and Practice of Autonomy, p. 36.
40. Friedman, "Social Self and Partiality Debates," p. 76 (emphasis in the original).
41. Sandel, Liberalism and the Limits of Justice, p. 179.
42. I take the example from Marilyn Friedman, "Autonomy and the Split-Level Self,"
The Southern Journal of Philosophy 24, no. 1 (1986): 19-35.
43. See John Rawls, A Theory of Justice (Oxford: Oxford University Press, 1971), and
Kymlicka, Contemporary Political Philosophy.
44. Sandel, Liberalism and the Limits of Justice, p. 183.
45. Benhabib and Cornell, Feminism as Critique, p. 13 (emphasis in the original).
46. Maclntyre, After Virtue, p. 32.
 Autonomy and the Social Self 71

47. Weiss, "Feminism and Communitarianism," p. 167.

48. See ibid.; Friedman, "Feminism and Modern Friendship"; Frazer and Lacey, Poli-
tics of Community.
49. Friedman, "Feminism and Modern Friendship."
50. Ibid., p. 234.
 3
FEELING CRAZY
Self-Worth and the Social Character of Responsibility1

Paul Benson

Crazy a. 1. Full of cracks or flaws; damaged, impaired, unsound; liable ro

break or fall ro pieces.
(Oxford English Dictionary)

I can hear you say, "What: a horrible, irresponsible basrard!" And you're
right. I leap to agree with you. Irresponsibility is part of my invisibility;
any way you face it, it is a denial. But to whom can 1 be responsible, and
why should I be, when you refuse to see me? And wait until I reveal how
truly irresponsible I am. Responsibility rests upon recognition, and recog-
nition is a form of agreement.
(Ralph Ellison, Invisible Man)

Feminist Interests in Responsibility

The notion of responsibility has figured prominently in feminist theorizing. Feminists
have explored ways in which received conditions of blameworthiness and excuse mask
and thereby sustain social arrangements that oppress women. Men who reap advan-
tages from such arrangements are commonly in a position to justify their conduct by
appealing to gendered social norms that grant men special prerogatives ("Lighten up!
Surely there is nothing wrong in my just looking, teasing . . .") or by professing their
innocent motives ("I didn't mean any harm by it"). Alternatively, men may exempt
themselves from responsibility by deflecting accountability away from individuals to
larger social systems over which individuals seem powerless ("It's not my fault; that's
just the way the system works"). In such ways, conditions of responsibility are inter-
preted so as to license male irresponsibility. Systematic distortion of responsibility is a
feature of most resilient forms of social oppression, not only gender-based oppression.
Responsibility also has been at the heart of much feminist practice and ac-
tivism. Feminists have sought to demystify responsibility by clearly naming male

72
 Self-Worth dr the Social Character of Responsibility 73

perpetrators of violence, abuse, exploitation, and harassment and by giving voice, as

ones who have been wrongly blamed for the harms inflicted on them, to their own
ethical identities and integrity. In the process of struggling with the paralyzing and
polarizing effects of self-blame and the reproach of other women, feminists have
also underscored the importance of women's taking responsibility for their feelings,
perceptions, desires, beliefs, and actions as a prerequisite for reclaiming and expand-
ing their personal and political agency. Part of the significance of consciousness
raising for feminism has been that it offers women a way to take responsibility, col-
lectively and individually, for their own interpretations of their experiences and
activities.2 In this vein, Virginia Held writes, "Feminist thought is committed to
keeping experience in mind, and feminist experience must include personal test cases
of a self, a female subject, taking responsibility for resisting the language and thought
and practices that confront us and maintain our subordination."3
Perhaps since responsibility has served these conflicting roles—at once a tool
for reinforcing oppressions leveled against women and also a valuable component
of resistance—philosophical work in feminist ethics has displayed considerable
ambivalence about responsibility. On the one hand, responsibility, conceived in
terms of interpersonal responsiveness, attention, and care, has been favored in
many formulations of an ethic of care.4 On the other hand, many feminists have
regarded responsibility with deep suspicion. The notion of personal moral respon-
sibility in particular has been associated with abstract, individualistic, or atomistic
moral psychologies that have their historical home in the liberal moral and political
theories of the Enlightenment. So conceived, the responsible person is the self-
governing, self-constituting, and self-sufficient agent whose natural freedom and
moral identity do not essentially depend on socially elaborated powers, roles, and re-
lationships. Many feminists have argued that such conceptions of agency obstruct
forms of critical social and psychological analysis useful for advancing women's in-
terests and frustrate the articulation of women's experiences of socially situated, in-
terpersonally constituted motives, decisions, and actions.5
Seen in this light, it is unsurprising that women's purported responsibility and
blameworthiness (for "choices" of body size and shape, of poverty, of unruly sexu-
ality and emotion, and of violence-ridden lives) should be a common theme of
moral discourse in male-dominated societies. Accordingly, many feminist ethicists
have taken pains to highlight the dangers of responsibility attribution and blame in
oppressive social contexts.6 Sarah Hoagland contends:
Once we begin considering questions of praise and blame among the oppressed,
we find the extremes of blaming the victim or victimism. As it stands now, some-
one who is harmed under oppression is held responsible for everything that hap-
pens to her—not only for her choices, but for the situation itself (women invite
rape, slaves never resisted slavery, jews [sic] were willing victims, and so on). Alter-
natively, she is perceived as a total victim, as if she were not making choices and
trying to survive and go on. ... Moral accountability as we understand i t . . . does
not present us with a viable concept of choice under oppression.7

The result is that extended treatment of conditions of personal moral re-

sponsibility has belonged largely to philosophical projects that are not motivated
74 Autonomy and the Social

explicitly by feminist concerns. My project in this article will be to hold some main-
stream accounts of moral responsibility more accountable by examining a theoret-
ically neglected form of diminished responsibility that shows those accounts to be
insufficient. This form of impaired responsibility comes to light by analyzing the
profound self-doubt that rigid feminine norms can instil when internalized. I argue
that standard conditions of responsibility oriented around agents' abilities of re-
flective self-control or the reasons-responsiveness of their actions turn out to be in-
sufficient for responsibility when studied in the light of persons' own standpoints
toward their agency, especially their sense of worthiness to answer for their actions.
In this way, I hope to respect Alison Jaggar's dictum that "from a feminist point of
view, the call to reflect on women's moral experience is politically and method-
ologically indispensable."8
Furthermore, the unorthodox condition of responsibility that I begin to de-
velop promises to bear fruit for feminist ethics by clarifying the extent to which re-
sponsibility is a social, or relational, matter. This condition highlights overlooked
ways in which responsible agency is not only shaped causally and developmentally
by social conditions but also intrinsically depends on various types of interpersonal,
moral relations.
I begin by considering a historically located example of a horrifying—and
philosophically illuminating—sort of self-doubt.

The Rest Cure and Self-Doubt

Let us imagine that Charlotte,9 a white, middle-class American woman who is com-
ing to adulthood in the 1880s, finds the rigid, conventional roles that women of her
class are expected to take up—wife, mother, and housekeeper—to be nearly in-
tolerable. She detests conventional feminine roles so thoroughly that she wants very
much to leave her husband and young child to pursue an artistic career, writing,
painting, and lecturing. But these interests are so alien from the standpoint of middle-
class Victorian norms of womanliness that Charlotte's friends and family fear for her
sanity. She is sent to a well respected physician, who diagnoses her as suffering from
a serious nervous disorder. The physician prescribes the rest cure: Charlotte is to re-
main in bed as much as possible, without visitors, and is not to read, write, or paint.
Charlotte concurs with the doctor's judgment, feeling that she really must be emo-
tionally ill if she cares about a potential artistic career more than her domestic
responsibilities. However, she simply cannot endure the rest cure and the doctor's in-
sistence on "moral medication."10 She feels that she may not survive it. She also be-
lieves that she has become an excessive burden to her husband and child. So she
leaves them and sets out in hope of launching her career. All the while she contin-
ues to feel that she must be quite emotionally deranged to be doing such a thing. In
fact, she regards her leaving as good evidence of her mental infirmity.
I want to consider some of the effects that Charlotte's internalization of the
restrictive norms of femininity of her day, and her resulting feelings of craziness,
could have on her capacity for moral agency. For my present purposes, what is sig-
nificant about Charlotte's predicament is not simply the fact that she experiences a
practical conflict that arises from conflicting ethical responsibilities; nor does my
 Self-Worth dr the Social Character of Responsibility 75

concern lie merely with the fact that this conflict arises from social expectations that
severely narrow the options Charlotte could pursue without facing condemnation
from her peers. Rather, I want to explore the fact that having internalized oppressive
social conventions, Charlotte decides to leave her family while doubting her very ca-
pacities to make such decisions in a morally competent way, and she regards that de-
cision as a symptom of her craziness. Charlotte's sense of derangement takes the
form of moral self-doubt. Furthermore, her sense of craziness may implicate an im-
portant element of her sense of moral worth, namely, her sense of her worthiness to
answer for her conduct. Charlotte's inner conflict about her capability to care about
both her domestic responsibilities and relationships and her artistic aspirations in
ways that seem to her and her peers to be appropriate might displace her sense that
she has the standing in the moral community that would allow her to give an ac-
count of herself as a moral agent in response to others' potential moral criticisms of
her. The feature of Charlotte's case that I concentrate on, then, is that her sense of
craziness comes about in such a way that she doubts her own moral capacities; and
this doubt jeopardizes that portion of her moral self-respect that consists in her
sense of her worthiness to answer for her actions.11
I examine some of the effects that Charlotte's attitudes about her craziness might
have on her capacity for moral agency by focusing on her responsibility for leaving
her family. In drawing attention to her moral responsibility, I am not supposing that
her action was wrong. Nor am I supposing that we are in a good position to deter-
mine what it would have been right (much less best) for her to do. However, this
does not mean that taking up the question of Charlotte's responsibility is artificial or
inappropriate. Given the divisions in Charlotte's attitudes toward her moral compe-
tence and her worthiness to give an account of her conduct, and because most of
her contemporaries (along with too many of ours, I fear) would be inclined to judge
her behavior as a serious moral wrong, there is a live question about what Charlotte's
action can reveal about her as a moral agent and about what range of responses to
her would be appropriate. Moreover, it will be instructive to see how three recent,
influential accounts of responsibility would treat Charlotte's conduct. As I argue in
the next section, none of these accounts can fully explain the significance that Char-
lotte's diminished sense of worth has for her responsibility.
Consider Jay Wallace's view of responsibility.12 A person is a morally account-
able agent, Wallace maintains, just when she possesses "the powers of reflective self-
control."13 These powers are general psychological capacities that admit of degrees
and are susceptible of influence by the agent's social and political circumstances. Re-
flective self-control consists in "the power to grasp and apply moral reasons" and
"the power to control or regulate [one's] behavior by the light of such reasons."14
The latter ability, in turn, requires a capacity for critical reflection, the power to
make choices through deliberation, and the capacity to translate choices into behav-
ior.15 According to Wallace, a person who possesses these normative powers can rea-
sonably be held to moral obligations, whether or not she actually exercises these
powers when she acts, and so can fairly be held accountable for what she does.16
Full responsibility for a particular action, on Wallace's view, demands that the
agent be accountable and also that she not have excuses (such as inadvertence, phys-
ical constraint, or coercion) for the action. Legitimate excuses exist only on those oc-
76 Autonomy and the Social

casions when an agent has not intentionally done anything wrong, when her action
does not display a culpable choice.
Applying Wallace's theory to Charlotte, we see that it is possible that Charlotte
does not have any ordinary excuses for her action. Notwithstanding the conflict she
felt toward her motives, she had her reasons for fleeing the rest cure—she believed
she would not survive it and that she had become unjustifiably burdensome to her
husband and child—and she acted for those reasons. She did not act inadvertently
or as a result of physical force. Had she submitted to the rest cure, she probably
would have had the excuse that she was coerced to do so, but that is precisely what
she did not do.
More important, nothing in Charlotte's moral self-doubt dictates her lack of suf-
ficient capacities of reflective self-control to be morally accountable by Wallace's
lights. It is possible that moral self-doubt could diminish these capacities, to be sure.
Her doubts about her competence could make her so confused that she could not rec-
ognize relevant moral considerations or apply them in her deliberations; or self-doubt
could obstruct her control. In the most extreme case, her self-doubt might threaten to
cloud Charlotte's consciousness with manifestly contradictory beliefs and aims, and
thereby undo her fundamental ability to maintain a minimally coherent standpoint
from which to deliberate, choose, and act.17 But my description of Charlotte also
leaves room for the possibility that despite her self-doubt, she retains a basic level of
competence to recognize and apply moral reasons, along with the capacity to reflect
critically on possible courses of action and to form intentions accordingly. The very
fact that Charlotte feels tremendous conflict about the conventions of womanliness
that she has internalized and intentionally defies the rest cure gives us good reason to
think that Charlotte has far greater capability for reflective self-control than she or her
peers give her credit for. Charlotte's attitude toward her emotional health may mean
that she does not exercise her reflective capacities as she otherwise might, but this is
compatible with her accountability, in Wallace's theory.
There is another possible way in which Charlotte's feelings of craziness and in-
competence might impede her reflective self-control, although this is a possibility
that Wallace apparently fails to recognize. Charlotte's self-doubt could alienate her
from her power to grasp and apply moral reasons. She might feel dissociated from
any exercise of that power, so that the results of any critical reflection she might
carry out would seem to her to be no more her own than decisions arrived at by
someone else. Divorced from her judgments of what morality counsels, those judg-
ments would not serve as grounds on which she could form, authorize, and take
responsibility for her choices. This is an important possibility to note since contem-
porary discussions of responsibility normally presume that persons are necessarily
engaged in or identified with their power of critical reflection in such a way that the
actual or potential operation of that power just is the full presence of the person in
her motivation and behavior.18 Thus, Charlotte's form of self-doubt may, but need
not, interfere with the satisfaction of Wallace's conditions for responsibility.
Consideration of John Martin Fischer's and Mark Ravi/za's theory of responsi-
bility likewise confirms that Charlotte may meet standard conditions of responsibil-
ity. Fischer and Ravizza argue that agents are morally responsible for their actions just
when they have actual causal control, or "guidance control," over them. 19 The most im-
portant element of such control is that agents' actions must issue from weakly reasons-
 Self-Worth dr the Social Character of Responsibility 77

responsive mechanisms. These are mechanisms that are counterfactually responsive to

some sufficient reasons to do otherwise. In other words, the mechanism that pro-
duces a bodily movement is weakly reasons-responsive when there exists some possi-
ble world, governed by the same natural laws as the actual world, in which the agent
has a sufficient reason to do otherwise, and the same type of mechanism operates,
leading the agent to do otherwise for that reason.20 As long as the agent meets stan-
dard epistemic requirements, such as that she knows what she is doing and knows
enough about the likely consequences,21 the reasons-responsiveness of the process
that actually produces her action suffices for her responsibility for the action.
Since, as noted, Charlotte's feelings of craziness and incompetence might not
actually disrupt her capacities for reflective self-control, the process that leads Char-
lotte to decide to leave her family, however serpentine it may be, could be counter-
factually responsive to some reasons for her to do otherwise. Imagine, for instance,
the world in which Charlotte not only feels that she is a burden to her family but
also believes that her child would never adjust to her leaving and would suffer per-
manently impaired emotional growth as a result. Holding fixed the process that
leads to Charlotte s decision to leave in the actual world, this additional belief may
have been sufficient for Charlotte to decide to stay home. Similarly, were Charlotte
to believe that she lacks enough talent for an artistic career, it is plausible to think
that she would not leave her family. Surely the mechanism that actually yields her
decision to leave can be weakly responsive to possible reasons for her to do otherwise
despite her fear that there really is something crazy about her defiant attitudes to-
ward her proper place in respectable society.
Whereas Charlotte's self-doubt could be compatible with being fully responsible
in both Wallace's view and Fischer and Ravizza's theory, a third influential account
would disagree. According to this third view, persons are morally responsible for their
actions just when those actions are capable of disclosing what morally matters to them.
This claim makes up the core of what Susan Wolf terms the "Real Self View" of re-
sponsibility.22 In this view, persons are responsible when their actions are properly at-
tributable to their "real selves," this being the case when persons are able to govern
their wills on the basis of their valuational systems and to govern their actions on the
basis of their wills.23 Gary Watson has endorsed such a position,24 holding that
we are responsible for our conduct not because of its causal complexity but be-
cause and when it is self-disclosing; because, as John Dewey put it, "that conduct
is ourselves objectified in actions." On this view excusing, mitigating and exempt-
ing conditions indicate circumstances that prevent or qualify self-disclosure.25

Thus far, our discussion of Charlotte's responsibility might be taken to indicate

that her decision to leave her family can disclose her values and commitments since
she may continue to have the power to control her conduct reflectively despite her
distrust of her own competence. Charlotte leaves her family intentionally. She is
aware of her reasons for leaving (her horror of the rest cure, her belief that she is
useless to her child and husband, and her devotion to art), and these reasons seem
genuinely to be her own reasons. So it seems that her leaving can disclose what really
matters to her, and, therefore, we can morally evaluate her in light of her conduct,
not merely appraise her behavior.26 At first glance, Charlotte seems to satisfy the self-
disclosure account of responsibility.
78 Autonomy and the Social

Charlotte appears otherwise, however, when we recall that her self-doubt can
lead her to feel that she is unworthy of answering for her action. Although Char-
lotte's doubts about her moral competence will not directly interfere with the self-
disclosive potential of her conduct,27 if they undermine her sense of worthiness to
give an account of herself, this will impede normative self-disclosure. If Charlotte
feels that it is no longer her place to answer for her actions, she will also feel im-
plicitly that it is not her place to express through her actions what really matters to
her. The social status of one who is worthy to give a moral account of her conduct
(by seeking to justify it, excuse it, admit fault, or the like) is also that of one wor-
thy to disclose her valuational system through her actions. So Charlotte's sense that
she has lost full standing in the moral community will bring with it a change in her
attitude toward her worthiness to disclose her values and commitments to others
through her actions. In turn, Charlotte will not regard leaving her family, or any
other actions of hers that would be assessed in relation to the norms of conven-
tional femininity of her day, as actions that could disclose her values and commit-
ments to others. But if she does not regard this portion of her conduct as poten-
tially self-disclosing, then, setting aside the possibility of unconsciously adopted
aims or values (for that is not at issue in Charlotte's case), neither can we legiti-
mately draw inferences about what really matters to her on the basis of her conduct.
Charlotte's conduct cannot be self-disclosing when she cannot regard it as being
so because of her sense of unworthiness to express and answer for her values. Thus,
Charlotte's sense of unworthiness can diminish her responsibility according to the
self-disclosure account.
This review of three current theories of responsibility and their handling of
Charlotte's form of self-doubt is instructive for three reasons. First, none of these
theories has been designed to treat agents' attitudes toward their own moral status
in a straightforward way. They attend to agents' powers and capabilities, or to the
character of the processes by which their decisions emerge, but not directly to the
persons' sense of what they are worthy of as moral agents. Second, although all three
accounts will grant that there are ways in which moral self-doubt conceivably could
impair responsibility, Wallace's theory and Fischer and Ravizza's account will be
more inclined to count Charlotte as responsible for her action than will the self-
disclosure account. The latter theory will be more sensitive to agents' sense of worth,
although this is a feature of the theory that its proponents have not noted. Accord-
ingly, this reason for a possible divergence between the self-disclosure view and the
other theories has not been appreciated. Third, as I hope to demonstrate in the sub-
sequent section, there is a very different way in which a diminished sense of self-
worth can bear on responsibility, independently of the considerations reviewed in
this section. If I am correct about this, none of the theories reviewed here supply
conditions of responsibility that are both sufficient and theoretically adequate.

A Self-Worth Condition of Responsibility

The argument that having a certain sense of one's worth is necessary for one's being
fully responsible for one's actions begins by underscoring the connection between
being responsible and being /Wa'responsible. Persons who are morally responsible
 Self-Worth & the Social Character of Responsibility 79

for their actions are properly liable to be held responsible for them. This means, in
part, that responsible agents can appropriately be objects of reactive attitudes such as
blame, indignation, or resentment when their actions run afoul of reasonable ex-
pectations.28 Furthermore, holding a person responsible enacts a moral relationship
with her in which one characteristically expects that she should respond to appropri-
ate criticisms of her. This expectation is not primarily descriptive, a belief about
the likelihood that the person will respond. It is a normative expectation concern-
ing standards of response for one who is fairly held responsible. Thus, being morally
responsible involves being worthy of a certain social standing, that of an eligible
participant in various kinds of moral exchange, such as offering reasons, seeking ex-
cuse, begging forgiveness, and so forth.29 As Gary Watson has suggested, we might
understand the conditions of moral responsibility as conditions of intelligible moral
address.30
This normative dimension of holding someone responsible is revealed, for ex-
ample, when we openly express blame (or some other central reactive attitude such
as indignation or resentment).31 When we openly blame someone, we characteris-
tically expect that the person blamed should respond in certain ways, by admitting
fault and attempting to set it right, say, or by attempting to show that her actions
were justified, despite appearances. We need not expect the person blamed to re-
spond right then and there. Sometimes the response is called for at another time
(after thoughtful consideration, for example) or in another setting (when the per-
sons to whom harm was done can be present, for instance). There may also be cases
in which the magnitude or depth of the wrong is so great that no standard response
is called for because none of the usual options—justification, excuse, admission of
fault, or apology—will do. The gravity of the wrong done in such cases seems to
outstrip the normal human repertoire for moral response—which is not to say that
failure to respond is acceptable but rather that no response can ever be enough.
Nevertheless, openly blaming someone standardly calls for her to respond because
blaming or, more generally, holding persons responsible is not merely a matter of
appraising the quality of their actions, motives, or traits. It is also a matter of call-
ing forth a moral relationship with the person in which appropriate blame de-
mands a response, and the person blamed is expected to recognize the legitimacy
of that demand.32
The appropriateness of holding someone responsible, therefore, will depend on
the appropriateness of holding the person to this characteristic normative demand
for a suitable response. This is one reason that, as Wallace recognizes, having capac-
ities for moral competence and control is necessary for being properly held respon-
sible. Persons who lack those powers (through no fault of their own) cannot reason-
ably be expected to give a suitable moral response to moral criticism; they are not
suitable candidates for the type of interpersonal exchange that responsible agency
warrants. Likewise, the appropriateness of the normal expectation of moral response
also requires that those at whom it is directed not regard themselves as unworthy of
engaging in just that sort of relationship. If Charlotte regards herself as unworthy of
giving an account of her action (and is not to blame for having this attitude),33 she
will not be an entirely appropriate object of the characteristic demand that she re-
spond properly if blamed for leaving her family. This is not a point about Charlotte's
80 Autonomy and the Social

general moral capabilities. She may, as I have noted, have the capabilities necessary
to give suitable moral responses to potential criticisms. The point is, rather, that it
is not reasonable to demand that someone participate in a certain type of relation-
ship when she has been made to feel so deeply that she is not worthy to engage in
that relationship.
To confirm the plausibility of this argument, it is helpful to remind ourselves of
how an agent's internalization of oppressive social norms can impede the develop-
ment of or undermine her sense of worthiness to answer for her actions. Let us
imagine a society with a rigid, hierarchical caste system that places members of the
lowest caste beneath consideration as moral equals. Persons of this caste are not im-
mune to evaluative appraisal, but it is not considered to be their place to attempt to
justify, seek excuse for, or admit fault for their purported failings. It is not only that
these persons' efforts to answer for themselves would make no difference to how
others treat them; I am supposing that any efforts they might make to speak for their
own actions would be perceived as transgressing established demarcations of social
status, and so would simply not count morally. Persons in the lowest caste are nei-
ther to accept nor refuse responsibility for their acts. They are simply to undergo the
judgments of their "betters" and submit to the consequences that attach to those
judgments.
Persons who have internalized the prevailing norms of this society and who are
unfortunate enough to be assigned to the lowest social stratum will feel, and will be
given much reason to feel, that it is not their place to answer for their conduct. This
attitude will be one element of the broader lack of moral self-respect, the failure to
recognize their fundamentally equal moral worth as persons, which they are likely to
suffer. This attitude will probably be secured by an ideological framework that in-
structs that members of this caste are infantile or bestial or impure, incapable of the
kinds of moral sensitivity, reasoning, and self-control that would warrant their re-
cognition as fully accountable agents. Like Charlotte, members of this class typically
will harbor grave doubts about their own competence as moral agents. This might
obstruct whatever moral competence they actually have (and the competence of
many in this caste will probably be severely underdeveloped, given their social loca-
tion), or it might dissociate them from their powers of moral reflection. But self-
doubt may also affect these individuals in a further way, as it does Charlotte: it may
engender the attitude that they are not worthy of being moral interlocutors, of an-
swering for their own actions in response to their betters' appraisals of them. This is
an attitude about one's position or status as a moral agent in relationship to others,
not merely about one's abilities as an individual to discern various reasons and mod-
ify one's choices accordingly.
This example supports the plausibility of my claim that persons who, through
no failing of their own, feel that it is not their place to give accounts of their actions
are also not appropriately held fully responsible for what they do. Charlotte's case
obviously differs from the caste example in notable ways. Charlotte has much greater
moral competence than members of the lowest caste are likely to have. She feels con-
flicted about the conventions of femininity that she has been brought up to em-
brace, and her conflict leads her intentionally to defy those conventions. Perhaps
most significantly, Charlotte probably has been treated in the past as being worthy
 Self-Worth & the Social Character of Responsibility 81

of answering for herself, particularly with regard to matters less fully regulated by
specific gender norms. So Charlotte may retain a sense of self-worth in relation to
other spheres of her life, as persons trapped in the lowest caste may not.34 But these
distinctive features of Charlotte's situation can only affect her responsibility if they
modify the appropriateness of holding her responsible. And the caste example rein-
forces the idea that an agent's socially instilled and legitimized attitude toward her
own self-worth can alter the appropriateness of holding the agent responsible. Thus,
even though Charlotte's responsibility and the responsibility of a person in the sub-
jugated caste will be assessed differently, the agent's sense of worthiness to answer for
her actions will directly affect her responsibility in each case.
Another, more familiar example suggests that common views about the effects
of trauma on responsibility reflect the belief that socially sensitive attitudes of self-
worth are necessary for full accountability. Consider a young adolescent who has
suffered through an unusually deprived and traumatic childhood. Depending on the
particular characteristics of her trauma, this young person may be incapable of ap-
preciating certain types of moral consideration, or she may be unable to control her
actions adequately on the basis of her reflective judgments. These deficiencies in her
competence would justify lessening the extent to which she is held responsible for
some of her actions.35 But childhood trauma is widely thought to affect its victims
in other ways as well, for example, by profoundly disrupting the development or
sustenance of their sense of personal worth. Among the components of self-worth
that can be obstructed by childhood deprivation is the sense of one's worthiness
to answer for one's actions, a normally important part of one's sense of equal moral
standing as a person and thus of one's relation to the rest of the moral community.
In this respect, the victim of a severely deprived upbringing may be similar to the
person confined to the invidious social hierarchy in the previous example; both may
have good reason to feel that it is not their place to take responsibility for their ac-
tions by answering for themselves. The common (albeit not uncontroversial) view
that the adolescent's traumatic childhood would lessen her responsibility for some of
her actions, even if she possesses basic moral competence, can be upheld if she has
a seriously impaired sense of her worth and if possessing a more robust sense of
worth is necessary for being fully responsible.36
My argument for admitting a self-worth condition of responsibility respects
more fully the relational character of holding persons responsible than do the three
accounts of responsibility discussed in the previous section. We can better under-
stand this feature of my position and clarify its value for feminist projects in ethics,
as well as mark the comparative shortcomings of the other three accounts, by noting
three overlapping social dimensions of responsibility that the self-worth condition
brings clearly to view. First, since the sense of worth required for full responsibility
concerns one's worthiness to give account of one's actions to others, this condition
serves to remind us that responsibility depends conceptually on there being others
who could morally criticize us and expect us to answer for our actions. Thus it
makes sense, as Wallace realizes, to approach responsibility by analyzing being held
responsible by others?7 This does not mean that we are responsible only in the actual
presence of others who might respond to our actions and to whom we could give an
account of ourselves. The point is just that the conceptual possibility of relating to
82 Autonomy and the Social

others in certain ways is necessary for responsibility, even in the remote instance
when such relations are not physically possible.
This first relational aspect of responsibility points beyond the purely concep-
tual, for it is also a developmental fact about us that only through being expected by
others to speak for our actions do we become responsible agents and acquire the req-
uisite sense of our worth. I also suspect that it is only through being held responsible
by others that we can come to hold ourselves responsible and learn how to answer to
ourselves since being held responsible by others seems to be a necessary step in learn-
ing how to hold anyone responsible and becoming susceptible in appropriate ways
to the range of reactive emotions.38
The conceptual and developmental primacy of being expected to account for
one's actions to others affords a theoretical framework that complements feminist
theories of the narrative construction—and reconstruction—of personal identity.
For example, as Susan Brison explains in her compelling discussion of the remaking
of selfhood in the wake of trauma, "In order to construct self-narratives, then, we
need not only the words with which to tell our stories but also an audience able and
willing to hear us and to understand our words as we intend them."39 Of course, in
the case of trauma survivors, this dependence on an audience is bound to be actual
and concrete, not merely conceptual.40 Nonetheless, recognizing in a theory of
responsibility the need for at least a potential audience to whom agents are expected
to give an account of themselves reveals a natural point of commonality between
narrative theories of selfhood and my view of responsibility.
This first relational feature of responsibility also helps to explain the evident
empirical fact that the relevant sense of worth typically is highly sensitive to the at-
titudes that others display toward the agent.41 If responsible agents' sense of self-
worth depends on (potentially) being expected to make sense of their actions to oth-
ers, it is understandable that this sense of worth would normally be sensitive to
others' expressed attitudes toward the agent's worth. This point coheres with femi-
nist insistence that an adequate view of moral agency must attend to the interper-
sonal context within which the agent is situated. Such attention is necessary not just
to characterize the choices the agent faces but also to reveal the role that others' at-
titudes may have in shaping the person's own sense of her status as an accountable
agent.
A second relational aspect of responsibility manifested in the self-worth condi-
tion is that publicly shareable norms must regulate both moral appraisal of an agent's
actions and the account of her conduct that she might give in response. Agents can
be in a position to respond to others' moral criticisms of them only if all parties to
this possible dialogue can share an understanding of the norms reflected in these
criticisms and the norms according to which responses can properly be made. The
social intelligibility of norms of moral appraisal and response does not require that
all parties must actually understand these norms equally well. Nor does it require
that any responsible agent must accept these norms or accept them for the same rea-
sons that others do.
Although it seems obvious that moral responsibility is possible only if everyone
to whom moral norms of criticism, justification, and excuse apply can understand
them, the significance of this second social feature of responsibility has not been
 Self-Worth 6- the Social Character of Responsibility 83

well appreciated. For instance, it highlights another respect in which justified attri-
butions of responsibility depend on the context of the social relationships within
which they are made. If someone is an intimate friend of mine, as opposed to being
just a familiar face in the grocery store or a total stranger at the bus station, what she
can understand me to hold her accountable for and what kinds of justification or ex-
cuse she can expect me to accept may differ considerably from these other cases. So-
cial relationships modify the kinds of moral dialogue that are possible among people
and, in doing so, influence the terms of moral responsibility that are possible be-
tween them.42 When persons are subjected to relations of subordination and domi-
nation, the public intelligibility of norms of criticism and response is likely to be
undermined, yielding contested attributions of responsibility. This is especially the
case when subordination is effected precisely by holding persons to conflicting ex-
pectations, ensnaring them in moral double binds that make it impossible for them
to reach a coherent understanding of how they could ever adequately defend or ex-
cuse their own conduct.43
A third relational aspect of responsibility illuminated by the self-worth condi-
tion is that being responsible is itself a matter of occupying a certain social position,
having the status of an eligible participant in a community of moral dialogue. Some-
one who regards herself as being in a position to speak for her own agency, should
others criticize it, and who also has the competence necessary for answering for her
actions (powers that include reflective self-control) is, then, in just such a position.
Whether others recognize it or not, she is in a position to take full part in a moral di-
alogue with others who can properly hold her responsible and whom she can hold
responsible. This point is particularly relevant to feminist ethics in two respects. It
shows why feminists, who are necessarily concerned about ways in which women are
denied the status of full social actors, also have reason to care deeply about whether
or not women are held fully responsible for their actions. Exempting women from
responsibility can in some cases serve sexist purposes by not respecting their status as
moral interlocutors.44 This third feature also brings to view a route by which femi-
nist accounts of the mechanisms used to deny or obscure women's rightful social sta-
tus can be extended to instruct us about the nature of responsibility. I have begun
to explore one branch of that route in the case of Charlottes socially manufactured
craziness.
Fischer and Ravizza's reasons-responsiveness account will allow social aspects of
a person's deliberations or circumstances to affect responsibility only to the extent
that those aspects influence her actual causal control over her behavior. This account
seems, at first, quite alien to many feminist interests in the relational dimensions of
women's moral selfhood. For instance, since it focuses on the counterfactual respon-
siveness of action-mechanisms rather than on the capabilities of agents (as Wallace's
view does), it can appear to neglect the integrity of individual agents considered as
wholes, not to mention the character of relationships among persons. However, a
reasons-responsiveness condition can apprehend some social features of responsibil-
ity that would be important for feminist ethics. Such a condition remains neutral
about the source and nature of reasons to act, thereby allowing action generated
through mechanisms that are responsive to emotional considerations or considera-
tions grounded in physical need to count as responsible. By refusing to accord priv-
84 Autonomy and the Social

ilege to modes of rational consideration that are emotionally cold, abstract, or im-
personal, the account makes room for rational responsiveness to the sorts of affec-
tively colored states that figure prominently in interpersonal attachments.
More important, Fischer and Ravizza's view is essentially historical. It ties re-
sponsibility to features of the process by which an action is generated, not simply to
features realized in the time slice during which the action is performed. 45 Taking se-
riously the history of actions permits the theory to be sensitive to elements of con-
text and narrative development, thus making the account conducive to feminist
elaboration.
Nevertheless, the fact that the reasons-responsiveness condition is rooted in fea-
tures of action-mechanisms that need not reflect the perspectives of agents them-
selves sharply limits the theory's serviceability for feminist ethics. As we saw in Char-
lotte's case, the way an agent regards herself as a moral agent can jeopardize her
responsibility even when the mechanism through which she acts remains counter-
factually sensitive to possible reasons to act otherwise. Moreover, Fischer and Rav-
izza examine the nature of an agent's control in conceptual isolation from the nexus
of relationships within which personal control becomes relevant to moral responsi-
bility. Control matters for responsibility because various degrees of control are re-
quired within the complex form of moral dialogue that defines what it is to be re-
sponsible. The reasons-responsiveness condition cannot fully reflect the normative
demands of such dialogue, as I have argued. Attending to agents' sense of worth, by
contrast, helps to advance feminist commitments to take women's experiences and
perspectives seriously. Of course, it does not follow that agents' feelings or beliefs
about their own responsibility are incorrigible.46
Wallace's account of the conditions of responsibility does much better than Fi-
scher and Ravizza's in recognizing the level at which social considerations contribute
to responsibility. For Wallace, having the powers of reflective self-control makes one
accountable because one is appropriately held responsible when one possesses those
powers (so long as one has no excuses in cases of wrongful behavior). Wallace con-
tends, as I do, that holding someone responsible is a matter of opening the possibil-
ity for a certain form of moral relationship with her. This means that relational con-
siderations operate not merely as incidental causal influences on responsibility; social
relationships also shape responsibility more profoundly.
The main hazards to accountability may be posed not by our physical and biolog-
ical nature but by the social and political circumstances in which we develop and
live. The conditions of responsibility I have identified . . . describe tin ideal that is
regulative of our social interactions, the ideal of a community of people capable
of participating constructively with each other in the exchange of moral criticism
and justification. No doubt many of the people we interact with conform to this
ideal sufficiently to make it fair to hold them fully accountable. . . . But approxi-
mation to the ideal is a matter of degree, and it is liable to be affected by such
common phenomena as childhood abuse, psychological trauma, and the persist-
ence of extreme poverty and violence in the midst of general affluence.47
Yet, like Fischer and Ravizza, Wallace ignores considerations having to do with
persons' experience of their own agency.48 He approaches responsibility primarily
 Self-Worth & the Social Character of Responsibility 85

from the standpoint of evaluation, the standpoint of holding persons responsible

(including oneself). He considers aspects of the agent's perspective on her own con-
duct only to the extent that they pertain to conditions framed from the evaluator's
standpoint. However, as I have contended throughout this discussion, failure to re-
spect the agent's perspective fully, particularly with regard to her sense of worthiness
to speak for her own conduct, occasions failure to understand the full effects of op-
pressive social conditions on responsibility and, accordingly, occasions failure to
grasp how deeply the relational aspects of responsibility run. The social relations
that affect responsibility are not only those that modify what we are capable of doing
but also those that mold who we regard ourselves as being.
Thus, the two criticisms I set before Fischer and Ravizza's reasons-responsive-
ness approach—that it focuses on control, to the neglect of persons' perspectives on
their own agency, and that it attempts to detach a conception of control from the
context of interpersonal moral relations that determines what responsibility is—are
not entirely independent of each other. Examining the relational context that de-
fines responsibility leads us to take seriously the agent's perspective on the exchange
of moral criticism and response. Conversely, taking seriously the agent's perspective
enables us to understand more fully the social character of responsibility.
Of the three accounts that I have considered in addition to my own proposal,
the self-disclosure view comes the closest to reflecting the three intrinsically rela-
tional aspects of responsibility that I have noted. This is not surprising since the no-
tion of self-disclosure naturally points toward the context of relationships within
which what an agent really cares about can be disclosed to others through her con-
duct.49 And, as we saw in the second section, a self-disclosure view of responsibility
could recognize the effect that diminished self-worth has on an agent's responsibil-
ity by arguing that adequate self-disclosure is possible only when an agent regards
herself as worthy of expressing through her actions what she values. Hence, if I am
correct that a theory's sensitivity to the relational dimensions of responsibility makes
it particularly valuable for feminist purposes, a self-disclosure account of responsi-
bility warrants much more feminist interest than it has received.
Nevertheless, it remains the case that a self-disclosure approach has not in fact
been developed in ways that explicitly draw out its relational implications. My dis-
cussion of the approach is (to my knowledge) the first to point out these implica-
tions and to consider the possible role of a socially articulated sense of self-worth in
self-disclosure. I suspect that this type of account has been insufficiently developed
along these lines because it was initially formulated by thinking about the internal
psychological conditions under which an agent's motives constitute her real wants
(as in Wolf's formulation of the Real Self View), instead of by conceiving self-
disclosure as a matter of entering into certain forms of moral relationship with oth-
ers who can appropriately expect the agent to answer for the features of her moral
self that her conduct may disclose.

Some Worries
I have mainly been concerned in this article to show three things. First, I have ar-
gued that three influential accounts of responsibility have failed to appreciate ade-
86 Autonomy and the Social

quately the role of persons' own perspectives on their agency. Taking those perspec-
tives seriously gives rise, I have proposed, to an unorthodox self-worth condition of
full responsibility.50 Second, I have used the self-worth condition of responsibility to
show that responsibility is sensitive to social relations in ways that other theories
have obscured. Finally, I have suggested that considering the intrinsically relational
aspects of responsibility that 1 have identified will help to advance feminist reflec-
tion about the significance of responsibility for liberatory approaches to ethics.
In closing, I want to discuss briefly some likely worries about the proposed con-
dition of self-worth. First, some may worry that because it can be difficult to gather
adequate evidence about whether persons have the requisite sense of worth, the pro-
posed view would make it very difficult to assess persons' responsibility. I admit this
but deny that it is a good reason to reject the self-worth condition. It is a feature of
many ordinary conditions of responsibility that their application can be extremely
difficult to verify, which is one reason why ascriptions of responsibility are so com-
monly contested. It would be unreasonable to demand that an adequate account of
responsibility make it easier than it actually is to determine whether, or to what de-
gree, a person is responsible for some action.
A second, related worry is that requiring a sense of self-worth for unqualified
responsibility would make it too easy to evade responsibility. Persons could plead
reduced accountability simply by denying that they felt worthy of answering for
their acts or, more extremely, by intentionally trying to sabotage their sense of worth.
I do not see why my proposal would have this consequence. For one thing, whereas
it is a simple matter to say that one does not have the requisite sense of worthiness,
what one says by no means settles the question of whether one actually has it or
not. Compare this with the ordinary condition of control. It is difficult to verify
the extent of an agent's control over her actions, and it is easy to deny that one has
much control. But this does not threaten the plausibility of requiring some sort of
control as a condition of responsibility. Furthermore, it would be very difficult to
destroy deliberately one's sense of worthiness to account for one's conduct with-
out damaging many other valuable aspects of one's moral agency, too. Even if one
could do so, it would be an irrational strategy for evading responsibility. An agent
who deliberately destroyed her sense of self-worth would normally be responsible
for that very act.
The proposed self-worth condition could also seem to permit unwarranted eva-
sions of responsibility because many agents who suffer serious blows to their self-
worth, through no fault of their own, do not appear to be less responsible as a result.
For instance, persons who were victims of abuse as children may become abusers as
adults, in part because of the dramatic loss of self-worth brought on by their early
experiences of abuse. Yet the fact that these abusers often feel relatively worthless as
persons does not seem automatically to lessen their responsibility for inflicting abuse
on others. 51
Such cases do not pose a genuine difficulty for my posirion, however. I have
made a claim only about the role that one particular component of our sense of
moral self-worth, namely, worthiness to give account of our actions, plays in our re-
sponsibility. There are many elements of our general sense of personal worth or es-
 Self-Worth & the Social Character of Responsibility 87

teem that are independent of this specific component of our self-regard.52 The ex-
ample of abuse victims who become abusers illustrates this. The self-contempt these
abusers are likely to suffer does not normally entail reluctance to give an account of
their actions. Abusers are, if anything, all too ready to advance justifications for
their violence (justifications that in fact often reflect indirectly their damaged self-
esteem and repressed rage). This is entirely different from persons like Charlotte,
whose socially inflicted self-doubt leads them to feel unworthy of speaking for them-
selves as morally answerable agents. Of course, there are other live questions about
the responsibility of victims of abuse (especially concerning the level of self-control
that it is fair to expect of them), but there is no general problem here for my pro-
posed condition of responsibility.
Third, one might object that my analysis of Charlotte's predicament compli-
cates matters unnecessarily. Would it not be simpler to argue that Charlotte's re-
sponsibility is diminished only to the extent that she has been subjected to coercive
or manipulative social influences? If this were the case, impairment of her responsi-
bility could be explained entirely in terms of disruptions of her ability to act freely,
without reference to her attitudes toward her worth. This might be adequate when
external threats or deceits force persons to conform to oppressive social practices.
But Charlotte does not act in conformity with social expectations, and in any case,
she has internalized many of the prevailing expectations that regulate women's judg-
ments and behavior (in her day). She views many of the social influences that oper-
ate on her as providing good evidence about how she should feel about herself. This
is precisely why she is so profoundly conflicted. Analyzing her responsibility solely in
terms of socially instituted threats and deception would radically oversimplify her
predicament. It is precisely because she does act through reasons-responsive pro-
cesses and can act on her own judgments despite her "craziness" that her situation
calls for more subtle analysis.
One might worry, alternatively, that diminished self-worth does compromise
moral agency, but not by reducing responsibility. Perhaps diminished self-worth just
makes it more difficult to discern what is morally acceptable and translate it into ac-
tion, so that we should not hold persons afflicted by this sense of unworthiness to
the same standards that we would expect others to meet. But this would not involve
reduced responsibility. We should still hold such persons accountable in relation to
the revised expectations.
There is much to agree with in this objection. It develops the point (noted in
the second section) that an agent's sense of worth could interfere with her capacities
of reflective self-control. However, this does not deflect my arguments that self-
worth may influence responsibility through other routes, too. It also underestimates
the seriousness of a plight like Charlotte's to think only that it is just a bit more dif-
ficult for her to know what to do than it would be for her doctor or husband. This
fails to appreciate how her self-doubt can change her relationship to the entire
process of deciding what to do in the face of others' expectations.
Finally, I should note the more sweeping concern that making self-worth a nec-
essary condition of full responsibility seems to be at odds with the political purposes
of feminism. The worry is that many women may feel unworthy of giving accounts
88 Autonomy and the Social

of their actions on equal terms with men, and so the proposed condition would, in
effect, condemn them to the position of pure victims, rendering them demoralized
and frustrating the possibility that they could take responsibility for developing re-
sistance to the oppressions leveled against them. In other words, if my proposal makes
full responsibility too difficult to attain under oppressive social conditions, wide-
spread acceptance of the view could disempower many women and make the
project of building women's collective and individual resistance just that much more
difficult.
I can only give a partial response here to a concern that really calls for a system-
atic account of how moral agency can be reconstructed from within oppressive con-
texts. Many others have offered significant proposals about whether attributions of
responsibility should be given an important role in fighting women's demoraliza-
tion.53 I restrict my comments to three points, which indicate that women who feel
unworthy to answer for their actions need not be pure victims.
First, as I noted in the previous section, the sense of self-worth that pertains to
responsibility is sensitive to social context. One's sense of worthiness to account for
one's actions can reasonably vary from one domain of normative expectations to an-
other, and so can vary from one network of relationships to another. Persons' sense
of their moral standing can vary as the context for action shifts from family rela-
tionships to the social demands of citizenship, and so forth. Thus, diminished re-
sponsibility that results from a depressed sense of worth in one domain of rela-
tionships, relative to one cluster of ethical expectations, does not necessarily entail
diminished responsibility in every area of activity, for all moral choices. Since op-
pressions typically burden some areas of activity more than others, their effects on
persons' responsibility are far from uniform. Persons who feel hopelessly victimized
in some respects can often find resources for reaffirming their self-worth in other
contexts, where their status as accountable agents receives greater social recognition.
A powerful illustration is the way in which patriarchal expectations that assigned
black women slaves in the United States virtually total responsibility for managing
slaves' domestic needs thereby provided a meaningful context of accountability, self-
worth, and power from which slave women could criticize and resist institutional-
ized slavery.54
Furthermore, as Charlotte's case shows, if self-worth can modify responsibility
while the agent's powers of reflection and self-control remain intact (though under-
used), it is hardly the case that diminished self-worth renders such a person a com-
pletely disempowered victim. On the contrary, if the agent retains her powers to rec-
ognize, apply, and act on moral considerations, she has tools that are essential for
analyzing and, ultimately, dismantling social systems that subordinate her. True,
these tools will be of little use to her until she can gain adequate trust in herself to
employ them. But it is hardly the case that a person in Charlotte's predicament has
to rebuild her capacities for moral agency from scratch.
Also, the task of reclaiming and sustaining the relevant sense of worthiness
from within oppressive conditions is far from impossible, however difficult it may
be. As I have noted, our sense of our fitness to give accounts for what we do is
highly sensitive to social recognition. Just as venomous, degrading, inferiorizing
social relations can break down our sense of our fitness to speak for ourselves as
 Self-Worth & the Social Character of Responsibility 89

moral agents, so also caring, dignifying, democratizing social relations can repair
damaged self-worth. Taken together with the preceding points about the context-
dependence of self-worth and the compatibility of normative competence with loss
of self-worth, this suggests that means for recovering a sense of accountability are
normally available even for those who appear hopelessly victimized.

Notes

1. I am grateful to the editors of this volume for making numerous helpful comments
on an earlier draft of this article. Whereas the content of the position I have defended is my
responsibility alone, whatever clarity I have managed to achieve in explaining it owes a good
deal to the editors' assistance.
2. This is compatible with Catharine MacKinnon's view that consciousness raising has
not only been a useful resource for feminists but also the distinctive methodology of femi-
nism. According to MacKinnon, "The key to feminist theory consists in its way of knowing.
Consciousness raising is that way." Catharine A. MacKinnon, Toward a Feminist Theory of the
State (Cambridge, Mass.: Harvard University Press, 1989), p. 84 (emphasis in original).
3. Virginia Held, Feminist Morality (Chicago: University of Chicago Press, 1993), pp.
14-15.
4. Margaret Urban Walker maintains, however, that a feminist ethic of responsibility is
more properly understood as a broad family of views, including but not restricted to an ethic
of care. See Walker's "Picking Up Pieces: Lives, Stories, and Integrity,' in Feminists Rethink the
Self ed. DianaTietjens Meyers (Boulder, Colo.: Westview Press, 1997), pp. 64-65.
5. For representative feminist criticisms of the abstract individualism that has been as-
sociated with personal responsibility, see Alison M. Jaggar, Feminist Politics and Human Na-
ture (Totowa, N.J.: Rowman & Allanheld, 1983), especially pp. 39-47, and Lorraine Code,
What Can She Know?: Feminist Theory and the Construction of Knowledge (Ithaca, N.Y.: Cor-
nell University Press, 1991), pp. 71-79.
6. See, for instance, Maria C. Lugones, "Playfulness, 'World'-traveling, and Loving Per-
ception," Hypatia 2 (1987): 3-19; Lugones, "On the Logic of Pluralist Feminism," in Fem-
inist Ethics, ed. Claudia Card (Lawrence: University Press of Kansas, 1991); Cheshire Cal-
houn, "Responsibility and Reproach," Ethics 99 (1989): 389-406; Sandra Lee Bartky,
"Shame and Gender," in Femininity and Domination (New York: Routledge, 1990); Susan
Wendell, "Oppression and Victimization: Choice and Responsibility," Hypatia 5 (1990):
15-46; Elizabeth V. Spelman, "The Virtue of Feeling and the Feeling of Virtue," in Card,
Feminist Ethics; and Anita M. Superson, "Right-Wing Women: Causes, Choices, and Blam-
ing the Victim," in "Nagging" Questions: Feminist Ethics in Everyday Life, ed. Dana E. Bush-
nell (Lanham, Md.: Rowman & Littlefield, 1995). For a thoroughgoing critique of blame as
demoralizing in the context of lesbian community, see Sarah Lucia Hoagland, Lesbian Ethics:
Toward New Value (Palo Alto, Cal.: Institute of Lesbian Studies, 1988), pp. 215-221. Bar-
bara Houston, "In Praise of Blame," Hypatia 7 (1992): 128-147, offers an opposing view.
While urging "a feminist transformation of blame" (p. 142), Houston nevertheless affirms
blame as necessary for developing women's moral agency, integrity, and sense of self in com-
munity (pp. 140-142).
7. Hoagland, Lesbian Ethics, p. 215.
8. Alison M. Jaggar, "Feminist Ethics: Projects, Problems, Prospects" in Card, Feminist
Ethics, p. 90.
9. I name her for feminist author Charlotte Perkins Gilman, whose life is reflected in
90 Autonomy and the Social

many (though not all) features of this case. I thank John Christman for reminding me
about Oilman's short story "The Yellow Wallpaper" in conjunction with my work on free
agency.
10. Oilman's neurologist, S. Weir Mitchell, used this term to describe the process by
which his female patients would gradually turn to him for moral guidance. Mitchell wrote,
"If you tell the patient she is basely selfish she is probably amazed and wonders at your cru-
elty. 'Jo cure the case you must morally alter as well as physically amend, and nothing else will
answer." See Ann J. Lane, To "Herland" and Beyond: The Life and Work of Charlotte Perkins
Oilman (New York: Pantheon Books, 1990), p. 117.
11. I regard the sense of one's worthiness to answer for one's conduct as one component
of one's respect for one's own personhood, one's "recognition self-respect," as Stephen Dar-
wall terms it in "Two Kinds of Respect," Ethics 88 (1977): 36-49. Lacking recognition of
self-respect does not entail that one fails to grasp one's worthiness to answer for one's con-
duct, however. And whereas failing to appreciate one's worthiness to give a moral account of
oneself will stand in the way of fully recognizing one's moral worth, it would be possible for
one to retain some appreciation of one's worth as a person nevertheless. For some helpful dis-
tinctions among different types and elements of self-respect and a survey of numerous ways
in which self-respect can be degraded, see Robin Dillon, "How to Lose Your Self-respect,"
American Philosophical Quarterly'29 (1992): 125-139.
12. R. Jay Wallace, Responsibility and the Moral Sentiments (Cambridge, Mass.: Harvard
University Press, 1994). I will not discuss theories that place incompatibilist requirements on
responsibility since 1 do not think that debates about the metaphysical compatibility of re-
sponsibility and deterministic explanations of human action have much to contribute to
feminist ethics.
13. Ibid., p. 157.
14. Ibid.
15. Ibid., p. 158.
16. Compare ibid., pp. 161, 183-186.
17. Compare John Christmans discussion of the ways in which obstacles to self-aware-
ness and inconsistent beliefs and desires may block an agent's autonomy. See "Autonomy and
Personal History," Canadian Journal of Philosophy 21 (1991): 1-24, especially 13-18.
18. I have discussed this mode of alienation from reflective powers in "Free Agency and
Self- Worth," Journal of Philosophy 91 (1994): 650-668, especially 657-659.
19. John Martin Fischer, "Responsiveness and Moral Responsibility," in Responsibility,
Character, and the Emotions, ed. Ferdinand Schoeman (Cambridge: Cambridge University
Press, 1987); John Martin Fischer and Mark Ravizza, "Responsibility and Inevitability,"
Ethics 101 (1991): 258-278; John Martin Fischer and Mark Ravizza, "Responsibility for
Consequences," in Perspectives on Moral Responsibility, ed. Fischer and Ravizza (Ithaca, N.Y.:
Cornell University Press, 1993); John Martin Fischer, The Metaphysics of Free "^'//(Cam-
bridge: Blackwell, 1994). (Fischer and Ravizza's most recent work, Responsibility and Con-
trol: A Theory of Moral Responsibility [Cambridge: Cambridge University Press, 19981, ap-
peared after this article was written.)
20. Fischer and Ravizza distinguish weak from strong reasons-responsiveness, where the
latter requires responsiveness of the act-mechanism to any possible sufficient reason to act
otherwise. See Fischer, "Responsiveness and Moral Responsibility," pp. 86-90; Fischer and
Ravizza, "Responsibility and Inevitability," pp. 269--270; and Fischer, The Metaphysics of
Free Will, pp. 166-168.
21. Fischer and Ravizza are primarily interested in the "freedom-relevant" component
of responsibility, that is, in the sort of control that responsibility requires. They discuss the
epistemic requirements of responsibility only briefly. See Fischer, "Responsiveness and Moral
 Self- Worth 6" the Social Character of Responsibility 91

Responsibility," p. 267, n. 10; Fischer and Ravizza, "Responsibility for Consequences," p. 338,
n. 22; and Fischer, Metaphysics of Free Will, p. 238, n. 4.
22. Susan Wolf, Freedom within Reason (New York: Oxford University Press, 1990),
chap. 2. Wolf does not endorse the Real Self View. She argues that it provides only a neces-
sary condition of responsibility, not a sufficient condition, since it does not entail the ability
to recognize and appreciate the right reasons for action.
23. Ibid., p. 33.
24. However, Watson has backed away from his early account of the role of valua-
tional systems in free and responsible agency. For that account, see "Free Agency," journal
of Philosophy 72 (1975): 205-220. He has argued more recently that actions for which a
person is responsible can disclose her values without necessarily reflecting her judgments
about what is valuable or worthwhile and without endorsing those values from some general
evaluational standpoint. See "Free Action and Free Will," Mind% (1987): 145-172, espe-
cially 149-150.
25. Gary Watson, "Responsibility and Normative Competence," unpublished paper,
delivered at the Pacific Division Meetings, American Philosophical Association, March 1992,
p. 11. John Dewey's statement is from his Outlines of a. Critical Theory of Ethics (New York:
Hillary House, 1957), p. 161. A revised version of Watsons paper appeared as "Two Faces
of Responsibility," Philosophical Topics 24 (1996): 227-248.
26. The self-disclosure account will grant, of course, that features of Charlotte's cir-
cumstances, such as the extreme psychological and social pressures enacted in the rest cure,
are relevant to a sound interpretation of what her actions reveal about what matters to her.
This view does not take the simplistic position that what persons care about can be directly
read from the actions for which they are responsible, independently of the circumstances and
history of those actions.
27. This assumes that self-doubt does not alienate Charlotte from her reflective capac-
ities, a possibility I considered when discussing Wallace's account. If such alienation were to
occur, the value judgments that Charlotte would arrive at through reflection would not nec-
essarily express what really matters to her. In that case, it is hard to see how Charlotte could
be capable of governing her actions through her valuational system and thereby disclosing
what she really cares about through her actions.
28. For an extended treatment of the conceptual relationship between the appropriate-
ness of holding persons morally responsible and their actually being responsible, see Wallace,
Responsibility and the Moral Sentiments, chap. 4. Wallace argues that a person is morally re-
sponsible for an action just in case it would be appropriate to hold her morally responsible for
it, where the relevant standards of appropriateness are moral norms of fairness.
29. I am indebted here to T. M. Scanlon, Jr., "The Significance of Choice," in The Tan-
ner Lectures on Human Values, vol. 8, ed. Sterling M. McMurrin (Salt Lake City: University of
Utah Press, 1988), especially pp. 166-176. Scanlon emphasizes connections between hold-
ing a person morally responsible and the person's inter- and intrapersonal responsiveness.
However, I do not endorse Scanlon's thesis that the forms of responsiveness necessary for re-
sponsibility are dictated by the content of moral judgments alone. For a valuable discussion
of Scanlon's account, see Wallace, Responsibility and the Moral Sentiments, pp. 74- 81.
30. See Gary Watson, "Responsibility and the Limits of Evil," in Schoeman, Responsi-
bility, Character, and the Emotions. As Jay Wallace reminds us, however, this does not mean
that holding someone accountable involves moral communication in each instance or always
requires an interest in having a moral exchange with the person (see Responsibility and the
Moral Sentiments, pp. 164-165).
31. I am not suggesting that blame, whether expressed or unexpressed, is strictly neces-
sary for holding someone responsible. I focus on cases of holding persons responsible for per-
92 Autonomy and the Social

ceived wrongs, hence on negative reactions such as blame or indignation, because these cases
best display the relational aspects of responsibility in which I am interested here. However, I
readily admit that philosophers would do well to spend more time reflecting on the role of
positive reactive attitudes, such as admiration, pride, or praise.
32. Sec Watson, "Responsibility and the Limits of Evil," pp. 264-265; Scanlon, "The
Significance of Choice," pp. 171 -172; and Wallace, Responsibility and the Moral Sentiments,
pp. 164-165.
33. I assume that Charlotte is not to blame for feeling unworthy of giving an account
of her actions because of the oppressive circumstances she faces. In such circumstances, a
person could not reasonably be expected to feel any differently about herself. A person whose
sense of self-worth erodes because of some character flaw under more favorable conditions
would be in a very different position, however.
34. I explore the importance of this point more fully in the last section.
35. In Freedom within Reason, chap. 4, Susan Wolf suggests that traumatic childhoods
may reduce responsibility primarily by impeding the development of adequate abilities to
recognize certain kinds of good reasons for acting. Wallace proposes instead that many kinds
of childhood deprivation reduce accountability by affecting the capacity to control behavior
(see Responsibility and the Moral Sentiments, pp. 231-233).
36. However, as I underscore in the last section, not just any sort of diminished self-
worth that is associated with traumatic upbringings will reduce responsibility, in my view.
The childhood trauma must specifically modify the sense of worthiness to give an account
of one's actions in order to engage the proposed condition of responsibility.
37. 1 argue shortly, however, that Wallace goes too far in the direction of analyzing re-
sponsibility from the third-person perspective of appropriately holding others responsible.
He omits due consideration of the agent's own perspective on her agency.
38. As Galen Strawson notes, admitting these developmental facts is consistent with
holding that "the true centre of one's commitment to the notion of human freedom [and re-
sponsibility] lies in one's attitude to and experience of oneself." See Freedom and Belief (New
York: Oxford University Press, 1986), p. 110.
39. Susan J. Brison, "Outliving Oneself: Trauma, Memory, and Personal Identity," in
Meyers, Feminists Rethink the Self, p. 21.
40. Ibid., p. 27.
41. This point is also important for understanding the political possibilities for restor-
ing accountability in oppressive contexts (sec the final section).
42. Christine Korsgaard's work on a Kantian account of responsibility suggests a differ-
ent route by which to arrive at the idea that responsibility is properly relative to interpersonal
relationships. See Christine M. Korsgaard, "Creating the Kingdom of Ends," in Creating the
Kingdom of Ends (Cambridge: Cambridge University Press, 1996). Korsgaard maintains that
responsibility should be conceived practically, as a matter to be decided, not a fact about per-
sons to be discovered. To hold someone responsible, in this view, is to decide to place oneself
in a relationship of reciprocity with the person (or to affirm the reciprocal relations one al-
ready has with her). Hence, Korsgaard reasons, "It may be perfectly reasonable for me to hold
someone responsible for an attitude or an action, while at the same time acknowledging that
it is just as reasonable for someone else not to hold the same person responsible for the very
same attitude or action" (ibid., p. 199). Whereas Korsgaard sees responsibility as relative
to relationships because of the decision to enter or continue relations of the sort in which
responsibility attributions make sense, my claim is that relationships can modify facts
about responsibility in virtue of affecting the shareability of particular norms of criticism
and response.
43. See Kathryn Pauly Morgan, "Women and Moral Madness," in Feminist Perspectives,
 Self- Worth & the Social Character of Responsibility 93

ed. Lorraine Code et al. (Toronto: University of Toronto Press, 1988), for a classic discussion
of how patriarchy traps women in unavoidable moral paradoxes and thereby deepens their
demoralization through feelings of "moral madness."
44. Cf. Houston, "In Praise of Blame."
45. Fischer and Ravizza argue that this dependence on historical features is not only
epistemic but also metaphysical. Their condition of responsibility is, therefore, genuinely his-
torical, not just apparently so. See John Martin Fischer and Mark Ravizza, "Responsibility
and History," in Midwest Studies in Philosophy: Philosophical Naturalism, ed. Peter A. French
(Notre Dame, Ind.: University of Notre Dame Press, 1994).
46. Here I disagree with a related claim that Jennifer Nedelsky makes about autonomy
in "Reconceiving Autonomy: Sources, Thoughts and Possibilities," Yale Journal of Law and
Feminism 1 (1989): 7-36. Nedelsky correctly notes that theories of autonomy typically have
ignored persons' feelings of autonomy. But she maintains, further, that the feeling of auton-
omy is "an inseparable component" of the capacity of autonomy (p. 25) and that someone
who feels autonomous (when autonomy is properly reconceived) must actually be autono-
mous (p. 24). For reasons implied in my discussion of Charlotte's capabilities (second sec-
tion), feelings of autonomy are neither necessary nor sufficient for the possession of capaci-
ties of autonomy. The important goal of granting greater authority to such feelings can be
met without making them immune to possible error.
47. Wallace, Responsibility and the Moral Sentiments, p. 234.
48. This is clear from Wallace's discussion of childhood deprivation. Ibid., pp. 231-233.
49. The fact that Watson is interested in both a self-disclosure analysis of responsibil-
ity (in "Responsibility and Normative Competence" and "Two Faces of Responsibility") and
a Strawsonian view that concentrates on the conditions of intelligible moral address (in "Re-
sponsibility and the Limits of Evil") is a further indication that the former type of account
may be capable of recognizing the importance of relational considerations.
50. Of course, giving due attention to persons' sense of who they are as moral agents
will be likely to do more than simply reveal the place of their sense of worthiness to answer
for their actions in their responsibility. 1 have focused on this particular element of agents'
standpoints because it is especially useful for bringing to view relational aspects of responsi-
bility that are significant for many feminist purposes.
51. I thank Catriona Mackenzie and Natalie Stoljar for urging me to address this
problem.
52. Compare note 11 on the relation of this specific component of self-worth to recog-
nition self-respect.
53. See the references in note 6.
54. See Angela Davis, "Reflections on the Black Woman's Role in the Community of
Slaves," The Black Scholar 3 (1971): 3-15.
 4

AUTONOMY AND THE FEMINIST

INTUITION1

Natalie Stoljar

Introduction
Feminist philosophers rightly have been suspicious of old-fashioned theories of au-
tonomy, in which it was equated with the masculinist ideals of substantive inde-
pendence and self-sufficiency. Lately, however, a different set of theories of auton-
omy has become dominant, namely, procedural theories. Many such theories argue
that it is the capacity for procedural, independence or independence of mind, rather
than the capacity for substantive independence, that is necessary and sufficient for
autonomy.2 In such accounts, an agent's preference or decision is autonomous if and
only if the process of formation of the preference or decision has satisfied certain
standards of critical reflection. Once a preference or decision has passed such formal
or procedural tests, it is autonomous, no matter what its content. Hence, procedural
theories arc "content-neutral."3
The implication of content neutrality, namely, that there is no restriction on the
content of preferences that may be chosen by autonomous agents, is an advantage
from some feminist perspectives, especially when contrasted with the idea that au-
tonomy presupposes preferences for substantive independence or self-sufficiency.
For instance, for feminists inspired by the care perspective of Carol Gilligan, it is im-
portant that autonomy be compatible with maintaining the relations of care and de-

94
 Autonomy and the Feminist Intuition 95

pendence that are characterized as valuable components of female agency. Jennifer

Nedelsky writes that "the most promising model, symbol or metaphor for autonomy
is child rearing. There we have encapsulated the emergence of autonomy through
relationship with others."4 Unlike self-sufficiency accounts of autonomy, procedural
theories allow that preferences for dependence and connection can be autonomous.
Moreover, as Marilyn Friedman points out, procedural theories are compatible with
another characteristic of autonomy noticed by Nedelsky, who says, "If we ask our-
selves what actually enables people to be autonomous, the answer is not isolation,
but relationships—with parents, teachers, friends, loved ones—that provide the
support and guidance necessary for the development and experience of autonomy."5
Finally, content neutrality allows different agents autonomously to opt for a range of
different life plans and conceptions of the good, thus respecting recent feminist con-
cerns to preserve the diversity among and the multiplicity of agents. Procedural the-
ories, therefore, are congenial to a feminist emphasis on the relational features of per-
sons in three senses: first, they are compatible with agents' desires to establish and
maintain valuable relationships of care and dependency; second, they allow for the
insight that capacities for autonomy and critical reflection are the product of family
influences, socialization, and so on; and third, they respect differences among agents,
especially with respect to life plans and conceptions of the good, which are pro-
duced by the different social contexts in which agents are embedded.
Despite the apparent advantages, however, feminists should be cautious about
adopting a purely procedural account of autonomy. In certain cases, even prefer-
ences satisfying the standards of critical reflection that are required by procedural
accounts would still be regarded as nonautonomous by many feminists. This is be-
cause such preferences are influenced by pernicious aspects of the oppressive con-
text. They therefore attract what I call thefeminist intuition, which claims that pref-
erences influenced by oppressive norms of femininity cannot be autonomous.
In the first section, I illustrate the feminist intuition by using examples from
Kristin Luker's study of women who take contraceptive risks.6 I suggest that the re-
sponse that such women are not autonomous is a manifestation of the feminist in-
tuition. In the second section, I argue that the feminist intuition cannot be ex-
plained by any of the conditions suggested as necessary for autonomy by procedural
theories. If I am right, we have to either reject the feminist intuition or reject purely
procedural theories of autonomy. In the final section, I propose that to explain the
feminist intuition, we need to endorse a substantive rather than a procedural ac-
count of autonomy.7 Substantive theories claim that there are further necessary con-
ditions of agents' autonomy in addition to the formal conditions derived from pro-
cedural theories. For example, some theorists argue that agents must have a sense of
self-trust or self-worth, in addition to relevant capacities for critical reflection; others
argue that agents' preferences must have certain specified contents. I argue that only
a strong substantive theory, namely, one that places restrictions on the contents of
agents' preferences, explains the feminist intuition. Although I cannot provide any
detailed development of such a theory here, I hope at least to show that feminists
committed to the feminist intuition must take on the task of articulating a strong
substantive theory of this sort.
96 Autonomy and the Social

Autonomy, Rationality, and the Feminist Intuition

In her book, Taking Chances, Luker describes a study of a group of about five hun-
dred women who attended an abortion clinic in California in the early 1970s. De-
tailed interviews were done with fifty women, who were selected on the basis of
medical records. Luker's aim was to investigate the relationship between abortion
and contraception, in particular to answer this question: "Why do women who can
presumably use freely available methods of contraception end up having unwanted
pregnancies which result in induced abortions?" 8 Among the women studied, the
unwanted pregnancies were not due to ignorance or lack of skills or access to con-
traception. Luker's study showed that over half had used a prescription method of
birth control, and 86% had used some method of birth control in the past. Yet 40%
used no contraception in the month prior to becoming pregnant, and 26% used
contraception inconsistently, including such methods as rhythm. 9
A typical response to the decision making of Luker's subjects is to label it irra-
tional or perhaps even crazy. Luker's aim in the book is to challenge this intuition, to
argue that the decision making leading up to the decision not to use contraception
or effective contraception is, although risky, nevertheless rational. According to
Luker, these "contraceptive risk-takers" engaged in a process of tacit bargaining with
themselves over the costs and benefits of using contraception and in the end chose
to take a risk. If she is right, her subjects' decision not to use contraception is not
prima facie irrational or crazy. There is no more significant occurrence of irrational-
ity among these women than there is in the population at large in situations of de-
cision making "under risk conditions" such as smoking of failing to wear a safety
belt in a car.10 More strongly, she argues that "their behaviour leading up to the un-
wanted pregnancy is both reasonable and logical given their own definition of the
situation."11
Luker provides a strong defense of her subjects' rationality. In so doing, she
challenges the stereotypical antifeminist idea that women's decision making, when it
is difficult to explain or does not conform to expectations, is likely to be irrational.
I argue here that the example of Luker's subjects also triggers the intuition that they
are not autonomous. However, the cases of rationality and autonomy are not sym-
metrical because it is ptecisely the intuition that the subjects are not autonomous
that is the feminist intuition. In the next sections, I sketch some connections be-
tween rationality and autonomy, and then I outline the reasons adopted by Luker's
subjects in the bargaining process and explain why it attracts rhe feminist intuition.

Rationality and Autonomy

Luker's study raises the issue of the relationship between rationality and autonomy.
Many theories of autonomy propose that rationality is a necessary condition of au-
tonomy or, more strongly and in the Kantian spirit, reduce autonomy to a capacity
for rational decision making. For instance, procedural theories of autonomy often
impose a rationality condition on autonomous preference formation, typically a
"minimal rationality" condition. 12 As one would expect of a condition of a purely
procedural theory of autonomy, the minimal rationality condition is itself purely
 Autonomy and the Feminist Intuition 97

procedural, requiring agents to manifest formal consistency among the relevant be-
liefs and desires. On the other hand, substantive theories of autonomy tend to re-
duce it to substantive rationality. In Susan Wolf's theory, for example, autonomy is
a capacity to form preferences that track objective standards of "Reason."13
More significant in this context, there are structural parallels between theories
of rationality and theories of autonomy. Both divide into the procedural and the
substantive. For example, in a recent response to Luker's argument, Elizabeth An-
derson in "Should Feminists Reject Rational Choice Theory?" identifies various the-
ories of rational choice. The formal theory describes rationality as the disposition to
maximize utility, namely, to weigh up preferences in the light of subjective assess-
ments of their costs and benefits and choose the option that maximizes the utility
of the overall scheme. Since the content of preferences is irrelevant to maximiz-
ing the overall scheme, the formal theory is purely procedural.14 On the other hand,
in the rhetorical theory of rational choice, an agent is rational when she has certain
substantive characteristics, for instance, those of self-reliance, autonomy, and self-
confidence.15 Thus, the rhetorical theory is substantive.
Anderson draws out the relationship between autonomy and rationality. She in-
terprets Luker as adhering to the formal theory of rational choice. Luker's subjects
are not irrational in this theory because rationality requires only that they order their
preferences to maximize utility. It does not matter that weighing up the costs and
benefits of contraception leads to taking a contraceptive risk or that the contents of
the preferences adopted by Luker's subjects may be criticized from a feminist point
of view. Anderson herself adopts the rhetorical theory, which she argues has greater
explanatory power because it offers a richer account of agency. In the rhetorical the-
ory, autonomy is one of the characteristics that a rational agent must display. Since
Anderson thinks that Luker's subjects are not autonomous, they cannot be rational
either.
The conception of autonomy that Anderson takes to be implicit in the rhetor-
ical theory of rational choice is as follows. She suggests, first, that autonomous
agents must regard themselves as authorized to act on their own interests and order-
ing of preferences, that they not "bow down to social convention, tradition, or even
morality," or "take other people's reasons for how they should act as their reasons
for action." Second, autonomous agents must "regard themselves as self-originating
sources of claims."16 The first condition has a stronger and a weaker formulation.
The stronger is that autonomous agents must not take others' reasons for action as
their own, and the weaker is that autonomous agents must regard themselves as au-
thorized to act on their own interests. The stronger formulation is false, however, be-
cause procedural theories (rightly) treat autonomy as compatible with socialization,
and acting on reasons generated through socialization is inevitably a case of taking
others' reasons as one's own. The question for all theories of autonomy is what kinds
of socialization are incompatible with autonomy. The weaker formulation, while
plausible, is not a purely procedural condition, and neither is the second condition
that Anderson describes, that is, that autonomous agents must regard themselves as
self-originating sources of claims. Procedural accounts would not typically conceive
failures to regard one's interests or claims as self-authorizing or self-originating as suf-
ficient to undermine autonomy. Such failures are closer to failures of confidence or
98 Autonomy and the Social

trust in one's ability to make claims than to flaws in the capacity for critical reflec-
tion. For example, Luker's subjects lack confidence in asserting their sexual agency
and, as a result, do not have a robust sense of their own authority in asserting their
claims. (I return to this issue in the final section.) Thus, the two conditions Ander-
son identifies seem to go beyond the conditions for autonomy required by recent
procedural accounts. The conception of autonomy implicit in the rhetorical theory
of rationality is a substantive conception.

The Feminist Intuition

I suggest that Anderson's reaction to the question of the autonomy of Luker's sub-
jects is an example of the feminist intuition. They are judged to be nonautonomous
because they are overly influenced in their decisions about contraception by stereo-
typical and incorrect norms of femininity and sexual agency. Unlike risk-takers in
other domains, such as those who smoke or fail to wear safety belts in a car, Luker's
subjects are motivated by oppressive and misguided norms that are internalized as a
result of feminine socialization. Because of the role of such norms in the decision
to take a contraceptive risk, Luker's subjects attract the feminist intuition.
Let us consider in more detail the kinds of reasons weighed up by Luker's sub-
jects in the tacit bargaining process at the time of the decision to take a contracep-
tive risk. She lists both the costs of contraception and the anticipated benefits of
pregnancy in the "bargaining set." For example, among the costs of contraception
are the social and cultural costs, the costs of obtaining and maintaining contracep-
tion, and the medical and biological costs. The social and cultural costs and benefits
described by Luker provide the best evidence of the kinds of social norms that are
motivating her subjects. Among the costs, there are the social costs of acknowledg-
ing intercourse by planning and using contraception. For example, there is the cost
of public censure for engaging in premarital sex. Consider the following excerpt
from an interview:

l: Why didn't you use more effective contraception?

R: I always thought about it, but never did anything about it. 1 used to think about the
pill, but my sister used it, she's married now and stuff, and my mother used to tell me
she'd die. She's really Catholic. But it seems as if most of my friends are on it.17

A second kind of social cost is that of flouting internalized religious or other

norms. This cost is to some extent removed by avoiding deliberate planning for sex
by not using adequate contraception:

l: Did you think you might get pregnant not using contraception?
R: I thought so, I mean, I knew it was a possibility. But there was this problem of my
religious background. If you are familiar with the Catholic Church it is against the
Church to use contraception or to have pre-marital sex. . . . Just using a contraceptive
seems like you're planning. 18
 Autonomy and the Feminist Intuition 99

A third cost identified by Luker is that of appearing to be a sexually active

agent.19 A pattern identified by Luker is ceasing to use contraception after the break-
down of a relationship with a husband or boyfriend. She observes that "continuing
contraception implies that one expects intercourse with a new partner, and this flies
in the face of a strong taboo."20
Luker argues that in many cases her subjects tacitly weighed up the costs of
obtaining and using contraception against what they anticipated would be bene-
fits of pregnancy. (The perceived benefits of pregnancy usually vanished once the
pregnancy became a reality.) Luker's subjects anticipated that pregnancy would
benefit them by proving fertility, establishing true womanness and increasing self-
worth, forcing the sexual partner to define his relationship to the woman, and
forcing significant others such as parents to pay attention to the pregnant woman.
For example:

I: You said you had a strong maternal urge. So you think that could have been a factor
in getting pregnant?
R: I think so, yeah. I don't know exactly, but taking a wild stab, I think that getting
pregnant means having someone who will take my love and care, 'cause lots of times I
think noone else wants it.21

The overall picture derived from these interviews is one of women motivated
in large part by the following norms: it is inappropriate for women to have active sex
lives; it is unseemly for women to plan for and initiate sex; it is wrong to engage in
or be seen to engage in premarital sex; pregnancy and childbearing promote one's
worthiness by proving one is a "real woman"; it is normal for women to bargain for
marriage by, for instance, proving their fertility to their partners or their partners'
families; and women are worthwhile marriage partners only if they are capable of
childbearing.
Contrast this picture with a hypothetical picture of risk taking by smokers. As-
sume that, like Luker's subjects, their risk taking is not explained by a lack of skills
or access to professional help. Imagine that many cases involve tacit bargaining,
paralleling the tacit bargaining of Luker's subjects. For example, suppose that the
financial and health costs of continuing to smoke are weighed against its benefits,
such as that smoking is pleasurable, that it is relaxing or therapeutic, or that it
would be financially onerous or otherwise too much trouble to seek professional
help to overcome the habit. For smokers who engage in a tacit bargaining process
of this sort, although we may judge them weak-willed or wrong in giving too little
weight to the medical costs, we are unlikely to judge that their decisions are not
autonomous because smokers are not typically opting to smoke on the basis of false
and oppressive norms. Notice, however, that when smokers (especially children) do
opt to smoke on the basis of false and oppressive norms, for example, in response
to images of smoking as glamorous and promoting one's desirability, the intuition
that they are not autonomous is triggered. Unlike risk taking among smokers, the
contraceptive risk taking of Luker's subjects attracts the feminist intuition because
the internalized norms motivating the decision to take a contraceptive risk have
100 Autonomy and the Social

criticizable contents. They are norms of religion, femininity, and sexuality that are
oppressive to women.

Procedural Theories
We saw in the preceding section that Luker's argument that her subjects are not ir-
rational succeeds if we take her to be endorsing a purely procedural theory of ra-
tional choice, namely, the formal theory described by Anderson. Because of the par-
allels between theories of rational choice and theories of autonomy, this suggests
that in procedural theories of autonomy, Luker's subjects should count as autono-
mous, or at least should not be ruled nonautonomous. In this section, I examine in
detail whether Luker's subjects would count as nonautonomous in procedural theo-
ries. In general terms, procedural theories claim that for a desire or preference to be
autonomous, it must satisfy certain standards of critical reflection. That is, critical
reflection must be autonomy conferring in the relevant way. I extract five necessary
conditions of autonomy from different procedural accounts: counterfactual, internal
coherence, endorsement, self-knowledge, and inhibiting factors. I argue in each case that
many of L,uker's subjects satisfy the condition and hence cannot be ruled nonau-
tonomous on that basis. Thus procedural theories are at best equivocal on the ques-
tion of whether Luker's subjects arc autonomous. They do not vindicate the feminist
intuition.

The Countcrfactual Condition

In providing an analysis of autonomous preference formation, Christman intro-
duces a counterfactual condition, as follows: an agent is autonomous with respect to
a certain preference or desire only if she did not resist its development when attend-
ing to the process of its development or would not have resisted had she attended to
the process.22 So, for example, suppose out of habit an agent chooses to go on vaca-
tion every year to the same location. The fact that the agent fails to attend critically
each time to the processes of formation of the preference does not rule out its being
autonomous unless, counterfactually, had the agent reflected on the process of de-
velopment of the preference, she would have resisted the development of the pref-
erence. Since in most cases of this sort agents would not have resisted the preference
had they attended to it, such preferences are not ruled nonautonomous on the coun-
terfactual condition.
Does the preference of Luker's subjects to take a contraceptive risk fail the coun-
terfactual condition? One approach is to take Luker's description of her subjects as
bargainers at face value. If her description is right, it is not plausible to claim that
they failed to attend to the process of formation of the preference. On the contrary,
they weighed up the costs and benefits of using contraception and decided to take
a risk. Since they attended to the process, there is no need to invoke the counterfac-
tual condition. On the other hand, since the bargaining described by Luker was
mostly tacit, it is likely that Luker's subjects did not sufficiently attend to their mo-
tivational structure, in particular to the influence of the norms of femininity and
sexual agency on the bargaining process; hence that they did not sufficiently attend
 Autonomy and the Feminist Intuition 101

to the process of formation of the preference. If they had attended, would they have
resisted?
To answer this question, it is useful to consider cases in which desires are
formed on the basis of external standards or internalized norms and as a result ap-
pear to be nonautonomous. For example, Feinberg distinguishes two paradigm pat-
terns of nonautonomous, or inauthentic, desires.23 The first is that of a "habitual
and uncritical conformist who receives his signals from some group whose good
opinion he needs, or from unknown tastemakers in the advertizing agencies and
public relations firms"; the second is that of an agent who is subject to "standards
[that] are implanted in the child by his parents, their authoritative source internal-
ized, so that they become his forever more."24 Suppose that both types of agents do
not sufficiently attend to the formation of their preferences when the preferences
are based on either the external standards to which they habitually defer or the in-
ternalized standards that have been operating since childhood. Would they have re-
sisted the formation of these preferences had they attended to the influence of the
standards or norms on their desires? It is unlikely that they would have, precisely
because the habits of deference and the internalized norms, that is, the values that
govern the agent's motivational structure, would themselves justify holding the rel-
evant desire. In other words, agents who attend to the influence of the values on
their preferences would be more likely to endorse the preferences than to resist
them.
Paul Benson makes the same point, using a different example:

Consider the eighteen year-old college student who excels in her studies, is
well-liked by her many friends and acquaintances, leads an active, challenging life,
yet who regularly feels bad about herself because she does not have "the right
look." . . . So, on top of everything else she does, she expends a great deal of time
and money trying to straighten or curl her hair, to refine her cosmetic technique,
to harden or soften her body, and so on.25

In claiming that the student is not autonomous, Benson expresses the feminist in-
tuition. He argues that the counterfactual condition cannot explain why the student
is not autonomous because even had she reflected on the fashion norms she has in-
ternalized, as well as the process of development of her preferences and its relation-
ship to the norms, she would not have resisted the process. As Benson explains, the
fact that the norms are internalized blocks the capacity of the agent to resist the de-
velopment of preferences based on the norms. If we consider only this procedural
condition, therefore, the student's decision to devote so much time and energy to
her appearance is procedurally competent and autonomous.
Many of Luker's subjects are influenced by internalized norms of feminine
sexual agency. Consider, for example, subjects who judge that one of the anticipated
benefits of pregnancy is to allow them to "bargain for marriage":

i: How did you feel about getting pregnant?

R: I thought I would probably get married, that he would want to. I never, ever thought
I'd go through an abortion. That never entered my mind at all.26
102 Autonomy and the Social

Although the case is somewhat underdescribed here, one way of interpreting this
woman's beliefs is that she is in the grip of feminine norms about the dependency
of women on men in sex, pregnancy, and marriage. Luker's phrase, "bargaining for
marriage," suggests that many of her subjects conceived of themselves as being un-
able to offer themselves as equal partners in marriage unless they brought some
stereotypical female value to the arrangement, especially the promise of children.
Moreover, they considered that once a man had "got them pregnant," he had some
obligation to fulfil his end of the bargain by providing her with institutional status
through marriage. Like Benson's student, these women are acting on internalized
norms, which they accept. Even if they had not sufficiently attended to the process
of formation of their preference to take a contraceptive risk, it is unlikely that they
would have resisted that process, and hence the counterfactual condition does not
rule out their autonomy with respect to the preference.

The Self-Knowledge Condition

A second condition of autonomous critical reflection that has been offered is that
the agent not be self-deceived in the formation of her desires and preferences. For
example, Diana Meyers claims that acting on one's authentic self, and hence
being autonomous, is the result of successfully exercising a number of skills, one of
which is the skill of self-discovery. She says that a typical autonomous person dis-
plays "active involvement in shaping herself and intimate knowledge of her evolv-
ing self."27 For Meyers, the authentic self is not something hidden and static, which
autonomous people successfully reveal to themselves; rather it is generated through
the exercise of the further skills of self-direction and self-definition. Christman also
includes a self-knowledge condition, which he describes as the requirement that
agents not tell themselves a "cover story" about the process of development of their
preferences.28
Recall that one of the costs of contraception described by Luker is the cost of
acknowledging oneself as an active sexual agent. If Luker's subjects are radically
self-deceived (as Anderson suggests29) because they do not acknowledge themselves
as active sexual agents, it is hard to see how they could satisfy the self-knowledge
condition. On the other hand, the motivations of Luker's subjects are diverse and
complex, and there is no uniform lack of acknowledgment to oneself of one's sexual
agency among them. Consider the following two contrasting cases:

i: Did you ever think about getting pregnant?

R: Sometimes I thought about it, but I didn't really pay too much attention, because we
weren't really into . . . we were goofing around . . . in an intimate . . .
I: Could you clarify that?
R: We started in on one thing and then we would go on to another thing and finally we
would have intercourse. We never said, "Well, we're going to do it today." It just hap-
pened.
(It "just happened" almost every night for over a year.)30
 Autonomy and the Feminist Intuition 103

I: You said that you had used the pill previously but had run out. Where did you get the
pills the first time?
R: A family-planning clinic in Southwest City.
I: Why didn't you get the prescription filled?
R: Because of my father. . . . We live in a small town, and the medical and dental peo-
ple are very close, and I couldn't go to another doctor without his finding out and I
think that would hurt him.31

A lack of acknowledgment to oneself, which amounts to self-deception, as in the first

case, should be distinguished from a lack of acknowledgment to others, as in the sec-
ond. In the second case, although the agent is influenced by norms of sexual agency
to the extent that she wishes to be seen to be observing them, she is not deceiving
herself about either the norms she subscribes to or the influence of the norms on her
preference formation. Her case is not one of «^deception.
Nor are different kinds of cases described by Luker obvious cases of self-deception.
One set of examples is of women who wish to engage in an active sex life while at
the same time being committed to religious or feminine norms that restrict women's
sexual agency. The motivations of such women are conflicting, and their decisions
not to plan for sex, and hence not to use adequate contraception, are attempts to
reconcile their conflicting desires. These women face an unfortunate dilemma, and
it may be weak-willed not to either wholeheartedly reject the norms and plan for
contraception or wholeheartedly accept the norms and forego an active sex life.
However, weakness of will does not amount to self-deception. Another set of exam-
ples is that in which women hope for pregnancy in the misguided belief that it will
promote their self-worth or increase their value as potential marriage partners. Al-
though these women are in the grip of false norms of femininity and sexual agency,
it does not follow that they are self-deceived about the influence of the norms on
their decision to take a contraceptive risk. Thus, although it is plausible that the au-
tonomy of the agent for whom sex '"just happened' every night for almost a year" is
extinguished as a result of her lack of self-knowledge, this is not the case for most of
Luker's subjects. In most cases, the decision to take a contraceptive risk is the prod-
uct of a complex array of factors, including a wish, based on pragmatic reasons, not
to be seen to be violating norms of female sexual agency.

The Internal Coherence Condition

It has been proposed that the desires and preferences of autonomous people must
be integrated in some sense. I examine two ways of spelling out the notion of inte-
gration in this and the following sections: integration as internal coherence and
integration as endorsement. Chrisrman's "minimal rationality" offers one kind of
internal coherence condition. His requirement demands that there be no mani-
fest inconsistencies among the beliefs and desires in the set of beliefs and desires
that contribute to the processes of reflection.32 Since minimal rationality is under-
stood as internal consistency among agents' beliefs and desires, not as consistency
between beliefs and the world, it is an internalist requirement of rationality. Marilyn
104 Autonomy and the Social

Friedman suggests another kind of internal coherence condition. She argues that
higher- and lower-order desires must be integrated in the sense that higher-order de-
sires are subject to revision in light of lower-order desires, and vice versa. In her ac-
count, autonomy requires a reflective equilibrium among higher- and lower-order
desires.33
To judge whether Lukcr's subjects violate an internal coherence condition, we
need a fine-grained account of the kinds of inconsistencies among desires that can
plausibly be said to be autonomy undermining. Many inconsistencies, especially
conflicts among first-order desires, are innocuous and should not count as auton-
omy undermining, even if contained within the desire set leading to the formation
of the relevant preference. Moreover, some types of incoherence seem to be required
by proper processes of rational reflection. Indeed, this situation is typical of mem-
bers of oppressed groups who are attempting to integrate the norms that govern
themselves with those that govern a nonoppressed outside world.34
Consider again the difference between the woman whose "decision" to risk sex-
ual intercourse without contraception is the product of paralyzing ambivalence and
the woman whose decision not to use contraception is the result of a conflict be-
tween the desire to have sex and the desire not to reveal her sex life to her father.
The former is plausibly an example of a manifest breakdown of internal coherence,
whereas the latter is closer to a conflict among first-order desires, which should not
be characterized as autonomy undermining. This is so in particular if the motivation
for her behavior is to be seen by her father as acting on the basis of acceptable
norms, rather than the norms of femininity or religion per se. The decision not to
use contraception in these particular circumstances is an imaginative way of bring-
ing the values of respect for one's parents into reflective equilibrium with first-order
desires.
Only a minority of Luker's subjects display a level of ambivalence that is suffi-
ciently extreme to count as a manifest violation of the internal coherence condition.
Yet many are torn between the norms of sexual agency that apply to women as an
oppressed group and the norms of sexual agency that should apply to them, that is,
the norms that apply to the relatively nonoppressed group of men. On the one
hand, the decisions of Luker's subjects are influenced by stereotypical ideas, such as
that sexual intercourse is appropriate for women only within marriage or a simi-
larly committed relationship; on the other hand, their decisions reflect the influ-
ence of more "liberated" notions of sexual agency. Does this constitute autonomy-
undermining internal incoherence?
Luker's subjects are often attempting to reconcile the norms of the two groups
to which they belong: the preliberated group, to whom the stereotypes of female
sexual agency are applicable, and the postliberated group of active sexual agents in
whom the difference between the sexes is not salient. This kind of situation is typi-
cal of members of oppressed groups. For example, Cheshire Calhoun argues that it
is precisely this kind of ambivalence that is experienced by women who are both
Latina and lesbian.35 The sexual norms imposed by Hispanic communities are often
irreconcilable with those of the gay and lesbian communities. Hence, for someone
in such a situation, experiencing ambivalence about both their sexual and their His-
panic identities is an inevitable consequence of reflecting on the situation in which
 Autonomy and the Feminist Intuition 105

they find themselves. Although Calhoun's argument is a discussion of internal co-

herence as a necessary condition of integrity, not of autonomy, her conclusions par-
allel those appropriate for the latter. She argues that the ambivalence or internal in-
coherence of such agents does not undermine their integrity because integrity in fact
requires that agents who are members of different groups simultaneously face up to
the ambivalence inherent in their experience. Similarly, the ambivalence experienced
by Luker's subjects is often a product of an attempt to liberate themselves from the
oppressive norms of sexual agency that are applied to the group of women. It does
not constitute a manifest breakdown in the capacity for critical reflection, which
would be incompatible with autonomy. Rather, a proper exercise of critical reflec-
tion requires acknowledgment of the incoherence inherent in one's circumstances.
Procedural theories rely on characterizing autonomous preference formation as
the result of processes of critical reflection. For internal incoherence to count as
being incompatible with autonomy, therefore, it must be sufficiently problematic to
count as a breakdown of the processes of critical reflection required by procedural
accounts. I suggest that only extreme breakdowns of coherence are problematic
enough to count.

The Endorsement Condition

An alternative way of understanding the integration condition is as that of iden-
tification or endorsement rather than simply consistency or internal coherence.
In Frankfurt's account, for example, first-order desires are autonomous just in
case they are endorsed by second-order desires. More precisely, a first-order desire
is autonomous only if an agent has a second-order desire (a second-order "volition")
that the first-order desire be effective.36 A good example of a failure of endorsement
being autonomy undermining is Frankfurt's example of the unwilling addict. The
unwilling addict has a desire to take drugs but a desire to desire not to take drugs.
He does not therefore desire that the first-order desire be effective; it is not endorsed
by a second-order volition. This suggests that his first-order desire is not under his
control and that it is not autonomous.37
Failure to satisfy an endorsement condition is a possible way of explaining the
prima facie nonautonomy of some of Luker's subjects. For example, many desire
premarital sex yet do not seem to endorse this desire at the second-order level. In
Frankfurt's terms, they do not "have the will they want" because at a second-order
level, they want their behavior to be governed by norms of femininity or religion
that are incompatible with having premarital sex. Thus, since a decision to take a
contraceptive risk is a product of an unendorsed first-order desire, it is not au-
tonomous.
For most of the subjects, however, the motivational structure is not as simple as
this. First, in many cases of the tacit bargaining that Luker describes, there is no fail-
ure to endorse the first-order preference to take a contraceptive risk. Consider a
woman who thinks that deliberate planning for sex is distasteful and also that get-
ting pregnant will have the benefit of testing her partners commitment to her. The
cost of using contraception together with the benefit of not using it lead her to de-
cide to take a contraceptive risk. In this kind of case the preference seems to be en-
106 Autonomy and the Social

dorsed by a set of norms, all of which lead in the same direction, that is, against
using contraception. Second, several of Luker's cases are ones in which the second-
order desires conflict, not those in which there is a failure to endorse a first-order de-
sire. Anderson observes that Luker's subjects are often "caught between contradic-
tory norms of femininity," for example, that it is not nice to have premarital sex and
that in situations of sexual intimacy, women should subordinate their desires to
those of their male partners. 38 In cases in which a woman's partner wishes to have
premarital sex, the norms come into conflict. If any of Luker's subjects, or women
like them, are in the grip of conflicting or contradictory norms such as these, a con-
flict of desire at the first-order level can be traced back to one at the second-order
level. It is not a case, as in the unwilling addict, of a conflict between an unendowed
first-order desire and a second-order volition because here both first-order desires are
endorsed at a second-order level.39 If I am right, the endorsement condition will not
explain the nonautonomy of many of Luker's subjects.

The Inhibiting Factors Condition

The final condition I wish to address is articulated by Christman. He claims that the
lack of resistance to the development of a preference or desire must not have taken
place (or would not have taken place) under the influence of factors that inhibit self-
reflection (unless exposure to such factors was autonomously chosen, in which case
that choice had to be made without such factors).40 There are at least two kinds of
factors that might inhibit self-reflect!on, external factors and internal factors. An ex-
ample of the former is offered by Christman's own "happy slave" case. Suppose a
"'happy slave' has expunged her desires for freedom only as a result of the oppres-
sive presence of the restraints she faces."/l1 In other words, the slave has adapted to
the oppressive circumstances imposed on her and no longer desires to be free of
them. Her desires are nevertheless not autonomous, even if she has not resisted their
process of formation, because the desires to adapt and remain unfrec were formed
only as a result of the presence of oppressive external factors that she could not
avoid. External factors that operate to inhibit self-reflection are not present in the
case of Luker's subjects. Unlike happy slaves, Luker's subjects are not subject to ex-
ternal constraints that make it impossible to extract themselves from an oppressive
situation. Their lack of resistance to the formation of the preference to take a con-
traceptive risk is not the product of adapting to circumstances that they cannot
avoid.
There are several kinds of internal factors that might inhibit self-reflection. The
classic examples are those of obsession, addiction, and compulsion. A different kind
of internal factor that inhibits self-reflection is one that derives from certain norms
of femininity themselves. For example, historically women have been characterized
as emotional rather than rational, as agents whose motivations are influenced more
by intuition than by careful critical scrutiny. If childhood socialization has encour-
aged the internalization of such norms, an agent may simply fail to develop the ca-
pacity to critically examine her own motivations in the way required for autonomy.
Thus, even if she does not resist the process of development of her desires, the lack
of a sufficient capacity to reflect on her motivational structure may be an inhibiting
factor, undermining the autonomy of those desires.
 Autonomy and the Feminist Intuition 107

The question of whether internal factors are present to inhibit Luker's subjects'
capacity for self-reflection is difficult to answer. As Meyers has pointed out, feminine
socialization typically hampers the development of the skills of critical reflection
that are essential for an agent to have a high degree of autonomy.42 Since there are
degrees of capacities for autonomy, however, feminine socialization will not extin-
guish female agents' autonomy altogether. Even women socialized through stereo-
typical feminine socialization will often have developed good capacities for critical
reflection and hence for autonomy. Thus it would be rash to conclude that Luker's
subjects have severely hampered critical capacities, to the extent that their prefer-
ences to take contraceptive risks, and indeed most of their other preferences, are au-
tomatically nonautonomous. Moreover, such a conclusion would fly in the face of
Luker's own description of her subjects as bargainers.

Normative Competence Theories

None of the five conditions that I have extracted from procedural theories of au-
tonomy vindicates the feminist intuition. If I am right, feminists who endorse both
the intuition and a procedural theory of autonomy are faced with a dilemma: either
they reject the intuition, or they reject the position that procedural conditions are
sufficient for autonomy. In this section, I look at two different substantive theories of
autonomy. Both are normative competence theories and both have been defended
(at different times) by Paul Benson. I suggest that only the stronger of the two the-
ories fully explains the feminist intuition, and hence unless feminists reject the in-
tuition, they must take on the task of defending this stronger theory.
The weaker normative competence theory offers the following necessary
condition of autonomy: agents must have a "sense of worthiness to act [which in-
volves] regarding oneself as being competent to answer for one's conduct in light
of normative demands that, from one's point of view, others might appropriately
apply to one's actions."43 One element of the conception of autonomy endorsed by
Anderson—that agents fail to be autonomous when they fail to regard themselves as
"self-originating sources of claims"— could be explicated by using a normative com-
petence theory of this kind. To say that agents must regard themselves as the self-
originating sources of claims suggests that agents must have a subjective sense of
their authority to answer for their own preferences and actions. Agents who lack this
subjective sense, through lack of self-confidence, self-trust, or self-esteem, will fail to
be autonomous.44 An example of an agent who lacks normative competence is Ben-
son's "gaslighted" woman. This woman has characteristics—she is passionate, ex-
citable, and "prone to emotional outbursts in public"—as a result of which her hus-
band and the scientific establishment believe that she is significantly unstable. Since
she trusts these external judgments of her character and behavior, her sense of her
own competence as an agent capable of answering normative demands placed on
her is seriously eroded. Her disorientation and self-conception as crazy undermine
her capacity for autonomy without undermining the capacities for critical reflection
required for autonomy by procedural theories.45
The stronger normative competence theory claims that what is required for au-
tonomy is "an ability to criticize courses of action competently by relevant norma-
tive standards."46 Benson argues that the effective internalization of false or irrele-
108 Autonomy and the Social

vant norms together with the fact that such norms are false or irrelevant diminish or
extinguish agents' capacities for autonomy with respect to decisions governed by the
norms. As Christman puts it, in this theory, normative competence, and hence au-
tonomy, is impaired because agents do not "have an understanding of the correct
norms applying to their actions."47 An example of an agent who lacks autonomy on
the stronger but not the weaker normative competence theory is that of Benson's
eighteen-year-old student who has internalized the norm purveyed by the fashion
industry that most women's natural physical appearance is deficient. The very inter-
nalization of the norm blocks her capacity to effectively criticize this "false construal
of [her] personal value," and hence she fails to be autonomous with respect to the
domain of her decision making that is governed by the norm.'18
The weaker normative competence theory might be thought, prima facie, to
explain failures of autonomy in Luker's subjects. First, certain subjects believe that
pregnancy and motherhood will raise their self-esteem and their standing as women
in the eyes of family members or society at large. This suggests that they have a low
sense of self-worth, which may hamper their normative competence. Second, An-
derson observes that Luker's subjects sometimes frame their actions as excuses, es-
pecially as excuses for not having sex with their partners. She argues that to do so
"is to concede his presumptive authority over one's actions" and hence to fail to re-
gard one's claims as self-authorizing or originating. 49 If she is right, perhaps this in-
dicates that Luker's subjects do not regard themselves—but rather their partners or
society—as normatively competent. Third, certain of Luker's subjects are ashamed
of their sexual agency, and as Benson points out, shame is sometimes sufficient to
erode the sense of self-worth necessary for normative competence. 50
Let us consider each reason in turn. Certain subjects do judge (at the time of
deciding whether to use contraception) that the benefits of pregnancy outweigh the
disadvantages precisely because they accept the norm that pregnancy and mother-
hood increase women's worthiness. If their low self-esteem involves a failure to re-
gard themselves as competent to answer others' expectations and demands, a deci-
sion to take contraceptive risks may fail to be autonomous. On the other hand,
accepting and acting on the norm that pregnancy increases self-worth suggests that
the women concerned take themselves as competent to answer the normative de-
mands that they regard as applicable to them. The wish to become pregnant is pre-
cisely a wish to increase their self-esteem by increasing their standing in the norma-
tive community to which they belong. The sense of self-worth that is lacking in
Luker's subjects is not tied in the appropriate way to a lack of a sense of normative
competence. Neither does framing a reason as an excuse convincingly show that the
agent lacks the relevant kind of self-worth. Framing a reason as an excuse could in-
dicate weakness of will or a desire to resolve conflict in a nonconfrontational way,
rather than a failure to regard oneself as normatively competent. Even if it does in-
dicate a failure of a sense of self-worth sufficient to undermine normative compe-
tence, only a small minority of Luker's subjects use the decision not to use contra-
ception as an excuse. Yet, by hypothesis, most of Luker's subjects' decision making,
even when not characterized as an excuse, is informed by oppressive and misguided
norms and hence is not autonomous. Finally, whether or not shame is sufficient to
undermine one's sense of normative competence is a matter of degree. Whereas many
 Autonomy and the Feminist Intuition 109

of Luker's subjects are ashamed to some degree of their desire for sexual agency,
most do not exemplify a level of breakdown of the sense of self-worth sufficient to
undermine their capacity to regard themselves as members of the normative com-
munity. In the case of Benson's gaslighted woman, the sense of self-worth that is
eroded is precisely the sense of worthiness to participate as a full member of the nor-
mative community. However, the shame experienced by Luker's subjects is typically
a result of failing to desire to live up to norms—those of religion or femininity—
that the subjects take as applicable to them. This suggests that whereas they are often
wrong in their judgment that those particular norms apply to them, they are not
failing to take themselves as normatively competent to answer the demands that
from their point of view others appropriately apply to them.
The intuition that Luker's subjects are not autonomous is underpinned not by
the lack of normative competence in the weaker sense but rather by the lack of nor-
mative competence in the stronger sense. Women who accept the norm that preg-
nancy and motherhood increase their worthiness accept something false. And be-
cause of the internalization of the norm, they do not have the capacity to perceive
it as false. Most of Luker's examples can be explained in these terms. The reason that
Luker's subjects are judged not to be autonomous is that the reasons weighed up in
the bargaining process—the costs of active sexual agency, as well as the benefits of
pregnancy—are often derived from false norms that have been internalized, such as
that women should not actively desire sex or prepare for sex in advance, that preg-
nancy is an expression of "real" womanhood, or that pregnancy is likely to lead to a
marriage commitment from one's partner and that this is a good thing. It is the con-
tent of these norms that can be criticized from a feminist point of view, not the way
in which Luker's subjects engage in the bargaining process. To vindicate the feminist
intuition that the subjects are not autonomous, therefore, feminists need to develop
a strong substantive theory of autonomy.51

Notes

1. I am grateful to Linda Barclay, Ian Gold, Catriona Mackenzie, and an audience at

the Department of Philosophy, Wollongong University, for helpful comments on earlier
drafts of this article.
2. Gerald Dworkin, The Theory and Practice of Autonomy (New York: Cambridge Uni-
versity Press, 1988), especially chap. 1; John Christman, "Autonomy: A Defense of the Split-
Level Self," Southern Journal ofPhilosophy 25 (1987): 395-420; "Autonomy and Personal
History," Canadian Journal of'Philosophy 20 (1990): 1-24; and "Liberalism and Individual
Positive Freedom," Ethics 101 (1991): 343-359.
3. For discussions of the notion of content-neutrality, see Christman, "Autonomy and
Personal History" and Paul Benson, "Free Agency and Self-Worth," Journal of Philosophy
91(1994): 650-668.
4. Jennifer Nedclsky, "Reconceiving Autonomy: Sources, Thoughts and Possibilities,"
Yale Journal of Law and Feminism 1 (1989): 7-36, 12.
5. Marilyn Friedman, "Autonomy and Social Relationships: Rethinking the Feminist
Critique," in Feminists Rethink the Self, ed. Diana Tietjens Meyers (Boulder, CO: Westview,
1997). Quote from Nedelsky, "Reconceiving Autonomy," p. 12. See Diana T. Meyers, Self, So-
110 Autonomy and the Social

defy, and Personal Choice (New York: Columbia University Press, 1989), for an extensive dis-
cussion of the ways in which socialization affects the development of capacities for autonomy.
6. Kristin Lukcr, Taking Chances: Abortion and the Decision Not to Contmcept (Berkeley:
University of California Press, 1975). I am indebted to a paper by Elizabeth Anderson,
"Should Feminists Reject Rational Choice Theory?" presented to the APA Eastern Division
Meetings, 30 December 1996, which suggested the idea of using Lukcr's examples to discuss
autonomy. I have benefited greatly from Anderson's discussion. All references to Anderson's as
yet unpublished paper are with permission.
7. Paul Benson defends substantive theories of autonomy in "Freedom and Value,"
Journal of Philosophy^ (1987): 465-486; "Feminist Second Thoughts about Free Agency,"
Hypatia 3 (1990): 47-64; "Autonomy and Oppressive Socialization," Social Theory and Prac-
tice 17 (1991): 385-408; and "Free Agency and Self-Worth." My approach in this article
owes a great deal to Benson's work. Sarah Buss defends a different kind of substantive theory
in "Autonomy Reconsidered," in vol. 19 of Midwest Studies in Philosophy, ed. P A. French, T. A.
Uehling and J I. K. Wettstein (Minneapolis: University of Minnesota Press, 1994): 95-121.
Substantive theories are discussed in greater length in Catriona Mackenzie and Natalie Stol-
jar, "Autonomy Refigured," in this volume.
8. Luker, 'Hiking Chances, p. 13.
9. Ibid., pp. 2o! 23-24.
10. Ibid., pp. 17,34-35.
11. Ibid., p. 17.
12. For example, Christmau, "Liberalism and Individual Positive Freedom," pp. 349-351.
13. For a discussion of autonomy as rationality, see Richard Lindley, Autonomy (Lon-
don: Macmillan, 1986), chap. 2. For Susan Wolf's theory of autonomy, see 1'reedom within
Reason (New York: Oxford University Press, 1990). In A Theory of justice, Rawls defines per-
sons as autonomous when they act from principles that express their nature as free and ra-
tional beings: A Theory of Justice (Oxford: (Oxford University Press, 1971), p. 515.
14. Anderson, "Should Feminists Reject Rational Choice Theory?" pp. 8-9.
15. Ibid., p. 10.
16. Ibid., pp. 10, 25. The latter formulation corresponds to Rawls's formulation of au-
tonomy in "Kantian Constructivism in Moral '1 'heory: The Dewey Lectures," Journal of Phi-
losophy 77 (1980), 515-572, 543.
17. Luker, Taking Chances, p. 46; for a full description of the bargaining process, see
chaps. 4 and 5.
18. Ibid., p. 45.
19. Ibid., pp. 47-48.
20. Ibid., p. 48.
21. Ibid., p. 68.
22. Christman, "Autonomy and Personal History" and "Liberalism and Individual Pos-
itive Freedom," p. 347.
23. foci Feinberg, "Autonomy," in The Inner Citadel: Essays on Individual Autonomy, ed.
John Christman (New York: Oxford University Press, 1989), p. 42.
24. Ibid.
25. Benson, "Autonomy and Oppressive Socialization," p. 389.
26. Luker, Taking Chances, p. 70
27. Meyers, Self, Society, and Personal Choice, p. 47.
28. Christman, "Autonomy and Personal History," p. 17.
29. Anderson, "Should Feminists Reject Rational Choice Theory?" p. 25.
30. Luker, 1'aking Chances, p. 4.
31. Ibid., p. 44.
 Autonomy and the Feminist Intuition 111

32. Christman, "Liberalism and Individual Positive Freedom," pp. 349-351.

33. Marilyn Friedman, "Autonomy and the Split-Level Self," Southern Journal of 'Phi-
losophy'24 (1986): 19-35.
34. For a detailed discussion, see Diana Tietjens Meyers, "Intersectional Identity and
the Authentic Self? Opposites Attract," in this volume.
35. Cheshire Calhoun, "Standing for Something," Journal of Philosophy 92 (1995):
235-260,236-241.
36. Harry Frankfurt, "Freedom of the Will and the Concept of a Person," Journal of
Philosophy 68 (1971): 5-20.
37. Notice that an alternative explanation for the lack of control here is that the desire
is the product of addiction.
38. Anderson, "Should Feminists Reject Rational Choice Theory?" p. 26.
39. The situation attracts one of the common objections to Frankfurt's theory, namely,
that the endorsement condition introduces a regress. When agents experience conflicts
among second-order desires, Frankfurt's theory must introduce a third-order level at which
one of the second-order desires must be endorsed for it, rather than the others, to count as
autonomous. See Gary Watson, "Free Agency," Journal of Philosophy 72 (1975): 205-220.
40. Christman, "Liberalism and Individual Positive Freedom," p. 347.
41. Ibid., p. 354.
42. Meyers, Self, Society, and Personal Choice.
43. Benson, "Free Agency and Self-Worth," p. 660.
44. For a development of the idea that a sense of self-trust is necessary for autonomy,
see Carolyn McLeod and Susan Sherwin, "Relational Autonomy, Self-Trust, and Health Care
for Patients Who Are Oppressed," in this volume.
45. Benson, "Free Agency and Self-Worth," pp. 655-657.
46. Benson, "Freedom and Value," p. 469.
47. Christman, "Liberalism and Individual Positive Freedom," p. 357.
48. Ibid., p. 389.
49. Anderson, "Should Feminists Reject Rational Choice Theory?" p. 26.
50. Benson, "Free Agency and Self-Worth," pp. 657-659.
51. As Benson has noted, there are prima facie objections to a strong normative com-
petence theory. He accepts the objection that a strong normative competence theory like his
or Susan Wolf's conflates autonomy with responsibility (ibid., p. 660). This is one reason he
now rejects the strong theory. However, Philip Pettit and Michael Smith offer a set of argu-
ments for a version of a normative competence theory in which autonomy is conceptually
linked to responsibility. Hence they implicitly respond to the charge that strong normative
competence accounts of autonomy conflate autonomy and responsibility. See "Background-
ing Desire," Philosophical Review 94 (1990): 565-592; "Practical Unreason," Mind 102
(1993): 53-79; and "Freedom in Belief and Desire," Journal of Philosophy 93 (1996): 429-
449. Moreover, there may be other ways of spelling out the strong substantive component of
autonomy than by using a notion of normative competence.
 5

INDIVIDUALS, RESPONSIBILITY,
AND THE PHILOSOPHICAL
IMAGINATION1

Genevieve Lloyd

The "Situated" Self: Feminist Critique of Individualism

Themes of connection and interdependence have been strong in feminist ethics, po-
litical theory, and epistemology. Many feminist philosophers have emphasized the
need for a relational approach to selfhood—the need to highlight the connections
between persons rather than what separates them, community rather than individ-
uality, sociability rather than the solitary self. The nature of the desired conceptual
shift is illustrated in, for example, Virginia Helct's Feminist Morality. Held describes
feminist reconceptualization of the concept of the self as a challenge to "standard
moral theory." "To be adequate," she says, "moral theories must pay attention to the
neglected realm of particular others in the actual relationships and actual contexts of
women's experience."2 Social relationships, especially that between "mothering per-
son" and child, move to the center of Held's refocused approach to ethical theory.
Feminist philosophers are urged to recognize the centraliry of human projects of
nurturance and growth to our understanding of moral subjects. Although some
feminists have suggested that friendship should be the central category to be con-
trasted with the "rational contracting" of standard moral theory, mothering relations
have for Held the advantage of being more distant from, and hence more challeng-
ing to, common preoccupation with contractual relations.3
I do not wish to deny that it can be an enlightening thought experiment to attempt

112
 Individuals, Responsibility & the Imagination \ 13

to think through what is lacking—what is distorted or remains unrepresented—

when we think about relations of nurturance or friendship through the idea of the
"contracting individual." That construct undoubtedly fits some forms of social in-
teraction better than others. In seeing where it does not fit, we become more reflec-
tively aware of areas of human life that have been neglected in ethical, political, and
epistemological theory. There is, nonetheless, something perplexing about the rela-
tions between such feminist attempts at reconceptualization and the theories they
try to challenge. There are unresolved methodological issues here. Can we really re-
place at will in our theorizing the model of contracting individual with that of
mothering person or friend and have the alternatives serve the same function as
those they replace? If not, what light does that throw on the point and scope of
moral theory and on how it might change under the impact of feminist critique?
The exercise of imagination is often central to feminist challenges to moral the-
ory. Often their implicit strategy is to take a theoretical model that has become so
familiar as to be invisible and confront it with undertheorized areas of experience. In
trying to think of friendship or nurturance in terms of contracting individuals, we
become aware of a misfit or distortion. In moving between concrete experience and
theoretical constructs of selves in relation, we are aware of a greater dissonance in
the area of human nurturance or friendship than in more commercial interactions,
which may indeed have provided the original context of the model. The dissonance
confronts us with the lack of appropriate theoretical models for neglected areas of
human interaction; it also makes visible things we might otherwise miss about com-
mon theoretical models of selfhood. The methodological point is spelled out by
Seyla Benhabib, in the introduction to Situating the Self, where she argues that the
Enlightenment conception of a "disembedded and disembodied" subject can be seen
to be an illusion when we confront it with the realities of human development,
which force us to recognize that "subjects of reason" can be "finite, embodied and
fragile creatures" such as human infants.4
In challenging received models of selfhood, feminist critics do not thereby offer
new conceptual models or alternative, competing theories of what is involved in
selfhood. Feminist philosophy does not change moral or political theory simply by
slotting mothering person, friend, or infant into the places traditionally occupied by
contracting individual. The confrontation of abstract models with concrete realities
is just the beginning of the exercise of reconceptualization. What must come next
is the challenge of thinking through what is involved in relations of interdepend-
ence to give content to the desired reconceptualization of the self.
In this article I want to open up some of these issues in the reconceptualizing
of selfhood by exploring a social phenomenon whose reality challenges, from a dif-
ferent direction, some of the same assumptions addressed in feminist discussion of
the "situated" self: the fact that individuals can take responsibility for things that lie
beyond the scope of their individual agency. The collective dimensions of responsi-
bility intersect with the collective dimensions of selfhood. By looking more closely
at the processes through which we make ourselves responsible for what we have not
ourselves done—the processes through which human beings enter relations of soli-
darity that challenge oppositions between self and other—we can gain insight into
how our models of selfhood might change to accommodate the realities of interde-
114 Autonomy and the Social

pendence. The crucial shifts, I argue, involve a new emphasis on the temporal di-
mensions of selfhood.

Imagining Responsibility
Although collective responsibility has received considerable philosophical attention
in recent years, there is a striking rigidity in the conceptualization of what is in-
volved in being a bearer of responsibility. Acceptance of the claim that responsibil-
ity can be a collective matter seems to have had little impact on how we think of
selfhood and agency. Even when responsibility is attributed explicitly to collectivi-
ties, the idea of the individual seems to have a grip on the philosophical imagina-
tion, which resists conceptual change. In trying to make sense of collective respon-
sibility, we may think of the collectivity as an overarching super individual, whose
responsibilities are not reducible to those of the individuals that compose it; alter-
natively, we may think of the collectivity as a fiction, whose responsibilities can ul-
timately be cashed out in terms of the familiar old individuals who have tradition-
ally borne them. We see these alternatives, for example, in how we think of
corporations when attributing praise or blame to them, holding them accountable
for policies that produce unjust or inequitable outcomes. In some circumstances the
responsibilities of the corporation are regarded as distributing to the individuals that
compose them; in others the collectivity itself is treated as having responsibilities
that do not thus distribute. Either way, our conceptualization of responsibility re-
mains centered on the idea of the individual.
Paradoxically, in its very insistence on claims of collective responsibility, philo-
sophical discussion seems to have entrenched the grip of the individual as a theoret-
ical construct. With regard to responsibility, collectivities seem to function as meta-
phorical extensions of individual persons. As bearers of responsibility they are
centers of agency and hence appropriate objects of praise and blame. There is noth-
ing amiss in this. Collectivities are appropriately held responsible for what they do;
and the possibility of moving between their responsibilities and the allocation of re-
sponsibilities to individuals is crucial to many important issues of accountability.
My concern here is with the idea of the individual through which we think these
collective responsibilities. How might we think of selfhood if we let our imagination
run from the collectivity to the individual rather than in the other direction? Imag-
inatively accommodating the realities of collective responsibility might change how
we think of both selfhood and responsibility. It would allow, I hope to show, some
shift of emphasis away from preoccupation with issues of praise and blame, as well
as some insight into less individualistic ways of thinking of selfhood.
By thinking through what is involved in the relations of sympathetic identifi-
cation and solidarity involved in friendship we can gain insight into the nexus be-
tween individual and collective identity—between a self and its "other selves."
Understanding the kinds of connection between selves that undermine easy oppo-
sitions between individual and group—between self-interest and concern for the
good of others—can clarify the limitations of prevailing notions of selfhood and
open up space in which we can begin to transform them. To bring the issues into
focus, I want now to look at two illustrations of the interactions between relations of
 Individuals, Responsibility & the Imagination 115

sympathetic identification and responsibility. The first example comes from the re-
flections on friendship and responsibility offered by Jacques Derrida in Memoiresfor
Paul de Man, the second from the depiction of solidarity and self-sacrifice in Pat
Barker's novel Regeneration.

Friendship and Solidarity

In "Like the Sound of the Sea Deep Within a Shell: Paul de Man's War," published
as an addition to the revised edition of Memoires for Paul de Man, Derrida addresses
the scandal of allegedly anti-Semitic journalism published by the young de Man in
occupied Belgium, between 1940 and 1942.5 Derrida's essay is a richly layered set of
reflections on interwoven aspects of responsibility, especially its connections with
friendship and with time. It is written out of the distress of bereaved friendship, in
response to the discovery of allegedly anti-Semitic remarks in articles written by de
Man in journals associated with Nazi collaborators. As well as being a response to the
scandal, the essay is also a highly self-conscious reflective act of taking responsibility
out of friendship with the dead. Themes of time, memory, and mourning run
through the essay. The connections between memory and responsibility had been al-
ready addressed in the earlier lectures in the volume, where Derrida's reflections on
bereaved friendship are woven into his rereading of de Man's own writings on mem-
ory, mourning, and autobiography. Memory and responsibility, Derrida suggests in
the preface to the French edition, cannot here be thought without one another:
"What is recalled to memory calls one to responsibility."6
Those themes from the earlier essays in the volume take on added poignancy
with the subsequent eruption of the scandal. Derrida's reflections on the impact of
later events on how the earlier ones are construed present the future rather than the
past as the primary perspective for the consciousness of temporality. The future
makes determinate what occurs in the present. The "absolutely unforeseeable," he
says, is always the condition of any event. "Even when it seems to go back to a
buried past, what comes about always comes from the future."7 The point here is not
the relatively straightforward one that from the perspective of the present what has
not yet happened will come from the future. The point is rather that the determina-
tion of what is happening now—the determinate content of what it is—depends
on what is now still future. What happens now is determinate only from a perspec-
tive not yet available. Yet it is in this indeterminate present that we have to answer, to
respond to what is happening. So we take responsibility, in situations we neither
choose nor control, by responding to unforeseeable appeals, that is, to appeals from
the other that are addressed to us "even before we decide on them." Derrida's re-
sponse, his "answering for" de Man is already implicit in their preexisting friendship,
when the appeal is still unforeseeable.8
Derrida responds to the scandal out of the promise implicit in a bereaved
friendship, which could not anticipate the ordeal the future held in store for that
friendship. He responds for de Man, as he says, not "judging or approving" of what
he did but speaking once again "of-him-for-him," at a moment when de Man is no
longer there to speak in his own name.9 Since this was a friendship imbued with in-
tellectual life, the response involves assuming an intellectual responsibility to metic-
116 Autonomy and the Social

ulously read the whole body of journalistic writings that form the context of the of-
fensive remarks: "Through the indelible wound, one must still analyze and seek to
understand." 10 De Man's legacy becomes "the gift of an ordeal, the summons to a
work of reading, historical interpretation, ethicopolitical reflection, an interminable
analysis."11 Derrida responds, then, with resolutely attentive reading and with an at-
tempt to make possible close readings by as many others as choose to read: "As
quickly and as radically as possible, it was necessary to make these texts accessible to
everyone. The necessary conditions had to be created so that everyone could read
them and interpret them in total freedom. No limit should be set on the discussion.
Everyone should be in a position to take his or her responsibilities."12
The taking of responsibility here is, in Derrida's analysis, not just a standing in
for—a speaking in the place of—the friend who, being absent, as it happens,
through death, cannot speak for himself. The responsibility enacts the structure of
bereaved friendship; it is an answering to as much as for the friend: "Before answer-
ing, responding for oneself and_/or that purpose, in order to do so, one must re-
spond, answer to the other, about the other, for the other, not in his place as in the
place of another 'proper self but for him."13 There is, for Derrida, an "impossibility"
about this responsibility—an impossibility that arises from its undecidability and
reflects the rich concept of "impossible mourning," which he has developed in the
earlier essays. In the death of the other we recognize the possibility of the impossi-
ble— the interiorization of the other in memory—which in turn encounters the re-
sistance of the other to the closure of that intcriorizing memory. 14
The shifts that occur on the uncertain borders of self and other—of interior-
ization and recognition of the other—are for Derrida central to the understanding
of friendship, of mourning, and of selfhood. Grief brings the attempt to incorpo-
rate the lost other into the self—an attempt that is thwarted by the reality of friend-
ship, with its demands for recognition of the other: "An aborted interiorization is at
the same time a respect for the other as other, a sort of tender rejection, a movement
of renunciation which leaves the other alone, outside, over there, in his death, out-
side of us."1 ^ The distinctions between self and other here take on an uncertainty
that is unresolvable, for it is an uncertainty that is the very core of friendship, of
mourning, and of selfhood itself. The distress the living friend feels is also the dis-
tress of the dead. It is in this area of uncertain boundaries—somewhere between the
self and the ambiguously other self of the friend—that the realities and the "im-
possibilities" of both friendship and selfhood are enacted. The ambiguities of self
and other here are expressed in our being called to &z&-responsibility in circum-
stances in which we can not be appropriately held responsible for what has been
done. Contributory guilt is in no way involved in our example; what is at stake is
not the issue of blame. Derrida's task in taking responsibility is not that of clearing
his dead friend of blame; he cannot know in advance that this is how it will go. The
task is to answer for and to his dead friend— to assimilate whatever is found to be
true into the ongoing, open-ended, and incompletable task of friendship beyond
death.
The complexities of solidarity with the dead and their bearing on issues of re-
sponsibility are at stake also in my second illustration from the first volume, Regen-
eration, of Pat Barker's trilogy on the experience and aftermath of shell shock in
 h

I n d i

Ihndividuals, Responsibility of the Imaginaion 117

World War I. Barker's central character, modeled on the war poet Siegfried Sassoon,
is a pacifist, committed to a mental hospital because no one knows quite what else to
do with him. Sassoon is readily able—on plausible, even if untrue, medical grounds
—to avoid, if he so chooses, being returned to the trenches. (The issue of whether
it is Sassoon that is irrational or the principles and practices of wartime Britain is an-
other central theme in the books.) His decision to return is against all self-interest,
nor is it motivated by any patriotic altruism. It is also at odds with the pacifist prin-
ciples Sassoon has embraced.
Barker provides hints to the explanation of her character's apparently strange
behavior in fragments of the war poems of the historic Sassoon: a solidarity with the
dead and with those awaiting death, which fills his imagination to the point where it
excludes all other motivation. Solidarity with the dead sustains the character's sense
of self, and with it a sense of responsibility to those no longer able to escape—"the
homeless ones, the noiseless dead"—and to those still living, but doomed to death
—the "Battalion in the mud:
Out of the gloom they gather about my bed
They whisper to my heart, their thoughts are mine.

"Why are you here with all your watches ended?

'From Ypres to Frise we sought you in the line."
In bitter safety I awake, unfriended. 16

The sense of responsibility here seems irrational. It has nothing to do with the
patriotic ideology of self-transcendence or self-sacrifice, which is supposed to sustain
the soldiers' lives and give meaning to their impending deaths. Those sentiments are
meaningless to Barker's character. He feels lost and angry among his compatriots who
are living in England, distanced from the reality of war deaths while they mouth pa-
triotic platitudes. All that engages him are the ties of companionship and loyalty
that bind him to the dead and to the about-to-die. Imagination cannot bridge the
gap between the vacuously alive and the pointlessly dead. He takes on responsibil-
ity out of solidarity with others, with whom he shares the strange middle ground be-
tween self and other—the uncertain borders of self articulated in Derrida's talk of
the "impossibility" of mourning and of responsibility. For Sassoon it is impossible
to continue safely alive while those with whom he shares this shifting selfhood are
dead or facing death. Let us now see how all this bears on the broader issues of col-
lective responsibility and selfhood.
Neither of these examples of taking responsibility out of friendship or solidar-
ity can be readily accommodated into our familiar ways of thinking of selves as
sharply bordered units. In both examples we see that individuals act out of a sense of
themselves as bound up with others, in ways that resist clear-cut divisions between
self and other. Their taking of responsibility is explicable only through relations that
forge identities, in situations in which there can be no clear-cut distinction between
self and other, between egoism and altruism. What emerges is a picture of the con-
struction and constant transformation of identities through relations of sympathetic
identification that underlie, but are also themselves made possible by, the taking of
responsibility. Responsibilities taken on in this way make visible relations of sympa-
11 8 Autonomy and the Social

thetic identification through which are constituted the identities of individuals as

parts of wider wholes. The examples suggest that the explanation of the fact that re-
sponsibility can be a collective matter lies in the fact that the formation of selfhood
is also a collective matter. Processes of formation and transformation of selves inter-
act with the assuming of responsibilities, in contexts of interdependence and friend-
ship; and this in turn creates further possibilities of subjectivity and agency.
The illustrations I have chosen are examples of friendship, solidarity, and loy-
alty. Responsibilities that arise in a context of close friendship operate in some re-
spects, of course, very differently from responsibilities in the context of larger and
more impersonal collectivities. Bur we can gain insight from Derrida's analysis of the
responsibilities that are integral to bereaved friendship into something more general:
how it is that we can appropriately take responsibility for what we have not ourselves
done. The operations of emotion and imagination involved in the formation of so-
cial wholes can perhaps be seen more clearly here than in cither the formation of
more impersonal ties in wider collectivities or in the more intense interweavings
of self and other in sexual love or parent-child relations. But wherever processes of
sympathetic identification operate, we can see similar enabling shifts between ideas
of self and other and similar implications for the temporal dimensions of selfhood.
Derrida's essay shows us that responsibility does not always rest on preexistent
identities. 1 he retrospective responsibility for the dead that Derricla describes resists
understanding through ideas of free contracts. Responsibilities can be projected
back into a past, which precedes the friendships that make them possible; and they
can depend on future perspectives, which are in principle unavailable at the time of
action. These aspects of responsibility demand models of selfhood that take seri-
ously the temporal dimensions of human consciousness — its complex shifts be-
tween past and future. These temporal dimensions of responsibility are intercon-
nected with temporal dimensions of selfhood.
The inherent temporality of selfhood—the multiple forming and re-forming
of identities over time and within the deliverances of memory—is an important
strand in the political dimensions of collective responsibility. Often those political
aspects of responsibility involve our taking responsibility for a past in which we did
not ourselves act, perhaps did not even exist. What our forebears did "comes from
the future," in Derrida's phrase—the future that is now our present. The retrospec-
tive determination of what they did—a perspective not necessarily available to
them—is an important part of what we now are; and it is our being what we now
are that grounds our taking responsibility in the present, even where there is nothing
in the past for which we can be blamed.
The point of bringing consideration of relations of friendship and sympathetic
identification to bear on the understanding of collective responsibility is not that
preexisting sympathetic identifications constitute a criterion for the attribution of
collective responsibility. The point is rather that: through understanding those pro-
cesses in which the taking of responsibility is intertwined with the formation of
identities, we can understand how it is that responsibilities can be a collective mat-
ter. Emotion plays an important part in the interconnection between responsibility
and the formation of identities. But the claim is not that the presence of emotional
ties justifies holding people responsible for what they have not themselves done.
 Individuals, Responsibility & the Imagination 119

Emotion figures rather in the understanding of how individuals come to appropri-

ately take responsibility for what they have not themselves done. Having understood
that process, we can also see that in some circumstances it may be appropriate to
hold people responsible for not having thus taken responsibility for the past. To clar-
ify what is at stake here, both for responsibility and for selfhood, it is helpful to re-
visit an old debate in which Joel Feinberg and Hannah Arendt address the ques-
tion of how emotion—especially sympathetic identification—bears on collective
responsibility.

Emotion, Individuals, and Community

In her response to Feinberg's treatment of collective responsibility, Arendt strongly
attacks the tendency to conceptualize collective responsibility in terms of shared
emotion.17 Her concern is to establish a firm distinction between guilt and respon-
sibility. Collective responsibility should be understood as a special case of vicarious
responsibility; talk of such "misplaced feelings" as vicarious guilt, in contrast, con-
fuses the issues. In postwar Germany, she says the cry, "We are all guilty," which at
first sounded so noble, actually served to exculpate to a considerable degree those
who actually were guilty. Guilt, unlike responsibility, is "strictly personal"; it always
singles out: "Where all are guilty, no-body is."18
What is relevant for my purposes in Arendt's criticism of Feinberg is her rejec-
tion of the notion of "sympathetic identification," which he invokes to explain the
formation of bonds of solidarity that make individuals identify themselves as part of
a group. Arendt argues that Feinberg's use of this notion blurs the crucial distinction
between guilt and responsibility. But despite Arendt's dismissal of the relevance of
sympathetic identification to collective responsibility, we can, in the light of Der-
ridas reflections, see a place for it in understanding the expansion of subjectivity and
agency that makes it possible for individual selves to take responsibility for what
they have not done. Such understanding, in turn, can help us to understand an im-
portant strand in the notion of collective responsibility.
Arendt's central idea is to insist on the political dimensions of collective respon-
sibility. The term "collective responsibility" and the problems it implies, she sug-
gests, owe their relevance and general interest to political predicaments, as distinct
from legal or moral ones, which always relate to persons and what persons have
done. These political issues, she argues, should not be allowed to disappear into just
a special case of matters subject to "normal legal proceedings" or "normal moral
judgments."19 What is at stake in such discussions of collective responsibility is
whether the individual is responsible because of his or her participation in a group
undertaking. In her own approach, which stresses in contrast the political dimen-
sions of collective responsibility, the whole point is that the individual is respon-
sible just because of his or her membership in the group. The responsibility be-
longs to us as members of a community, and we can escape it only by leaving the
community—that is, since we cannot live without being in some community or an-
other, by exchanging one community for another. So the only really nonresponsible
people, she concludes, are the outcasts—refugees, the stateless.
It is a striking feature of Arendt's analysis of the political form of collective re-
120 Autonomy and the Social

sponsibility that it shifts the focus away from issues of blame and guilt to the need
for remedy. The concept of collective responsibility, she observes, belongs primarily
in situations in which Socrates's question of whether it is worse to do or to suffer evil
ceases to be important—where what matters is not who should be blamed but what:
should be done about it. But what is at stake here is not a generalized response to the
presence of injustice or oppression but the assumption of a responsibility that is
specifically "ours," despite our not being to blame for the injustice.20
Arendt's political form of collective responsibility highlights for us the distance
of this concept from the model of the self as "contracting individual." We do not
acquire these responsibilities through any decisions we make as individuals, nor do
we acquire them by contracting into a group whose actions or policies we thereby
accept. We acquire them by being born into a community. These are responsibilities
we have just by virtue of being who we are. But if we think of what it is to be a self
in terms of a model of self-contained individuals — capable of freely opting in and
out of collectivities and contracts — it is difficult to understand how we come to
have such responsibilities. In The Human Condition, Arendt introduces a concept of
"natality" as a kind of second birth, an initiation of words and deeds—a principle
of beginning through which we become members of a community. To make sense
of the claim that we can be born into responsibilities, she suggests, we must think of
the beginning of a self in terms of entry into a community rather than the physical
facts of birth.
Such an entry cannot be captured through the idea of a contracting individ-
ual, which assumes a preexisting self. Arendt's account of these collective respon-
sibilities seems to demand a complementary account of selfhood. What must a
self be if it is possible for it to have such responsibilities? What model of selfhood
can accommodate this political form of collective responsibility? How can it be
that simply by being born into a community we can be regarded as taking on its
burdens of responsibility for past injustices—injustices in the production of
which we played no role? Here we see a convergence between the demands that
arise from feminist critiques of individualism and the need for a theoretical con-
struction of selfhood and agency that can accommodate the political dimensions
of collective responsibility articulated by Arendt. By bringing the two sets of con-
cerns together, we can perhaps make some progress toward a reconceptualization
of selfhood and agency that can accommodate both. The consideration of the op-
erations of solidarity and sympathetic identification that Arendt dismisses as irrel-
evant to the political dimensions of collective responsibility can, at another level,
contribute to our understanding of the kind of theoretical construction of self-
hood that is needed.
T have argued that we can give content to Arendt's version of collective respon-
sibility by supplementing it with an understanding of the processes of sympathetic
identification that allow individuals to take on responsibilities for what they have
not themselves clone — including actions whose agency lies not only with others but
also with the irrevocable absence of the dead. Being attuned to the operations of
sympathetic identification can help us see the nexus between self and other that un-
dermines easy oppositions between the individual and the group. It can help us see
the join between individual and group identity, as well as the points where the op-
 Individuals, Responsibility dr the Imagination 121

position between self-interest and concern for the good of others breaks down. Let
us now see what emerges for the reconceptualization of selfhood.

Temporality and the Self

Talk of blurring of the boundaries between self and other quickly becomes vacuous
when we think in terms of a spatial imagining of the self; and spatial metaphors are
powerful in our prevailing political ideals of selfhood, helping to reinforce the grip
on the philosophical imagination of the idea of the sharply bordered individual.
Jean Bethke Elshtain has remarked on the interconnections between ideals of indi-
vidual and political sovereignty in "The Risks and Responsibilities of Affirming Or-
dinary Life." "What rights have become to individuals in the modern West," she sug-
gests— namely, "marks of a sovereign self"—sovereignty is to states: "the sovereign
state is immanentised, thereby constituting a view of citizens, or political subjects, as
sovereign selves."21 Spatial metaphors are caught up in this movement of thought.
The sovereignty of the self takes in aspects of the sovereignty of the state, which re-
inforce our thinking of selves as sharply bordered unities modeled on spatial terri-
tories; and the projection operates also in the opposite direction—from the idea of
a discrete, self-contained individual to the idea of a sovereign state.
What might it be to think otherwise of the relations between individuals and
the collectivities to which they belong, and how might it affect the ways we think of
ourselves and of our responsibilities? If we think of agency, letting our imaginations
run from the collectivity to the individual rather than in the other direction, the self
takes on something of the complexity and multiplicity of a collectivity. But to avoid
this collapsing into unhelpful metaphors of blurred spatial borders, we need to make
a shift that Derrida's reflections have already prepared us for—a shift from space to
time in our imagining of the self. Stressing these temporal dimensions of selfhood
can help clarify what is involved in taking responsibility for a past for which we can-
not be blamed. Blame for the unjust treatment of Australian indigenous people
might belong entirely in the past, but that does not remove present collective re-
sponsibility, of the kind Arendt is articulating. Nor is our collective responsibility in
the present removed if it can be shown that past policies were blameless by the stan-
dards and beliefs that prevailed at the time. We are still responsible in the present,
from the perspective of which the determinate content of the evil done by the group
with which we are identified is now accessible.
The sovereign, bounded self described by Elshtain is a construct modeled on
space—on parts, wholes, and borders. We are responsible, in this way of thinking of
selfhood, for what lies on our side of the borders and for the changes that come
about beyond those borders as a result of what we do on our side. But we have seen
that there are complexities of selfhood and related complexities of responsibility
that are not captured by this spatial model—complexities that arise from the tem-
poral aspects of being a self. These complexities spring not from the self's relations
with an external world but from its own inner relations with time—from its not
being confined to the here and now or even to what has ever been present to it.
Selves struggle for coherence, for unity—struggle to make themselves well-
functioning temporal, no less than spatial, wholes, brought together out of the rich
122 AutonomyandtheSocial

but confusing deliverance of memory and imagination. In their relations to time,

the limits of selves are less stable than in their spatial dealings with the world. The
capacities for memory and imagination, although they may be preconditions for ra-
tional agency, are also sources of instability, of a lack of fixity. The capacity to have
a past and to reflect on it is crucial to selfhood; and having, in the relevant sense, a
past is not something that admits of tidy borders between the individual existence
that is mine and the collective existence that precedes me, into which I am born.
The fact that a self is intrinsically something that has a past means that there is an
internal multiplicity of selfhood, an open-ended source of possibilities for what a
self can appropriate and enact. For each self there is a multiplicity of narratives into
which it can insert itself and from which it can rupture.
A self is born into a future in which it will make individual decisions, for which
it will be held responsible, praised or blamed. But it is also born into the past of its
communal life—a past that both precedes it and awaits it; a past of collective memory
and imagination—which must be reckoned with in the present. The endless multi-
plication of possibilities for what a self can be and do is kept under restraint by the
limitations of, and on, the here-and-now self that interacts spatially with the world.
Both spatial and temporal dimensions of selfhood are essential. They do not
compete for exclusive lights to represent the truth of selfhood, and there is room for
disagreement about which of them best helps the theorization of different facets of
human action and interaction. They arc more like two registers between which we
move in attempting to understand selfhood. Neither can be reduced to the other;
any more than our responsibilities can always be reduced to those that rest on hap-
penings for which there is some kind of individual to blame.

Notes

1. For helpful comments on earlier versions, I am grateful to Peta Bowdcn, Lorraine

Code, Penelope Dcutscher, Moira Gatens, Catriona Mackenzie, Ross Poole, Amelie Rorty,
Robert: Sparrow, and Natalie Stoljar.
2. Virginia Held, Feminist Morality: 'transforming Culture, Society, ami Politics (Chicago:
University of Chicago Press, 1993), p. 59.
3. Ibid., p. 72.
4. Scyla Benhabib, Situating the Self: Gender, Community and Postmodernism in Con-
temporary Ethics (New York: Routledgc, 1992), pp. 4-').
5. Jacques Derrida, Mernoires for Paul de Man, rev. ed., trans. Cecile Lindsay, Jonathan
Culler, Eduardo Cadava, and Peggy Kamuf (New York: Columbia University Press, 1989),
pp. 155-263.
6. Ibid., p. xi.
7. Ibid., p. 160.
8. Ibid., p. 164.
9. Ibid., p. 166.
10. Ibid., p. 216.
11. Ibid., p. 229.
12. Ibid., p. 219.
13. Ibid., p. 30.
 Individuals, Responsibility & the Imagination 123

14. Ibid., pp. 149-150.

15. Ibid., p. 35.
16. Pat Barker, Regeneration (London: Penguin, 1992), p. 189.
17. See Joel Feinberg, "Collective Responsibility," Journal of Philosophy 65, no. 14 (July
1968): 674-688, and Hannah Arendt, "Collective Responsibility," in Amor Mundi: Explo-
rations in the Faith and Thought of Hannah Arendt, ed. James W. Bernauer (Dordrecht: Mar-
tinus Nijhoff, 1987), pp. 43-50.
18. Ibid., p. 43.
19. Ibid., p. 44.
20. Here, as elsewhere throughout her work, Arendt makes a sharp distinction between
the political and the ethical. For a helpful discussion of the rationale for the sharpness of
Arendt's version of this distinction and of the problems it raises, see Seyla Benhabib, "Judg-
ment and the Moral Foundations of Politics in Hannah Arendt's Thought," in Situating the
Self pp. 121-147.
21. Jean B. Elshtain, "The Risks and Responsibilities of Affirming Ordinary Life," in
Philosophy in an Age of Pluralism: The Philosophy of Charles Taylor in Question, ed. James Tully
(Cambridge: Cambridge University Press, 1994), p. 76.
 6

IMAGINING ONESELF OTHERWISE1

Catriona Mackenzie

In the literature on autonomy, a few theorists have suggested in passing that im-
agination might play a role in autonomous reflection, deliberation, and action. In
her account of autonomy competency, for example, Diana Meyers mentions imagi-
nation as one of the capacities involved in its exercise and suggests that one way
in which autonomous agents can test and refine their self-portraits is through
imaginative enactment. 2 It has also been suggested that one way in which oppres-
sive social relationships and institutions may impair autonomy is by restricting
agents' imaginative repertoires. Thus, Paul Benson writes that oppressive social-
ization "can erode competence at: rational consideration by restricting persons'
capacities for imagining with sufficient sensitivity and seriousness major alter-
ations in the prevalent gender system." 3 Aside from these few remarks, however,
the possible relationship between the various activities of the imagination and
the capacity for autonomy has not been the subject of serious investigation in the
literature. When one reflects on the extent to which our waking mental life is
taken up with different kinds of imaginative thought, this neglect is striking and
rather puzzling. My conjecture is that this neglect is due to a tendency to think of
critical reflection in overly rationalistic terms, at the expense of a recognition of
the extent to which critical reflection can be prompted by the imagination and by
emotion, desire, and bodily feelings.
This article is largely exploratory, but the exploration is guided by two central

124
 Imagining Oneself Otherwise 125

aims. First, I want to lay the groundwork for an investigation of the connection be-
tween autonomy and the imagination by investigating the role played by imagina-
tive thought in self-understanding, self-reflection, and practical deliberation about
the self. To focus my analysis, I concentrate on one kind of imaginative activity only,
namely, imagistic or representational thinking. However, the analysis could be ex-
tended to other kinds of imaginative mental activity, for example, mental conversa-
tions between an agent and an imaginatory interlocutor. My argument is that a va-
riety of different kinds of representational and imagistic thinking play a central but
often overlooked role in the processes of self-reflection and deliberation. This role is
an ambivalent one. The various activities of the imagination do not always promote
adequate self-reflection or ideal deliberation. The imagination can be, and often is,
delusional. But because of its affective force and cognitive power, imaginative men-
tal activity is crucial to the various processes by means of which we try to sort out
what we want; what matters to us; and what ideals, goals, and commitments shape
our lives.
My account of the connections among imagination, self-definition, and delib-
eration is developed in the first two sections of this article. In the first, I outline
Richard Wollheim's analysis of a person's point of view and explain the role played
by this notion in Wollheim's accounts of the mental activities of visually remember-
ing, imagining, and previsaging.4 In the second, I use Wollheim's analysis of a person's
point of view to provide an integration account of the process of self-definition. I
then argue that imaginative mental activity is central to this process.
The second aim of this article is to suggest that an understanding of the role
played by imaginative mental activity in self-reflection and in deliberation about the
self can provide some clues about the connection between failures of autonomy and
dominant cultural metaphors, symbols, images, and representations or, in short, the
cultural imaginary. Feminist philosophers, literary critics, anthropologists, and film
theorists, among others, have drawn attention to the way in which normative stereo-
types of gender relations and cultural fantasies of sexual difference are enforced and
perpetuated through metaphors, symbols, and visual representations.5 They have
also stressed the need for alternative representations by means of which women can
structure and understand sexual relationships and their own experience. Even in a
context of formal legal equality of opportunity, social reform has limited power to
reshape people's lives and opportunities if the cultural imaginary is predominantly
phallocentric.
One question that is raised by these analyses, however, is how cultural metaphors,
symbols, images, and representations can structure the experiences, self-concepts,
and gender identities of individuals. Psychoanalysis provides one kind of explana-
tion. In the third section, I sketch out an alternative explanation, one that is not hos-
tile to psychoanalysis but, rather than focusing on the formation of the unconscious,
focuses on the self-formative activities of self-reflection and deliberation. Drawing
on and extending the account of self-definition developed in the second section, I
argue that self-definition is not purely an introspective activity but also depends on
social recognition. Understanding the social dimensions of self-definition helps ex-
plain the connection, particularly in oppressive social contexts, between the imagi-
native projects of individuals and the cultural imaginary. The suggestion I develop is
126 Autonomy and the Social

that in self-reflection and deliberation, our own imaginative activities, our abilities
to imagine ourselves otherwise, draw on a cultural repertoire of images and repre-
sentations. When this cultural repertoire is predominantly phallocentric, the cultur-
ally available images on which women can draw seriously constrain their imaginative
possibilities and hence their self-conceptions.

Imagistic Thinking and Point of View

A great deal of our waking mental life consists of imagistic or representational think-
ing of a variety of kinds. In event or experiential memory, for example, we visualize
the event or experience we remember, sometimes as a series of unfolding scenes,
sometimes as a succession of disjointed snapshot images. Sometimes one particu-
lar image can come to stand for, or be emblematic of, a whole phase of our lives. In
future-directed anticipations, or previsaging, we represent our future selves to our-
selves, doing, thinking, and feeling what we believe we will be doing, thinking, and
feeling at that point in the future. And in imaginative visualizing, or what Richard
Wollheim calls iconic imagination, we imagine ourselves, either as ourselves or as
someone else, doing what we want or hope to be doing or being the kind of person
we might want to be.
Often we simply find ourselves engaged in a sequence of imagistic remember-
ing or imagining or anticipating, and because imagistic thinking has this involutary
and associational character we can be quite surprised by what we find ourselves
thinking. However, we can also engage voluntarily in these different kinds of repre-
sentational mental activities, for a range of purposes. We visually rehearse events in
the past to convince ourselves that they really happened or to try to understand their
significance; we try to visualize the face of a friend who has died or from whom we
are separated to remind ourselves of that person; we recall encounters with others to
interpret the meaning of their gestures. In prcvisagement or future-directed antici-
pation, we represent our future selves and activities to ourselves, perhaps because we
dread a possible future situation or something we will have to do. By actively prefig-
uring the future in this way, we try to prepare ourselves emotionally for what we
dread or to think out different ways in which we might be able to handle the situa-
tion. Or in other cases, we previsage the future because we might want to experience
a foretaste of anticipated future pleasures.
These examples bring out two important features of mental imaging of this
kind. First, whether it is voluntarily elicited or whether it surfaces involuntarily,
mental imaging has affective force. In visualizing the past, the activity of visualizing
very often reactivates at least some of the emotions that we originally experienced
and the bodily sensations kindled by them. For example, visualizing a past experi-
ence that was embarassing not only rearouses the embarassment but also our cheeks
may start to burn. We can find ourselves smiling with joy in anticipation of a future
encounter or close to tears as a result of some imaginative sequence, like imagining
the death of a child. It is this capacity to arouse and rekindle emotions arid bodily
sensations that contributes to the psychic force of mental imagery. The second fea-
ture of mental imaging of this kind is that partly because of its affective force, it has
cognitive power. We use mental imaging not only as a mnemonic device or to pro-
 Imagining Oneself Otherwise 127

vide us with pleasurable relief from the humdrum of our daily lives but also as an
aid in understanding ourselves and others.
In The Thread of Life, Richard Wollheim develops a detailed analysis of imag-
istic thinking to explain its affective force and cognitive power within our mental
and bodily lives. In this section I outline this analysis, focusing in particular on
Wollheim's notion of a person's point of view and on his account of the connections
and differences among iconic imagination, event memory, and previsaging. In the
next section I then draw on and develop aspects of this analysis to explain the role
played by imaginative mental activity in self-definition and self-understanding.

Centrally Imagining
In his discussion of iconic imagination Wollheim draws a distinction between acen-
tered and centered visualizing or imagining.6 When I am visualizing or imagining
acentrally, no point of view is represented in what I imagine. When I am imagining
or visualizing centrally, I am imagining from the point of view of someone who is
represented in the scene imagined. For example, I can acentrally imagine a scene in
which someone is receiving an Oscar. I might be able to imagine the scene in a great
deal of detail: I can visualize the audience, the auditorium, the face of the actor re-
ceiving the award, and the dignitary presenting it, as well as what they are both
wearing. But this visual sequence unfolds in my mind as though I am watching the
scene from a distance, uninvolved. By contrast, in centrally imagining, I represent to
myself the same scene but this time from the point of view of one of the characters
in it, the protagonist. For instance, I imagine that I am the famous actor who is re-
ceiving the award.
I want to highlight four features of Wollheim's account of centrally imagining
that will be vital to my discussion later. First, although in centrally imagining I rep-
resent the imagined scene from the point of view of the protagonist, the protago-
nist that I imagine need not be the real, empirical me. In some cases the protagonist
will be me, and I will imagine the scene from my point of view. In these cases, the
act of imagining will not require me to leave behind my character, dispositions,
ideals, or my specific body. Indeed, these may inform the nature of the scene visu-
alized, although the act of imagining may involve foregrounding, perhaps exagger-
ating, certain of my dispositions, characteristics, or capacities and deemphasizing or
overlooking others. However, I can also centrally imagine a scene from the point of
view of a protagonist who is not me. In these cases I am imagining being someone
else, for instance, I can imagine being Emma Thompson receiving an Oscar.
In imagining being someone else I suspend my point of view. I don't believe
that I am the person I imagine; I just suppose it.7 Wollheim explains how this is pos-
sible by likening centrally imagining to a kind of internal theater, in which the
imaginer is simultaneously internal dramatist, internal actor or protagonist, and in-
ternal audience. In the imaginative project, these roles are of course not distin-
guished: "we represent the internal narrative to ourselves even as we concoct it ...
and even as we respond to it."8 As internal dramatist, the imaginer assigns to herself,
as protagonist, lines and actions but also thoughts, feelings, and experiences. These
thoughts, feelings, and experiences not only provide the background that makes
128 Autonomy and the Social

sense of what she says and does but also provide the perspective through which the
other characters appear in the narrative. The other characters appear from her point of
view. As internal actor, the imaginer represents to herself the actions, lines, thoughts,
feelings, and experiences of the protagonist as though they were her own. The qual-
ification "as though" is the important distinguishing mark between centrally imag-
ining being someone else, where the point of view I represent is someone else's, and
centrally imagining myself, where 1 am the protagonist and where the point of view
I represent is my own.
In characterizing the internal audience, Wollheim argues that in the activity of
centrally imagining, the internal audience is akin to an empathic audience in a the-
ater rather than to either a detached or a sympathetic audience. On the one hand,
a detached audience is emotionally uninvolved with the characters, as in Brechtian
drama. Thus although it makes favorable or unfavorable judgments about the char-
acters, these judgments do not give rise to any affective response on its part. In con-
trast, a sympathetic audience is emotionally involved with the characters, and its
favorable or unfavorable judgments do give rise to the appropriate affective re-
sponses. But these responses arise from judgments about the worth or otherwise of
the various characters and their actions. On the other hand, the response of an em-
pathic audience is based on affect, rather than judgment, and in particular on an af-
fective response toward a particular character, whom it selects as protagonist and
with whom it emotionally identifies. This kind of emotional accord or resonance
on the part of the audience need not be accompanied by a favorable judgment of
the protagonist's actions or character. One can emotionally identify with a protag-
onist whom one may nevertheless judge to be despicable. In centrally imagining,
the imaginer is like an empathic audience toward her own imagining, and it is her
emotional identification with the protagonist she invents or imagines that plays the
decisive role in selecting that character as the protagonist of the imagining. Cen-
trally imagining being someone else is adopting another's point of view and emo-
tionally identifying with it.
The second important feature of "Wollheim's account of centrally imagining is
his insight that this kind of imaginative project has cogency. The mental activity of
representing to oneself the thoughts, feelings, and experiences of the protagonist, as
though they were one's own, will leave one in the condition—cognitive, amative,
affective—that one would be in were one actually to have the mental states one
imagines. It is its cogency that gives centrally imagining its psychic force. The degree
of cogency of an imagining will, of course, vary according to the extent of the
repertoire assigned to the protagonist. In centrally imagining myself, my repertoire
will be extensive, and hence the imagining is likely to have a grearer degree of co-
gency. In centrally imagining someone else, my repertoire is likely to be more im-
poverished and the imagining less psychically cogent. Representing to oneself some-
one else's point of view, then, involves assuming or adopting that person's repertoire,
to the extent that one can.
The third noteworthy feature of centrally imagining is Wollheim's distinction
between internal audience and internal observer. As internal audience, the imaginer
is parr of the imaginative project and is moved cognitively, conatively, and affec-
tively by the imagining in the same way as she would have been moved by the event
 Imagining Oneself Otherwise 129

that the imagining simulates. The internal observer, however, is outside the imagi-
native project and reflects on and reacts to it with emotions and judgments that may
be quite different than those aroused by the imagining. Wollheim gives as an exam-
ple an erotic daydream in which the imaginer may come to feel sexually aroused by
an imagining, but in reflecting on the daydream and the desire it has aroused may
react in a whole range of different ways—surprise, embarassment, disgust, or in-
creased arousal. Thus even when I am the protagonist of an imagining, and hence
even when the point of view represented in the imagining is my point of view, that
point of view need not coincide with my self-conception. I return later to the issues
of cogency and to the role played in self-understanding by the possible noncoinci-
dence between the point of view of an imagining and a person's self-conception.
Fourth, and finally, Wollheim makes the important point that we initiate cen-
tral imagining in the service of two distinct intentions—either to gain knowledge
or to experience pleasure or satisfy desire. The imagining will bear different relations
to our total body of beliefs, desires, and emotions, depending on which of these in-
tentions it serves. When iconic imagining expresses desire, the repertoire that is in-
voked in the imagining is intentionally restricted, and beliefs, desires, and emotions
that may conflict with the desires that are seeking satisfaction in the imagining are
filtered out "so that nothing in my psychology that could disrupt the pleasing effect
of imagining desire satisfied is afforded recognition."9 In other words, the informa-
tion we supply to ourselves about the background of the desire is minimized to en-
sure pleasure. By contrast, when we initiate imagining to gain knowledge about our-
selves or about our relations with the world, the repertoire on which we draw in the
imagining must be as rich and complex as possible if the imagining is going to yield
for us the kind of understanding we are seeking. Wollheim does not pay close at-
tention to the issue of how imagining can work in the service of self-understanding,
but I attempt an explanation in my account of the relationship between self-definition
and imagining oneself otherwise.

Point of View: Past, Present and Future

So far I have mentioned the role played by adopting a point of view in centrally
imagining, but I have not explicated what constitutes a person's point of view. This
is in fact the central concern of Wollheim's book. His analysis of a person's per-
spective or point of view is phenomenological in character; it involves understand-
ing what he calls the process of life, or what it is to live the life of a person. 10 On
his analysis, understanding this process hinges on understanding the relationships
among a person's present mental states and dispositions, the events that constitute
her past, and the future toward which she projects herself. These relationships are
not a matter of contingent connection or association. The past and the future inhere
in our present and they do so in two interrelated ways.11 First, because our present
perspectives or points of view trace out a particular temporal trajectory, they are
characterized by a present awareness of the influence of our past and a present con-
cern for our own future. To live the life of a person, as Wollheim puts it, is to live at
a crossroads between the present, into which we enter; our past, which influences us;
and our self-concern for our own futures.12 It is the internal relations among past,
130 Autony and the scoial
present, and future mental and bodily states, in particular an agent's awareness of
these relations, that account for the qualitative distinctness of her perspective or
point of view.
The second way in which past and future inhere in the present is in the content
of our present mental states, which include memories of past states and previsage-
ments of future states. Wollhcim argues that iconic representations, in the form of
experiential memory and previsagement, and in particular the psychic force of these
representations, play a pivotal role in giving rise to our sense of ourselves as creatures
persisting through time and to the sense that the various events that occur in our
lives are connected to ourselves in a noncontingem way. My later account of self-
definition draws on Wollheim's temporal account of a person's point of view and on
his explanation, outlined below, of the role played by experiential memory and pre-
visagement in constituting that point of view.
The crucial similarity between iconic imagination and experiential memory is
that because of their iconicity, they both have cogency and, in particular, affective
force. The activity of remembering an event in my life by representing it to myself
not only preserves my knowledge of the event but also rekindles the emotions asso-
ciated with that event, leaving me to some extent in a similar condition as the one I
was in when 1 experienced the event. Thus experiential memories preserve and trans-
mit the psychic force of the original events that they represent and in so doing bring
us under the influence of the past. Wollheim's explanation for this phenomenon
turns on a dispositional account of experiential memory. One of the ways in which
experiential memory differs from imagining is that a memory state stands in a de-
pendence relation to an earlier event. This dependence relation is a causal relation
in two ways. First, the memory state is causally related to an earlier event via a dis-
position. The disposition is an effect of the original event, and the disposition in
turn manifests itself in the memory state. Second, the role of the disposition is to
keep the event causally alive, to preserve and transmit its causal effect, and this role
is furthered by the manifestation of the disposition in the memory state. Thus "de-
pendence is at once an instance of causality and an agent of causality: it exemplifies
and diffuses one and the same causal influence."13 However, because the causal rela-
tion between earlier event and memory state is mediated by a disposition—which
has a history in the history of the person and stands in changing relations to other
dispositions, as well as to beliefs, emotions, and desires—the causal influence of the
eaiiier event is not always preserved in its original form; it changes over time. Fur-
thermore, the causal relation between disposition and occurrciit memory state is not
just a one-way relation. The manifestation of the disposition in an occurrent mem-
ory state can modify or reshape the disposition, altering its strength, for example, or
its content. Wollheim observes that "the feedback from mental state to disposition is
an essential element in the way in which we try to control the lives that we lead."14
One way in which this might occur is through the agency of the internal observer,
who may respond to the emotions elicited by the memory with quite different emo-
tions. I return to this point later.
Another important way in which experiential memory differs from iconic imag-
ination has to do with point of view. Experiential memories, Wollheim argues, are
 Imagining Oneself Otherwise 131

paradigmatically centered. However, unlike iconic imaginings, which can represent

from a point of view other than that of the imaginer, the only point of view from
which an experiential memory can represent is the point of view of the rememberer.
This follows from the dependence relation between the earlier event and the mem-
ory state: an experiential memory is a memory of an event that the person remem-
bering has lived through; it is a memory of her experience, and in remembering it,
the person remembers it as she experienced it, that is, from her point of view. Woll-
heim's arguments against the possibility or intelligibility of quasi memories, that is,
memories in which a person allegedly centrally remembers an event that another
person lived through, are important in illuminating what it is to have a point of
view.15 His arguments bring to the fore the way in which memories are not just
backward looking but also forward looking; that is, they exert the influence of the
past on the present and the future, and in so doing help constitute our points of
view.16 The central argument is that quasi memories are not consistent with the dis-
positional character of memories. Because dispositions stand in complex networks
of relations with other dispositions and with beliefs, desires, and emotions, for the
idea of quasi memories to be intelligible the following would need to be explained:
how to make sense of the idea that parts of this complex network could be trans-
ferred intact from one person to another; how to determine how much of this com-
plex network would have to be transferred from one person to another for a partic-
ular memory to be transferred; how to account for those of my current dispositions
that conflict with the network of dispositions within which the quasi memory is em-
bedded. Against the suggestion that we should think of quasi memories as akin to
centrally imagining being someone else, Wollheim points to a salient difference be-
tween imagining being someone else and imagining oneself. To imagine being some-
one else, I must select and adopt that person's repertoire—at least to the extent of
my knowledge. In imagining being myself, there are no such epistemic limits and no
assumption of repertoire is involved; I simply draw on my own repertoire. I want to
describe what Wollheim calls drawing on my own repertoire as "unselfconscious
identification" and to say that what distinguishes my point of view from another
point of view I might assume is that I identify with my point of view in an unself-
conscious way. In the case of memory, unselfconscious identification is explained by
two features of representational memory that are inconsistent with the assumption
of a repertoire: first, the causal link between disposition and mental state; second,
the involuntary character of much remembering.
Whereas experiential memory reveals the extent to which our present is shaped
by the influence of the past, previsagement is a manifestation of self-concern, that
is, of our concern for our own futures. As with experiential memory, in centered
previsagement I identify unselfconsciously with the protagonist or subject of the
representation; I can only centrally previsage my future from my own point of
view. However, previsagement differs from experiential memory in two significant
ways, which reflect the asymmetry of past and future. First, previsagement involves
no dependence relation: the future has no causal hold over the present, and there
is no direct causal relation between the present representation and the future that
it represents. Second, and following from the first, whereas in experiential memory I
132 Autonomy and the Social

represent to myself a past that I have experienced, in previsagement I can and very
often do represent to myself a future that does not take place. In this respect, pre-
visagement is more akin to iconic imagination than to experiential memory. A fur-
ther feature that previsagement shares with both experiential memory and iconic
imagination is its affective force. Because previsagement is a manifestation or mode
of self-concern and because of its iconicity, the activity of previsaging our own fu-
tures has the power to arouse in us the kinds of emotions we would feel were we to
experience the events we previsage. I return to these issues later.
In addition to Wollheim's three modes of imagining from one's own point of
view—centrally imagining, experiential memory, and previsagement—I want to
suggest two others that are closely akin; in fact, they are species of iconic imagin-
ing. The first, which 1 call countcrfactual speculation, stands somewhere between
experiential memory and iconic imagination. In countcrfactual speculation, we
speculate about the course our lives might have taken had certain features of those
lives been different. In these kinds of speculation, we imagine that at some point
in the past, for example, at the juncture at which an important decision was made,
the course of our lives diverged from the course that our lives have actually taken.
From that point on, we imagine a different past for ourselves and hence a different
present and future. What characterizes this mode of imagining is that although
we imagine ourselves living a different life, we nevertheless identify unselfcon-
sciously with the protagonist of the imagining; the life we imagine is still recog-
nizably our own.
The second mode of imagining, which I call future-directed fantasy, stands some-
where between previsagement and iconic imagination. In this mode of imagining,
1 represent to myself a future I might have or might have had. So, once again, I
imagine from my point of view a life that is still recognizably mine but one that di-
verges at some point in the future from the life I am likely to lead or believe 1 will
lead. Thus an adolescent fantasizes that her talent for acting will be spotted; whether
or not she has such a talent is irrelevant to the fantasy, and she imagines a future life
of fame and pleasure, without really believing that such a future is a genuine possi-
bility for her. The difference between counterfactual speculation and future-directed
fantasy, on the one hand, and experiential memory and previsagement, on the other,
is therefore that they manifest different dispositions. In the latter case, the represen-
tation of my past or my future is accompanied by a belief that what I represent did
happen to me or is likely to happen to me. In the former case, I just suppose that
this might have been my past or that this might be my future, without believing that
what I represent was or will be a genuine possibility.
I have discussed the role played by different kinds of mental imaging in the
phenomenology of a person's perspective or point of view. Following Wollheim, I
have characterized the point of view of a person as a network of interrelated emo-
tions, beliefs, desires, and mental and bodily states and dispositions, shaped by the
influence of the past and directed by self-concern for the future. I have characterized
the phenomenological distinctiveness of that point of view as unselfconscious iden-
tification with that network. In the following section I provide an account of self-
definition that centers on the triadic relationships among a person's point of view;
her self-conception; and her ideals, values, commitments, and cares. I then explain
 Imagining Oneself Otherwise 133

the role played by the different modes of imagining in the process of self-formation
or self-definition.

Self-Definition, Deliberation, and Imagination

Diana Meyers has argued that self-definition is one of the capacities necessary for
autonomy.17 I develop, concurrently, three interrelated suggestions concerning self-
definition. The first is that the process of self-definition should be thought of as a
process of negotiation among three related but distinguishable elements of the per-
son: her point of view; her self-conception; and her values, ideals, commitments,
and cares, in short, what matters to her. The aim of this process of negotiation is to
achieve a kind of reflective equilibrium among these different elements of the self.18
A reflective equilibrium is achieved when these elements are integrated in a relatively
stable way, that is, when there are not serious and persistent conflicts among them.
Thus, I am proposing an integration account of self-definition. I regard the kind of
integration I describe as a necessary, if not a sufficient, condition for autonomy.19
Second, I characterize the process of formation of a person's self-conception as a pro-
cess whereby, through reflection guided by her values, ideals, commitments, and
cares, a person constitutes certain elements of herself, or certain features of her point
of view, as external to herself while appropriating others. Third, I argue that the var-
ious modes of imagining already described play an important but overlooked role
in the formation of a person's self-conception and in the process of achieving an in-
tegration among the different elements of the self.

Integration and Self-Definition

Following Wollheim, I have characterized a person's point of view as a network of
interrelated emotions, beliefs, desires, and mental and bodily traits and disposi-
tions, shaped by the influence of the past and directed by self-concern for the fu-
ture. A person's self-conception delimits that part of the network that the person re-
gards as defining herself or with which she identifies. Wollheim's metaphors of the
empathic internal audience and the reflective internal observer help clarify the re-
lationship between a person's point of view and her self-conception. Recall that the
difference between internal audience and internal observer is that whereas both are
internal, they stand in different relations to the imaginative project. The internal au-
dience is caught up within the imaginative project and emotionally identifies with it.
The internal observer, on the other hand, stands outside the imaginative project;
may react to it emotionally in a range of different ways; and reflects on and evaluates
the desires, emotions, and beliefs represented or manifested in the imagining.20
Elaborating from these metaphors, we can identify two salient features of the re-
lationship between a person's point of view and her self-conception. First, like inter-
nal audience and internal observer, both my point of view and my self-conception
are internal to me in the sense that both are constitutive aspects of my identity. To
characterize a person's identity solely in terms of one of these aspects rather than
both would be mistaken because both motivate our actions and our responses. Thus,
on the one hand, our actions are often motivated, in ways that we either may not be
134 Autonomy and the Social

aware of or cannot change, by mental or bodily dispositions and habits, desires,

character traits, and so on that do not accord with our self-conceptions.21 On the
other hand, part of what it is to be an agent is to be able, at least within certain lim-
its, to modify oneself in line with one's self-conception.
Second, although both our points of view and our self-conceptions arc consti-
tutive of our identities, like the internal observer we shape our self-conceptions by
reflecting on different aspects of our points of view. This reflective activity can be
thought of as a process of externalizing, distancing, or dissociating ourselves from
certain aspects of ourselves and appropriating others. 22 The activity of externalizing
an clement of oneself, for example, a desire, involves reconfiguring the network or
structure of beliefs, emotions, dispositions, desires, and so on that constitute one's
point of view in order to lessen, perhaps eventually eliminate, the motivational pull
of this desire. So to externalize an clement of oneself is to reject it as a motivating
factor in one's actions. To appropriate an element of onself is to acknowledge and
accept it as one of the aspects of one's identity that guides one's actions. It is in this
sense that we identify ourselves, or our self-conceptions, with those aspects of our-
selves that we appropriate. In some cases, appropriation may involve endorsement.
But appropriation may also be consistent with equivocal endorsement or in some
cases just with acceptance.
The metaphor of the internal observer brings into focus two central features of
externalizing and appropriating, namely, that they involve both emotional response
and evaluation. In reflecting on and reacting to our own desires, habits, character
traits, and so on, we respond to these affectively—with pride, shame, or embarass-
ment, for example. These emotional responses are both guided by and shape our
self conceptions. So, for example, we might feel shame when we find ourselves act-
ing on a desire that conflicts with our self-conception. But our emotional responses
and our self-conceptions are in turn guided by our valuations and our judgments
about what matters. Several philosophers have argued that our identities or self-
conceptions are shaped by what matters to us. Thus Williams talks of a person's
identity as being bound up with her "ground projects" and commitments; 23 Ronald
Dworkin aligns the person with her critical, as opposed to experiential, interests,
that is, with those interests that incorporate a person's critical or reflective judgments
about what constitutes a good life, what is important, what is worth doing and valu-
ing, what ideals she should live her life by, and so on;24 and Frankfurt suggests that
the boundaries of a person's will are defined by what she cares about.25 For the pur-
poses of this discussion, I ignore some of the philosophical differences among these
accounts and draw on some of the intuitions that unite them in order to clarify the
evaluative dimension of externalizing and appropriating.
Two central intuitions are shared by these views. The first is that our pro-
jects, cares, values, or critical interests guide the process of distinguishing our
self-conceptions from our points of view. The judgment of whether we wish to ap-
propriate or dissociate ourselves from a desire, disposition, or characteristic is related
to our assessment of whether this element of our psychology strengthens or under-
mines our commitment to these cares, values, or critical interests.26 So the formation
of a person's self-conception could be understood as the attempt to bring the differ-
 Imagining Oneself Otherwise 135

ent aspects of the self into line with what matters to her. At the same time, making
judgments about whether we wish to appropriate or dissociate ourselves from cer-
tain aspects of ourselves is a process that also involves engaging in a reflective assess-
ment of our values, ideals, cares, or in short what matters to us. In reflecting on
whether a certain desire or set of desires, for example, undermines our commitment
to what we care about, we are also reflecting on whether what we care about is worth
caring about or worth caring about in the way or to the extent that we do.27
The second intuition is that what we are and what matters to us are not simple
matters of choice. To say that what we are is not a simple matter of choice is to say
that the network or configuration of emotions, beliefs, dispositions, and desires that
constitute our points of view is, to some degree, not voluntary. Our identities are
shaped in fairly determinate ways by our various characteristics, by the relations be-
tween these characteristics and our social context, and by what matters to us. To say
that what matters to us is not a simple matter of choice is that say that to a certain
degree, we just find certain things mattering to us. This may be because we are dis-
posed in certain ways by the manner in which different aspects of our identities, for
example, our temperament and talents, reinforce one another; what matters to us
may be connected with commitments to others, for example, parents, that are not
entirely of our choosing;28 or, what matters may be the result of significant events in
our particular histories or of decisions we made in the past that are now no longer
a matter of choice. Thus, we cannot simply choose to abandon our cares or to give
up what matters to us. Or rather, we cannot do so without forfeit or loss. Certainly
what matters to a person may change, perhaps because of a decision she has made
or because of an event or action that has intervened to disrupt the reflective equilib-
rium she had established among different aspects of herself.29 But something that
has mattered usually cannot simply cease to matter. It can only do so, or come to
matter in a different way, as a result of a process of readjustment of the elements of
the self.
However, although what we are and what matters to us are not under our im-
mediate voluntary control, this should not be taken to imply that we are passive
with respect to ourselves. Self-definition or self-formation is a matter of actively ne-
gotiating the relationships among one's point of view, one's self-conception, and one's
values. A reflective equilibrium among the different aspects of one's self is achieved
when these elements are integrated in a relatively stable way, that is, when there are
not serious and persistent conflicts among them. The notion of stability does not
imply that there cannot be tension or inconsistency within or among the different
elements of the self, that the self is seamless. Nor does it imply that an integrated self
is static. In fact the process of self-integration is an ongoing and dynamic process
precisely because of inevitable tensions and inconsistencies within the self and be-
cause the different elements of the self are constantly undergoing transformation.30
The notion of stability does imply, however, that an agent who is persistently inter-
nally divided or whose sense of self is seriously fragmented cannot achieve the kind
of reflective equilibrium necessary for unified agency.31 By unified agency I mean
the kind of practical unity necessary to deliberate, make decisions and choices, and
act. 32
136 utonomy and the Social

Imagination, Deliberation and Self-Definition

For the remainder of this section I want to develop the idea that imaginative men-
tal activity plays a crucial role in self-definition, as I have characterized it, and in the
deliberative processes that precede self-formative decisions. To explain this role, I
want to recall two points that were made earlier in the discussion. The first is that
mental imagery, because of its iconicity, has psychic force or cogency; it is able to
rearouse emotions or simulate the effect of emotions. The second is that in dis-
cussing the relationship between a person's point of view and her self-conception, I
likened this relationship to that between Wollheim's internal audience and internal
observer. This comparison brought to the fore two central features of the kind of re-
flection that is involved in negotiating the relationships among our points of view,
our self-conceptions, and what matters to us, namely, that reflection involves both
an affective and an evaluative response.
Putting these two points together, we can begin to see the role played by the var-
ious modes of imagining in the process of self-definition. By virtue of its power to
rearouse or simulate emotions, imaginative mental activity initiates self-reflection by
prompting an emotional response and, through that, an evaluative judgment. The
cognitive power of this process is reasonably evident in the case of experiential
memory. Because experiential memories are representations of experiences we have
lived through, such memories can enable us to understand and reshape ourselves
by enabling us to understand the ways in which our present points of view, self-
conceptions, values, ideals, and commitments have been influenced by our past. The
memories rearouse the emotions associated with the original experience. But the in-
terval between the past and the present and the way in which the web of disposi-
tions within which the memory is embedded has been modified by the history of
the person during that interval, enable the person to respond to the original emo-
tions with a further set of emotional reactions, which I call reactive emotions. These
reactive emotions then prompt evaluative judgments, which enable the person to
gain some understanding of the significance of past experiences and to gain some
measure of control over the psychic force of those experiences. For example, the
process of sifting through and evaluating memories, externalizing some and appro-
priating others, is one way in which a person can come to terms with a traumatic
event in her past and reestablish some kind of equilibrium among the different ele-
ments of herself.33
However, il might be thought that if there is some truth to the claim that
iconic mental representations have cognitive power and motivational force, if it is
true that they are indeed able not only to yield self-understanding but also to inau-
gurate a reconfiguration of the various elements of the self, that this will be true
mainly of those cases of imagining that are accompanied by belief, namely, experi-
ential memory and previsagement. On the other hand, it might seem that because
they typically manifest desire, iconic imagining, future-directed fantasy, and coun-
tcrfactual speculation are modes of imagining that are more likely to be invoked in
the service of self-deception than as an aid to self-understanding. Although I try to
show that imagining of these varieties can contribute in a positive way to practical
deliberation about the self, this objection certainly has force. The imagination can
 Imagining Oneself Otherwise 137

be delusional, and its role in self-understanding is ambivalent. There are a number

of ways in which imagining can block, rather than aid, self-understanding and prac-
tical deliberation about the self. The first, discussed previously, is when the imaginer
intentionally restricts the repertoire invoked in the imagining, so that no beliefs, de-
sires, or emotions that might threaten the pleasure she seeks gain admittance to
the imaginative project. This kind of restriction of the imaginative project can, of
course, also occur when experiential memory and previsagement are invoked in the
service of desire. The second way is when a person's imaginings and the repertoire
she invokes in her imaginative projects are limited by the culturally available images
and representations that provide the raw material on which we draw in our own
imaginative activities. I turn to this issue later.
To understand the role that the species of imagining that includes counterfac-
tual speculation, future-directed fantasy, and iconic imagination can play in self-
definition, I want to focus a bit more closely on the emotional response prompted
by an imagining. My emotional response to an imagining involves a number of in-
terwoven components, including my response to the content of the representation,
my response to the emotions that I perceive to have been aroused or simulated in me
by the representation, and my response to what I perceive the representation and the
emotions it generates in me disclose about me. For example, let us say that I imagine
myself being told of the death of someone I believe I love. This imagining might
arise involuntarily or it might be invoked deliberately to represent to myself my own
feelings for this person. But let us say that as the imagining unfolds, I unexpectedly
find myself representing myself, as the protagonist, as experiencing somewhat mixed
emotions to this event: dread, grief and sorrow, to be sure, but also relief. As internal
audience, the imagining simulates in me the effect of these different emotions. As
internal observer, I respond to these different emotions with reactive emotions. My
reactive emotions include those that are in accord with the emotions simulated by
the imagining, for example, a heightening of sorrow and dread. These reactive emo-
tions are direct responses to the content of the representation, that is, to the repre-
sentation of the loss of the loved person, and they are also responses to the emotions
simulated by the imagining. Furthermore, since I believe I love this person, these
emotions concur with and confirm my self-conception. However, in this case my re-
active emotions will also include a response to the apparently anomalous emotion
simulated by the imagining, the emotion of relief, which takes me by surprise as the
imagining unfolds. My reactive emotions to this emotion may include bewilder-
ment, shame, and anxiety. These reactive emotions will in turn prompt self-reflec-
tion and evaluative judgment—of the imagining and the anomalous emotion of re-
lief represented in it, of the various reactive emotions that arose in me in response to
this simulated emotion, and of my self-conception and my sense of what matters.
The judgment I arrive at may be that this simulated emotion is indeed anomalous
and external to me, and so the judgment confirms my self-conception. However, it
may be that this imagining sets in train, or at least forms part of, a process of com-
ing to realize that I do not love this person any more, and so of reconfiguring my
self-conception and my sense of what matters.
This example shows that in imagining, we present aspects of ourselves to our-
selves for reflection and evaluation, not only directly or indirectly in the content of
1 38 Autonomy and the Social

the representation, but also because their cogency enables such representations to
provide a window into our own emotional states, our points of view, and our self-
conceptions. In those cases in which imagining is invoked in the service of self-
knowledge, we allow as much knowledge of ourselves as we can into the representa-
tion in order to make this window as wide as possible. In those cases in which
imagining is invoked in the service of desire, we narrow this window, sometimes to
guarantee that the imagining will afford maximum pleasure, sometimes to ensure
that we see of ourselves only what we want to see. Thus we manipulate our imagin-
ings so that they conform to, or at least do not conflict with, our self-conceptions.
However, even in these cases, we very often find ourselves unable to control the
imagining to the extent that we want—beliefs, desires, and emotions that conflict
with the intention of the imagining can creep in as the imagining unfolds and can
provide a sometimes unwanted window into those aspects of ourselves that we
would rather not see.
The example also shows that representational imagining can provide this kind
of window into the self because it can abstract us from our habitual modes of un-
derstanding ourselves and our relations with others. By putting ourselves at a remove
from these habitual modes of understanding, we are able to reflect: on and evaluate
them, and so to test: our satisfaction with them. But this is not the only way in which
representational imagining can aid self-understanding and self-definition. By re-
moving us from the habitual, imagining also opens up a space within which we can
try out different possibilities for ourselves—different possibilities of action, desire,
emotion, and belief. This trying out of different possibilities or postures of the self
is a central feature of countcrfactual speculation and future-directed fantasy. In as-
suming different postures, say in imagining a different past for ourselves or fantasiz-
ing about the future, we hold certain elements of ourselves stable and play around
with others. Thus we place in the foreground certain aspects of our identities, for ex-
ample, certain ideals, characteristics, or talents, and downplay others. As internal ob-
server, we then respond, emotionally and evaluatively, to these alternative represen-
tations of ourselves. Through this process we start distinguishing those possibilities
that may be genuine possibilities for us from those possibilities that are not: really
thinkable for us at all. An obvious example is the kind of previsaging of different
possibilities that we engage in when we are trying to make a decision between alter-
native courses of action—for example, choosing between two jobs in two different
cities. In making decisions of this kind, much of our deliberative activity involves
imaginative projection. We represent to ourselves the different kinds of life we be-
lieve we would live, given the different options, and by evaluating our responses to
these representations we gradually get ourselves into a position to make a decision.
For reasons that I have already indicated, this imaginative playing around with
our identities, or imagining ourselves otherwise, does not always promote adequate
self-reflection or ideal deliberation. In the case of previsagement, counterfactual spec-
ulation, and future-directed fantasy, an additional problem is that these kinds of
imagining can be more or less impoverished. Our ability to imagine a different past
for ourselves, for example, peters out fairly quickly because we simply do not know
enough about what our lives would have been like had they taken a course such as
the one we imagine. Similarly, although in the kind of deliberative previsagement
 Imagining Oneself Otherwise 139

that precedes decisions, we build in as much knowledge of ourselves and of the dif-
ferent possibilities as we can to make the imagining as rich and informative as pos-
sible, in retrospect these imaginings often strike us as impoverished. Thus, after the
event, we can find ourselves thinking that we might have made a different, and bet-
ter, decision if only we had known ourselves better or taken into account a consid-
eration that in retrospect seems so obvious or predicted an unforeseen outcome of
our actions.
Nevertheless, bearing in mind these reservations, I contend that our ability to
imagine ourselves otherwise—that is, our ability to imaginatively distance ourselves
from our habitual modes of self-understanding and to envisage, in imaginative rep-
resentations, alternate possibilities for ourselves—plays an important role in practi-
cal reflection and deliberation about the self, and hence in self-definition.

Imagination, Social Recognition, and the Cultural Imaginary

From the way I have been describing imaginative projection so far it may seem as
though our individual lives are completely discrete and self-contained; as though our
mental and bodily lives, our memories, our self-concern for our own futures, and
the repertoires on which we draw in our imaginings refer only to our own experi-
ences and points of view; and as though these exist in some kind of vacuum. But,
of course, this is not so. We are social creatures who are formed and transform our-
selves in our intimate and nonintimate relationships with other people.34 We be-
come persons and live our lives in particular social, cultural, and historical commu-
nities. Our sense of our lives as temporal, our points of view, our self-conceptions,
and our values, are therefore shaped by these relationships and these communities.
So, too, are our imaginative mental activities. So how can the relationship between
our individual imaginative projects and the social world in which we live our lives be
characterized? Or, since this is a rather large question, let me narrow it by asking
how we can understand the relationship between our own imaginative representa-
tions of ourselves and cultural images and representations. In particular, how can a
person's imaginative projects, and hence her capacity for self-transformation and au-
tonomy, be stymied by the dominant cultural imaginary? These are also very large
questions that cannot be explored here adequately, but I want to sketch out one pos-
sible answer. The first part of my answer involves an analysis of the role played by
social recognition in self-definition and autonomous agency. The second part draws
some connections between social recognition and the social imaginary.

Social Recognition and Intrapsychic Integration

Previously, I suggested that self-definition should be understood as a process of at-

tempting to achieve a dynamic integration of the different elements of the self. Here
I want to argue that social relationships more generally, and social recognition in
particular, play a crucial role in achieving this kind of integration, and hence in
achieving autonomy. There are two crucial ways in which social recognition is nec-
essary for self-definition. First, it is necessary because self-knowledge is crucial to
self-definition and because we achieve self-knowledge in social relationships. Self-
140 Autonomy and the Social

knowledge involves knowing which aspects of one's point of view—which desires,

characteristics, traits, and so on—one identifies with; knowing what one values and
cares about; and knowing how one feels and what one wants. It is precisely this kind
of self-knowledge that is lacking in the case of agents whose self-conceptions are
fragmented or who experience persistent conflict between different aspects of them-
selves. Most of the time this kind of self-knowledge is tacit and taken for granted.
However, in crises or when we face difficult decisions, not: only does it need to be-
come more explicit but also our knowledge of ourselves and our self-conceptions are
often seriously challenged. 35 Thus an enlarged or altered self-understanding may
emerge during the process of deliberating and deciding—if all goes well. If all does
not go well and self-knowledge only emerges after a decision has been made and ac-
tions undertaken, the result is often serious regret. Regret reveals the importance of
self-knowledge to self-definition. An agent who persistently regrets her decisions is
an agent whose capacity for autonomy is significantly compromised.36 However, the
affective, evaluative, and imaginative processes of reflection—by means of which
we clarify what we value, distinguish our self-conceptions from our points of view,
and so achieve self-knowledge—cannot be purely introspective. Our emotional re-
sponses to aspects of our identities such as our temperamental characteristics—
responses, for example, of shame or pride—are shaped by and responsive to the es-
timations and responses of others. And these responses, at least in part, form the
basis for our judgments about ourselves. Thus we come both to know and to define
ourselves in our interactions with others.37
Second, social recognition is necessary for self-definition because a sense of self-
worth is necessary to the achievement of a reflective equilibrium among the differ-
ent aspects of the self and because self-worth is bound up with social recognition. In
what follows I want to elaborate this second point by explaining the connections
among integration, self-worth, and social recognition. 38 Having a sense of self-
worth is necessary for integrated and self-defining agency for at least two reasons.
One reason is that agents are motivated to act only if they have a conception of their
actions as effective, as making a difference. This in turn requires diat agents have a
conception of themselves as capable of effective action—as having the necessary ca-
pacities, talents, and attributes. Without a sense of her own worthiness as an agent
and of the worthiness of her capacities, her desires, and her beliefs, an agent will not
be able to conceive of herself as capable of effective action.39 Another reason is that
an agent's capacity to resolve internal conflicts is tied to whether or not she thinks
that what she does or thinks or feels makes a difference. It is also tied to her sense of
whether or not she thinks that it is important for her to be able to define and express
herself, in other words, with whether or not she thinks that who she is matters. If an
agent has little or no sense of self-worth, she will think that who she is or what she
does makes no difference, and hence she will have no motivation for resolving inter-
nal conflicts, for trying to establish a reflective equilibrium among the different ele-
ments of her self.
However, having a sense of self-worth is not an all-or-nothing matter. Since our
identities are complex, an agent may have a reasonably strong sense of self-worth
with respect to certain aspects of herself, for example, certain talents or capacities,
 Imagining Oneself Otherwise 141

but a rather fragile sense of self-worth with respect to other aspects of herself, for ex-
ample, her physical appearance or her ability to form intimate relationships.40 More
important, self-worth is not an all-or-nothing matter because self-worth is funda-
mentally social. Agents live their lives in a number of overlapping but distinct social
spheres, including the spheres of intimate interpersonal relationships—such as fa-
milial relationships, love relationships, and close friendships—as well as those of
nonintimate social contacts and acquaintances; the sphere of work; the spheres of
sport, social clubs, and artistic pursuits; the spheres of group-based ethnic or com-
munity identities and social life; the spheres of political activity and participation;
and so on. These different spheres bring out different, sometimes conflicting, as-
pects of agents' identities and reinforce or undermine these aspects. An agent's sense
of the worthiness of different aspects of her identity is bound up with the extent to
which the social sphere in which those aspects are salient reinforces or undermines
the relevant aspect of her sense of self-worth.
Another way of putting this point is to say that an agent's sense of self-worth is
bound up with social recognition. I would suggest that there are three intercon-
nected but distinct types of social recognition. The first, most fundamental type,
which is obviously Kantian in flavor, involves recognition of the agent as a human
being worthy of respect. It involves recognition of the agent as a person whose life
matters to her and to others and as a being capable of feeling, thought, and self-
defining agency. It is this kind of recognition that is violated by extreme forms of
oppression, such as slavery or genocide, and by acts of violence committed by oth-
ers, whether in their impersonal forms in warfare or in their more direct, personal
forms in rape, assault, abuse, and murder. The power of this kind of violation is at-
tested to by Susan Brison's remark that "one assailant can undo a lifetime of self-
esteem."41 The second type of social recognition involves recognition of the worth
of the social group to which the agent belongs—where social groups may be de-
fined in terms of class, racial, ethnic, gender, cultural, or religious identity or inter-
sections of these.42 This second dimension is clearly much more tied to sociopoliti-
cal structures and to social norms and expectations than the first. Systematic denial
of social recognition of this kind tends to characterize oppressive social relations.
The third type of social recognition operates at a more directly personal level and in-
volves recognition of the worth of different aspects of an agent's identity, including
her talents and capacities, mental and bodily traits and dispositions, emotions and
desires, temperamental characteristics, and so on; recognition of the worth of her
self-conception; recognition of her values, commitments, and ideals; and recogni-
tion of the worth of her social relationships. One of the consequences of a lack of
social recognition of this kind is that it undermines an agent's sense of her own wor-
thiness at a personal, social, and political level. I suggest below that in oppressive so-
cial contexts, this third kind of social recognition is systematically withheld from in-
dividual members of oppressed groups or is available to them only in very truncated
and restricted forms.
I have argued that a sense of self-worth is necessary for integrated agency, and
hence for autonomy, and that self-worth is tied to social recognition of all three
kinds. However, the relationship between social recognition of the third kind and
142 Autonomy and the Social

self-worth is ambivalent as far as autonomy is concerned. On the one hand, because

self-worth is tied to social recognition, we emotionally invest ourselves in those as-
pects of our identities and self-conceptions and those social spheres that provide so-
cial recognition and that reinforce our sense of self-worth. This investment provides
the kind of direction needed for autonomy. On the other hand, this kind of invest-
ment of our self-worth can be perilous and undermining of autonomy because it
leaves us highly vulnerable to changes in our personal and social circumstances; wit-
ness the profound disorientation experienced by many men upon retirement, if
their sense of self-worth has been overly invested in work, or by many women when
their children grow up and leave home.
It is perhaps because of this ambivalence that philosophers have traditionally
thought of autonomy as requiring independence from social norms, expectations,
and recognition. I think this is a mistaken inference. Rather, the ambivalent rela-
tionship between self-worth and the third kind of social recognition illuminates
the importance for autonomy of a robust sense of self-worth, that is, a sense of
self-worth that is not overly dependent on a narrow range of forms of social recog-
nition and that is not overly invested in a narrow range of attributes, capacities, re-
lationships, and so on. For example, the discovery of infertility, at least in our so-
ciety, is commonly extremely devastating to a woman's general sense of self-worth.
But just how far-reaching and generally debilitating this feeling is depends on the
extent to which a woman has invested her sense of self-worth in maternity. This in
turn will depend on the importance of maternity within the central social spheres
of her life, on what other social opportunities are available to her, and on whether
there are other meaningful spheres of activity in her life from which she can gain
social recognition. 45
A diversity of forms of social recognition is typically absent in contexts of op-
pressive socialization, fn such contexts, only certain aspects of the identities of
agents are afforded recognition. Furthermore, their scope for self-defining agency is
restricted to a limited range of social spheres. Given the connection between self-
worth and social recognition, the effects are obvious. One effect is that the achieve-
ment of self-worth will be overly invested in certain aspects of the self and in certain
social relationships, namely, those that conform to social norms and seriously re-
stricted social roles, with the result that other aspects of the self are repressed or
stunted. Some agents who arc able to conform their self-conceptions to these re-
stricted norms and roles may in fact be quite integrated as agents and have a reason-
ably strong sense of self-worth—which is why integration is a necessary, but not a
sufficient, condition for autonomy.4'' More commonly, however, the effect will be
some degree of internal conflict, alienation, or fragmentation of the self, as agents
struggle to achieve a sense of the worthiness of their capacities in social contexts
where there is restricted scope for their expression and where these are afforded lit-
tle social recognition. In this kind of context it is not surprising that agents may vol-
untarily and apparently rationally make choices, like the choice to be a surrogate
mother or to have a nose job or to undergo a sex-change operation, that promise to
resolve these conflicts and provide them with greater social recognition. It should
also not be surprising that the attempt to achieve integtation by such means does
not always succeed or enhance the agent's autonomy.
 Imagining Oneself Otbenvise 143

Social Recognition and the Cultural Imaginary

For the remainder of this article, I sketch an explanation of how a restricted or op-
pressive cultural imaginary may limit an agent's capacities for imaginative projec-
tion, and in so doing impair her capacities for self-definition, self-transformation,
and autonomy. Previously, I suggested that in imagining from someone else's point
of view we assume someone else's repertoire. In imagining from our own point of
view, we draw on our own repertoire—that is, on our beliefs and desires and our
own experiences as these have been sedimented in memories, mental and bodily dis-
positions, and habits. Imagining involves elaborating from this repertoire. I have
also argued that our points of view and those of others are developed in the context
of particular social relationships and particular social contexts. Furthermore, our
sense of self-worth and our self-conceptions are shaped by, and responsive to, social
recognition. One of the most important ways in which social recognition is ex-
pressed or withheld is through cultural images, representations, symbols, and meta-
phors. These representations reflect, incorporate, and instantiate, often in sublimi-
nal but nevertheless powerful ways, social and cultural understandings of agents'
worth, especially understandings of the significance and worth of various kinds of
identities—such as gender, ethnic, racial, and religious identities.45 Drawing these
points together, I want to make three suggestions to explain the connections among
the cultural imaginary, an agent's imaginative projections, and the impairment of
autonomy in oppressive social contexts.
First, the repertoire on which we draw in our imaginary self-representations is
mediated by the available cultural repertoire of images and representations. The so-
cial psychic force of this imagery mirrors the psychic force of mental imagery; it
shapes our emotions, our desires, and our beliefs. So the medium by which the cul-
tural imaginary informs our beliefs and shapes our desires is through our own rep-
resentations; cultural imagery latches on to the individual psyche, as it were, by grip-
ping the imagination.
Second, the ambivalent role of cultural imagery vis-a-vis our imaginations mir-
rors the ambivalent role of imagining in self-understanding. On the one hand, I
have argued that the activity of imagining can abstract us from our habitual modes
of understanding and open up a space within which to envisage new possibilities of
self-definition and self-understanding. Imagining can do this because we are not re-
stricted to representing only what we actually are or what we think is actually possi-
ble. We can represent what we might want to be, what we wish to be possible, or just
what might be or might have been possible. This is what is liberating about imagin-
ing. Innovative cultural imagery plays a similarly liberating role. In representing
what might be possible, it abstracts us from our habitual understandings of our-
selves and others and so begins to loosen the grip of dominant imagery—which is
why alternative representations of gender relations and sexual difference have been
so vital to feminist efforts to restructure our social understandings. On the other
hand, representations can act like compulsions to constrain the imagination, enforce
habitual patterns of thought, and stymy self-understanding and self-definition. I
suggested earlier that imaginative projects that are initiated in the service of desire
can have this effect. The imaginer deliberately rules out beliefs, emotions, and de-
144 Autonomy and the Social

sires that might conflict with the desire that seeks satisfaction in the imagining ot
that might conflict with her self-conception, and in so doing narrows the window
that the imagination can open into the self. Similarly, cultural imagery enlisted, con-
sciously or unconsciously, in the service of the desire for domination or the desire
to perpetuate the status quo draws on a curtailed cultural point of view to restrict
the repertoire in terms of which the culture can represent itself.46 When these re-
stricted cultural representations grip the imaginations of individuals, the effect is to
narrow the range of the repertoires on which we can draw in our imaginative proj-
ects and so to curtail our imaginative explorations of alternative possibilities of ac-
tion, emotion, belief, and desire. Having a restricted repertoire is, of course, quite
consistent with having a vivid imagination. There arc all sorts of tediously repeti-
tious imaginative permutations and combinations that can be elaborated from a sin-
gle theme, as Hollywood knows all too well.
Third, given the connection between an agent's sense of self-worth and social
recognition, there is a strong incentive for agents to identify with those cultural rep-
resentations of their identities that seem to afford greater social recognition and to
incorporate these representations into their self-conceptions and their imaginative
projections. Even if these representations are oppressive, in the sense that they pres-
ent agents with severely curtailed avenues for achieving social recognition, the fact
that these avenues afford the main means of achieving social recognition neverthe-
less provides agents with a strong incentive for identifying with them. It may also
provide them with a strong incentive for resisting innovative cultural imagery. Thus,
whereas oppressed agents may have more or less rich fantasy lives within a restricted
repertoire, their desires and capacities to seriously imagine alternative possibilities for
action, emotion, and desire, that is, to seriously imagine alternative lives, are likely to
be underdeveloped.
There are three different but interrelated levels at which socialization can im-
pede autonomy: first, at the level of the processes of formation of our beliefs, de-
sires, patterns of emotional interaction, and self-conceptions; second, at the level of
the development of the skills and abilities that constitute what Diana Meyers calls
autonomy competence; third, by frustrating a person's ability or freedom to act
upon or realize her autonomous desires or an autonomously conceived life plan.
For good reason, feminists have had a lot to say about how restricted social oppor-
tunities curtail autonomy at the third level, and Meyers has investigated the way in
which socialization can hamper the realization of autonomy at the second level—
by hampering the development of autonomy competence. Understanding the role
played by imaginative representation in self-definition and understanding the rela-
tionships among our individual imaginative projects, social recognition, and the
cultural imaginary are crucial parts of understanding how socialization can impede
autonomy at the first level—at the level of the processes of formation of our be-
liefs, desires, patterns of emotional interaction, and self-conceptions. What it can
help explain is why, in oppressive social contexts, the capacities of agents for au-
tonomous action can be impaired by their own inabilities to imagine themselves
otherwise.
 Imagining Oneself Otherwise 145

Notes

1. My thanks to Sarah Bachelard, Peter Menzies, Diana Meyers, and especially Natalie
Stoljar for helpful comments on earlier drafts of this article.
2. Diana Meyers, Self, Society, and Personal Choice (New York: Columbia University
Press, 1989), part 2, section 4. In this book, Diana Meyers develops the notion of autonomy
as a competence that comprises an array of skills and capacities. She provides a useful sum-
mary of her account in "Personal Autonomy and the Paradox of Feminine Socialization,"
Journal of Philosophy 84 (1987): 619-628. See also the discussion of autonomy competency
in the introduction to this volume.
3. Paul Benson, "Autonomy and Oppressive Socialization," Social Theory and Practice
17 (1991): 385-408, 396-397.
4. Richard Wollheim, The ThreadofLife (Cambridge: Cambridge University Press, 1984).
5. A number of feminist philosophers, for example, have analyzed the alignments be-
tween philosophical conceptual sttuctures and symbolic or metaphorical cultural representa-
tions of masculinity and femininity. In The Man of Reason, 2nd ed. (London: Routledge,
1992), Genevieve Lloyd analyzes the historically variant symbolic associations between rea-
son and masculinity in the history of philosophy; in The Philosophical Imaginary, trans.
Colin Gordon (Stanford, Cal.: Stanford University Press, 1989), Michele Le Doeuff analyzes
the conceptual role played by a range of different philosophical metaphors, including
metaphors of the masculine-feminine distinction; in Speculum of the Other Woman, trans.
Gilligan C. Gill (Ithaca, N.Y.: Cornell University Press, 1985), Luce Irigaray focuses on cer-
tain textual metaphors, for example, in the work of Plato, Hegel, and Freud, among others,
to suggest that Western theoretical discourse is structured by a phallocentric representation of
sexual difference in which the feminine is always represented in relation to masculinity as lack
or complement. In "Gendered Reason: Sex, Metaphor and Conceptions of Reason," Hypa-
tia 6 (1991): 77-103, Phyllis Rooney locates feminist concern with metaphorical and sym-
bolic representations of reason and femininity in the context of theories of metaphor, such as
those of Max Black, that argue for the view that metaphor has conceptual force. See Black,
"Metaphor," in Models and Metaphors (Ithaca, N.Y.: Cornell University Press, 1962). For dif-
ferent developments of Black's account, see also Paul Ricoeur, The Rule of Metaphor, trans.
Robert Czerny (London: Routledge & Kegan Paul, 1978), and Eva Kittay, Metaphor: Its Cog-
nitive Force and Linguistic Structure (Oxford: Clarendon Press, 1987).
6. In "Imagination and the Self," Bernard Williams makes a similar distinction between
"simple visualing," in which a scene is visualized from no particular point of view, and "par-
ticipant imagery," in which the imaginer is a participant in the visualized scene. Williams,
"Imagination and the Self," in Problems of the Self (Cambridge: Cambridge University Press,
1973). Although Wollheim does not explicitly discuss Williams, his analysis is clearly an en-
gagement with the issues raised by Williams.
7. Here Wollheim's analysis disagrees with Williams, who argues that one cannot co-
herently imagine being someone else. Williams's argument centers on the analysis of imag-
ining being Napoleon.
8. Wollheim, Thread of Life, p. 71.
9. Ibid., p. 87.
10. Wollheim's concern with analyzing what he calls the process of life, or what it
means to live the life of a person, seems to be a response to reductive accounts of personal
identity, such as that proposed by Derek Parfit in Reasons and Persons (Oxford: Oxford Uni-
versity Ptess, 1984). Parfit claims that a person is no more than a brain, a body, and a series of
interrelated mental and physical events and that a person's life can be described, without
omission, in purely impersonal terms. For critiques of Parfit that provide different but related
146 Autonomy and the Social

accounts of exactly what is omitted by an impersonal description, see Christine Korsgaard,

"Personal Identity and the Unity of Agency: A Kantian Response to Parfit," Philosophy and
Public Affairs 18 (1989): 101-132, reprinted as chap. 13 of Korsgaard, Creating the Kingdom
of Ends (Cambridge: Cambridge University Press, 1996); also Paul Ricocur, Oneself as An-
other, trans. Kathleen Blarney (Chicago: University of Chicago Press, 1992), studies 5 and 6.
11. T he temporal dimensions of selfhood are also discussed by Genevieve Lloyd in "In-
dividuals, Responsibility, and the Philosophical Imagination" in this volume.
12. Wollheim, Thread of Life, p. 31.
13. Ibid., p. 98.
14. Ibid., p. 100.
15. Again the target here seems to be a view of memory like that of Parfit. Parfit pro-
vides an account of ordinary memory as a subclass of quasi memories. Ordinary memories
are quasi memories of our own past experiences, as opposed to other people's past experi-
ences. The argument hinges on viewing memories as memory traces, raising the theoretical
possibility that one person's memory traces might be copied from one person's brain and
transplanted into another person's brain. The argument centers on the example of Jane's
quasi memories of Paul's memories of Venice, as a result of an operation that involves copy-
ing and transplanting a discrete section of Paul's memory traces into Jane's brain. Parfit ar-
gues that Jane quasi-remembers Paul's memories of Venice from Paul's point of view, that is,
in the first-person mode of presentation. However, where these quasi memories differ from
ordinary memories is that whereas the latter involve the first-person mode of presentation ac-
companied by a belief that these memories are memories of one's own experiences, Jane's
quasi memories are not accompanied by this belief (Parfit, Reasons and Persons, pp. 220 — 222).
On Wollheim's account of memory, the idea that one could quasi-remembcr somebody else's
experience from their first-personal point of view is incoherent because it violates the neces-
sarily dispositional character of memory. Wollheim's discussion centers on an example that
is similar to Parfit's, the example of quasi-remembering being his own father as a boy walk-
ing to school through the streets of Breslau (Thread of Life, pp. 112-115). In his discussion
of Parfit in Oneself as Another, Ricocur is also critical of Parfit's notion of quasi memory,
claiming that in reducing memory to the memory trace, Parfit has stripped memories of their
distinctive phenomenological character of always being someone's memories. Like Wollheim,
Ricoeur connects this phenomenological character with temporality.
16. Susan Brison makes a similar point about memory in "Outliving Oneself: Trauma,
Memory and Personal Identity," in Feminists Rethink the Self, ed. Diana Tietjens Meyers
(Boulder, CO: Westview, 1997). Her discussion of the effects of traumatic memory brings
out the way in which memories shape our points of view. A central feature of traumatic
memories is that they block agents' abilities to project themselves into the future, hence giv-
ing rise to the feeling among trauma survivors of having outlived themselves, a feeling cap-
tured in the metaphor of having stayed on the train one station past one's destination. In Bri-
son's analysis, this blockage arises from the way in which traumatic memories invade the
present, dominating the person's point of view and disconnecting her from her memories of
her past before the traumatic event. My analysis of self-formation suggests that traumatic
memories form part of a person's point of view that cannot be integrated into the person's
self-conception. The process of recovery involves the attempt to achieve some measure of in-
tegration. Iwo recent novels that powerfully evoke and reflect on the interweaving of mem-
ory and the present in the aftermath of trauma are the three novels forming Pat Barker's The
Regeneration Trilogy (London: Viking, 1996) and Anne Michaels, Fugitive Pieces (London:
Bloomsbury, 1997).
17. Meyers, Self, Society, and Personal Choice.
18. Amelie Rorty and Richard Wong use the notion of "reflective equilibrium" to char-
 Imarinin? Oneself Otherwise
O o J
147

acterize the process whereby agents shape their self-identities. I am adopting their usage of
this term. Rorty and Wong, "Aspects of Identity and Agency," in Identity, Character, and
Morality: Essays in Moral Psychology, ed. Amelie Rorty and Owen Flanagan (Cambridge,
Mass.: MIT Press, 1990).
19. Many theorists of autonomy see some kind of intrapsychic integration as a neces-
sary condition for autonomy (for a fuller discussion see the introduction to this volume). I
concur with this view, although I do not provide an argument for it here. However, for rea-
sons that are now familiar in the literature and that I discuss later, although I regard integra-
tion as I characterize it in this section as a necessary condition for autonomy, it is not a suffi-
cient condition. My account of integration is probably closest to Marilyn Friedman's in that,
like her, I resist a hierarchical account of the different elements of the self and see integration
as involving a mutual process of reflection among these different elements. Friedman, "Au-
tonomy: A Critique of the Split-Level Self," Southern Journal of Philosophy 24 (1986):
19-35. Where my account differs from Friedman's is in its characterization of the different
elements of the self. My account also shares similarities with the accounts of Diana Meyers,
who stresses the importance of conceptualizing integration within the context of a concep-
tion of the self as dynamic and capable of self-transformation. Meyers, Self, Society, and Per-
sonal Choice, part 2, sect. 2, and "Intersectional Identity and the Authentic Self?: Opposites
Attract" in this volume.) What makes the self dynamic, in my view, is first, the fact that the
self is internally differentiated and not a seamless unity and, second, the fact that the self is
formed and continually transformed in the context of relations with others, relations of con-
nection and differentiation. I discuss these points more fully later.
20. It is important not to read the metaphors of the internal audience and an internal
observer too literally, as implying different entities within the self. The point of the meta-
phors is to capture the idea that we can take up different attitudes toward ourselves and to-
ward our own emotional states and to show that these different attitudes are characterized by
different degrees of involvement.
21. Because I stress the importance of both point of view and self-conception to an
agent's identity, my view is not vulnerable to the charges made against hierarchical accounts
by Irving Thalberg, for example, that they give privilege to only the "higher-level" aspects of
the self and so beg the question against Freudian accounts of our psychic structure. Thai-
berg, "Hierarchical Analyses of Unfree Action," Canadian Journal of Philosophy 8 (1978):
211-225. On my account, unconscious desires form part of a person's point of view.
22. My account of externalization recalls Frankfurt's notion of externality in "Identifi-
cation and Externality," in The Identities of Persons, ed. Amelie Rorty (Berkeley: University of
California Press, 1977), and "Identification and Wholeheartedness," in Responsibility, Char-
acter and the Emotions, ed. Ferdinand Schoeman (New York: Cambridge University Press,
1987). Both essays are reprinted in Frankfurt, The Importance of What We Care About (New
York: Cambridge University Press, 1988).
23. See especially his critique of utilitarianism in "A Critique of Utilitarianism," in
J. J. C. Smart and Bernard Williams, Utilitarianism: For and Against (Cambridge: Cambridge
University Press, 1973).
24. Dworkin's distinction between critical and experiential interests arises in the context
of his discussion of euthanasia in Life's Dominion: An Argument about Abortion and Eu-
thanasia (London: Harper Collins, 1993), chap. 7.
25. See especially "Identification and Externality," "The importance of what we care
about," "Identification and wholeheartedness," and "Rationality and the unthinkable," in
Frankfurt, The Importance of What We Care About. See also "The Necessity of Ideals," in The
Moral Self, ed. G. Noam andT. Wren (Cambridge, Mass.: MIT Press, 1993).
26. It is important to note the difference between dissociation and disavowal. Whereas
148 Autonomy and the Social

disavowal involves denial, dissociation can involve accepting a desire or a character trait as an
element of one's makeup, without endorsing it. The process of acceptance, without endorse-
ment, can sometimes be very important to our ability to achieve a relatively stable integration
among the different elements of the self since a disavowed desire is very likely to lead to se-
rious internal conflict and self-divison. A similar point is made by Jean Grimshaw, who ar-
gues for the importance of conceptualizing autonomy in such a way that it does not require
agents to disown or disavow their desires. As Grimshaw points out, in many cases acceptano:e
of aspects of oneself that one may not particularly like or endorse is more likely to promote
autonomy than disavowal. Grimshaw, "Autonomy and Identity in Feminist Thought," in
Feminist Perspectives in Philosophy, ed. Morwenna Griffiths and Margaret Whitford (Bloom-
ington: Indiana Uuniversity Press, 1988). Susan Brison's discussion of coming to terms with
traumatic memories is an example of the difference between dissociaton and disavowal and
highlights the importance of the former to self-formation. Brison, in "Outliving Oneself,"
suggests that an important part of remaking herself in the aftermath of trauma involved dis-
sociating herself from her traumatic memories. This dissociation was not a matter of denying
that the remembered traumatic event happened to her and was part of her subjective history
but rather a matter of dissociating her self-conception from the event and her memories of it.
27. It is because I see this process of reflection as involving a three-way process of ne-
gotiation among our points of view, self-conceptions and cares, values and ideals, with no
particular element of the self given primacy, that I regard this account as non-hierarchical.
28. Anne Donchin's article in this volume, "Autonomy and Interdependence: Quan-
daries in Genetic Decision Making," draws attention to the nonvoluntary nature of biologi-
cal and genetic ties and focuses on their implications for conceptions of autonomous deci-
sion making in genetic contexts.
29. For further discussion of the way in which the kind of fragmentation of the self
that is a consequence of trauma can change a person's sense of what matters, see Brison,
"Outliving Oneself," and "Surviving Sexual Violence: A Philosophical Perspective," Journal of
Social Philosophy 24 (1993): 5-2.2.
30. Amy Mullin argues for a related account of integration as involving a process of nc-
gotation between diverse aspects of the self. Mullin, "Selves, Diverse and Divided: Gan Fem-
inists Have Diversity without Multiplicity?" Hypatia 10 (1995): 1-31.
31. The importance of this kind of intrapsychic integration to agency is made partic-
ularly clear by the effects of trauma, which is often characterized as involving loss, fragmen-
tation, or "dismemberment" of the self, sometimes to the point of feeling that one's former
self has died.
32. The notion that unified agency is a practical, not a metaphysical, matter is dis-
cussed by Korsgaard in "Personal Identity and the Unity of Agency."
33. For a detailed discussion of this issue see Brison, "Outliving Oneself."
34. As Annette Baier points out, persons are "second persons," in a number of senses: in
the sense that we only become persons after a long period of dependency on other persons
from whom we learn what it is to be a person; in the sense that we can only acquire and sus-
tain self-consciousness because we know ourselves as persons among other persons; in the
sense that we each only learn to distinguish ourselves as an "I" in the context of being ad-
dressed by another as a "you" and addressing another as a "you"; and in the sense that our self-
consciousness is a historical, narrative consciousness, which emetges through the "acquisition
of a sense of outselves as occupying a place in an historical and social order of persons, each
of whom has a personal history interwoven with the history of a community." Baier, "Garte-
sian Persons," in Postures of the Mind: Essays on Mind and Morals (Minneapolis: University of
Minnesota Press, 1985), p. 90. Baicr's notion of second persons thus connects the social or
relational dimensions of personhood with its historical dimensions through the notion of
 Imagining Oneself Otherwise 149

memory as recollection or narrative self-consciousness. These connections are also central to

Genevieve Lloyd's reading of the narrative self in Being in Time: Selves and Narrators in Lit-
erature and Philosophy (New York: Routledge, 1993) and to her account of the social dimen-
sions of responsibility in "Individuals, Responsibility, and the Philosophical Imagination," in
this volume.
35. Susan Brison's reflections on trauma suggest that when one's sense of self has been
"dismembered," and the various elements of the self are in disequilibrium, this tacit self-
knowledge is also disrupted. The survivor of trauma no longer knows who she is, what she
feels, or what aspects of her former self she can continue to identify with—which is why she
is unable to integrate the various aspects of her identity.
36. Diana Meyers also argues that persistent regret is a good indicator that a decision
or choice was not autonomous. See Self, Society, and Personal Choice, part 2, sect. 1. As Ber-
nard Williams makes clear, however, this kind of "agent-regret" may be the result not only of
failures of self-knowledge but also of bad luck. Thus, achieving an equilibrium between the
different aspects of the self might be more a matter of luck than we find comfortable.
Williams, "Moral Luck," in Moral Luck (Cambridge: Cambridge University Press, 1981).
37. This is, of course, a central theme in Hume's moral psychology, a theme empha-
sized in Annette Baler's interpretation of Hume. David Hume, A Treatise of Human Nature,
ed. L. A. Selby-Bigge, rev. P. H. Nidditch (Oxford: Clarendon Press, 1978), books 2 and 3;
Baier, A Progress of Sentiments: Reflections on Hume's Treatise (Cambridge, Mass.: Harvard
University Press, 1991).
38. Once again, the experience of trauma illuminates the role played by social relation-
ships and social recognition in intrapsychic integration. Trauma is characterized not only by
self-alienation but also by a sense of loss of connection with intimate and nonintimate oth-
ers. One symptom of this is that the agent's usual trust that others are not disposed to harm
her has been shattered; another is that the agent feels that her experience of trauma puts a
gulf between her and others who have not undergone the same experience. This gulf leads to
the feeling, on the part of the agent, that she cannot be herself with other people. This feel-
ing is a persistent emotion for Billy Prior, one of the central characters in Pat Barker's The Re-
generation Trilogy, whose sense of alienation from civilians who have not experienced the hor-
rors of the trenches sometimes manifests itself in an overwhelming violent hatred of them.
For many of the characters in Barker's novels, including Prior, the only way to escape this
sense of alienation is to go back to the trenches and to the solidarity in the face of horror that
unites the soldiers there. For Brison, being able to overcome the feeling of alienation from
others and to reconnect with them was one of the most important factors in regaining her
sense of self, or rather remaking herself. Constructing a narrative of the experience and hav-
ing others listen to it and, in doing so, acknowledge the pain she had undergone were essen-
tial to being able to reconnect with others and, through them, with her self.
39. My discussion at this point is indebted to Paul Benson's analysis of self-worth in
"Free Agency and Self-Worth," Journal of Philosophy 91 (1994): 599-618.
40. Benson makes a similar point in "Autonomy and Oppressive Socialization."
41. Brison, "Outliving Oneself," p. 30.
42. On the notion of intersectional identity, see Kimberle Crenshaw, "Demarginalizing
the Intersection of Race and Sex: A Black Feminist Critique of Antidiscrimination Doctrine,
Feminist Theory, and Antiracist Politics," in Feminist Legal Theory, ed. Katharine T. Bartlett
and Rosanne Kennedy (Boulder, Colo.: Westview, 1991).
43. Natalie Stoljar's discussion of autonomy and contraceptive risk taking in "Auton-
omy and the Feminist Intuition," in this volume, provides another example of the way in
which an agent's capacities for autonomous decision may be compromised if her sense of
self-worth is overly dependent on a narrow range of forms of social recognition.
1 50 Autonomy and the Social

44. As Benson remarks, in raising this objection against Marilyn Friedman's integration
account of autonomy, "An integration view detects threats to autonomy only when the total
internali/ation of autonomy-inhibiting socialization fails to take hold or begins to break
down" ("Autonomy and Oppressive Socialization," p. 95).
45. For a recent discussion of the role played by cultural imagery in shaping identities,
see Diana Meyers, Subjection and Subjectify: Psychoanalytic feminism and Moral Philosophy
(New York: Routlcdge, 1994).
46. In Wild Swam (London: Harper Collins, 1991), a personal and cultural autobiog-
raphy of life in Mao's China, Jung Chang vividly describes how a whole country can be in the
grip of such imagery.
 7

INTERSECTIONS IDENTITY AND

THE AUTHENTIC SELF?
Opposites Attract!1

Diana Tietjens Meyers

I start from the conviction that autonomy is not a figment of the philosophical
imagination; rather it is a phenomenon that most people have some experience of
and that they commonly value (in my opinion, rightly so). Nevertheless, the term
"autonomy" is not in everyday usage, and, it must be admitted, it is a philosophical
term of art. Still, it does not follow that people have no way of talking about auton-
omy or that they don't talk about it. Colloquially, the exhortations "Be true to your-
self!" and "Don't cave in!" express the value people place on autonomy, and the dec-
larations "She lives by her own lights," "Now I know what I really want," and "I feel
solid [right in my skin, at one with myself] about this" voice their achievement of
autonomy. Think, too, of the soaring refrain of the Sinatra tune, "I did it my way"
—a paean to autonomy. Likewise, there are idiomatic expressions associated with
heteronomy. People may bemoan their lack of autonomy—"I feel at sea [adrift],"
"I'm at odds with myself," or "How could I have given in?" And they may stand up
for their autonomy and claim it as a right—"Mind your own business!" or "Butt
out!" When people feel confused, irresolute, pressured, and so on, they may lament
their lack of autonomy or protest incursions on it. When people are clear about
what they truly want, who they deeply care about, what they genuinely believe in,
and so forth, and when they are able to act on these desires, affections, and values,
they may attest to their own autonomy.
In philosophical treatments of autonomy, this satisfying, sometimes exhilarat-

151
I 52 Autonomy and, the Social

ing, experience of self-understanding and self-realization has been crystallized in

the ideas of authenticity and self-governance. This way of conceptualizing the phe-
nomenon of autonomy has, alas, proved susceptible to hyperbolic distortion. Self-
understanding has been taken to presuppose a transparent self; congruence of self
and action has been taken to presuppose a unitary, homogeneous self; self-governance
has been taken to presuppose unfettered independence from other individuals, as
well as from the larger society.
To the extent that this caricature has seized the philosophical imagination, fem-
inists have charged, the autonomous individual has been reduced to an androcentric
phantasm. 2 Yet, despite feminist objections to the self-originating, self-sufficient,
coldly rational, shrewdly calculating, self-interest maximizing, male paragon of au-
tonomy, and despite feminist wariness that reclaiming autonomy will prove anti-
thetical to the project of revaluing interpersonal capacities that are conventionally
coded feminine, many feminist writers continue to invoke ostensibly discredited
values like self-determination in unguarded writing about the needs of women and
the aims of feminism. 3 It is notable, too, that other feminists have rallied to the ex-
plicit defense of autonomy. 4 This revival is by no means surprising, for feminists
must account for the control women exert over their lives under patriarchy, for their
opposition to subordinating social norms and institutions, and for their capacity to
bring about emancipatory social change. The concept of autonomy promises to be
helpful in these endeavors.
The philosophical problem of autonomy takes on texture and depth when it is
situated in a realistic context. The reality I propose to inject into my discussion of
autonomy is the fact that enormous numbers of people are assigned to social groups
that are systematically subordinated. The wonder is that despite this subordination,
some of these individuals are exemplars of autonomy, and few of them altogether
lack autonomy. There are autonomous dissenters and revolutionaries and legions of
individuals who autonomously craft private lives within the confines of oppressive
regimes.5 That this is so belies two prominent, seemingly incontrovertible assump-
tions about autonomy. First, although it seems undeniable that an autonomy-friendly
environment is necessary for autonomy, this is not the case. Horrible as they are, so-
cial and economic structures that funnel individuals into a preordained status, that
regiment their life trajectories, and that penalize nonconformity need not defeat au-
tonomy. A social and economic environment that makes a wide range of attractive
options available to all individuals is conducive to, not necessary for, autonomy. Sec-
ond, intuitively it seems that psychic fragmentation would preclude autonomy. Yet,
recent work on the social positioning, vulnerability, and subjectivity of subordi-
nated people suggests that as long as oppressive social structures and discourses sur-
vive, which is to say, for the foreseeable future, their identities will not be unified and
coherent. Still, many of these individuals enjoy autonomy in at least some parts of
their lives. Indeed, one implication of the line of thought I develop is that multiply
oppressed individuals are in some respects better positioned to make autonomous
judgments concerning issues of social justice and policy, and hence are better posi-
tioned to exercise autonomous moral and polirical agency, than multiply privileged
individuals are.
 Intersectional Identity and the Authentic Self 153

Of course, recognizing that autonomy is possible despite structures of domina-

tion and subordination is no reason for complacency. On the contrary, to advocate
expanded opportunities for autonomy is to endorse significant social and economic
change. Nevertheless, I leave this aspect of autonomy aside and focus on the ques-
tion of how autonomy is possible for individuals whose identities are shaped by
structures of domination and subordination.
Specifically, I take my cue from work by feminists and critical race theorists on
women's diversity and intersectional identity. The idea of intersectional identity is
premised on the general philosophical thesis that who one is depends on one's social
experience. However, the intersectional conception is specific to societies that ex-
hibit certain kinds of social stratification, for it derives from a social-psychological
view about how individuals internalize gender, sexual orientation, race, class, and
ethnicity in sexist, homophobic, racist, classist, and xenophobic societies. This view
does not purport to capture the whole of personal identity but rather those aspects
that are conditioned by membership in subordinated or privileged social groups. It
emphasizes that people are categorized according to gender, sexual orientation, race,
class, and ethnicity and that these multiple ascriptions interact—sometimes com-
pounding one another's effects and sometimes creating inner divisions and con-
flicts.6 On this view, who we are—what we are like and how we think and act—is
significantly influenced by social systems of domination and subordination. Of
course, intersectionally constituted subjects may or may not notice attributes that
stem from crosscutting systems of domination and subordination, and when they
notice such attributes, they may or may not understand them as consequences of
these social structures. People are not introspectively clairvoyant, but to the extent
that they fail to make these connections, their self-concepts are incomplete and pos-
sibly distorted.
Although Lacan originated the intersectional view of the self, I take the liberty
of ignoring Lacanian doctrine. Instead, I draw attention to the tropes feminists and
critical race theorists are currently using to articulate the concept of intersectionality.
This approach is justified, in my view, because the impulse to fashion novel imagery
hints at points of dissatisfaction with Lacan's theory and because these theorists' im-
agery is suggestive of heretofore buried agentic potential. After reviewing in the first
section a number of tropes of intersectional identity and characterizing the experi-
ences and capacities they symbolize, I take up the topic of autonomy, and I seek to
capitalize on the richness of these intersectional tropes.
To clarify the problem of autonomy once it is placed in an intersectional frame-
work, I set out several obstacles to achieving autonomy as an intersectional subject
(in the second section). I then focus on two aspects of autonomy—self-knowledge
and self-definition—and argue that intersectional identity obliges us to rethink the
concept of authenticity (in the third section). If ignorance of one's intersectional
identity impairs autonomy, intersectional identity and hence the internalized norms
of the groups to which one is assigned are attributes of the authentic self. But since
intersectional identity is constituted in part through a process of self-definition, the
authentic self is an evolving self that is not chained to conventional group norms.
Although it is easy to discard concepts of authenticity that exclude the influence of
154 Autonomy and the Social

social experience or that rule out personal transformation, it is extremely difficult to

furnish an account of authenticity that does not succumb to these pitfalls. To bring
out the nature of this difficulty, I examine Harry Frankfurt's well-known views. Al-
though Frankfurt's account of the authentic self accommodates intersectional iden-
tity in some respects, I argue that its reliance on organizing identity constituents into
an integrated order ultimately defeats it (in the fourth section). 1 propose, then, that
we redirect our attention — away from the internal structure of the authentic self
and toward the process of constituting an authentic self. Thus, I recommend view-
ing the authentic self as the collocation of attributes that emerges as an individual
exercises self-discovery and self-definition skills (in the fifth section). Applying the
skills that comprise autonomy competency, intersectional subjects analyze their po-
sition in social hierarchies, interpret the psychic impact of their social experience,
and reconfigure their identities as members of social groups. In so doing, they con-
stitute authentic intersectional identities.

The Trope of Intersectlonality

Intersectionality is currently the reigning feminist metaphor for complex identities
insofar as they arc constituted by race, ethnicity, class, and sexual orientation to-
gether with gender.7 The trope of intersectionahty has numerous variants taken
from a wide assortment of domains, including mathematics, cartography, anatomy,
and botany. While all of these images share the aim of symbolizing an identity
drawn from diverse sources that may give rise to conflicting desires and rival alle-
giances, they highlight different potentialities and liabilities of such identities.
Kimberle Crenshaw's primary image of intersectionahty may seem antisepti-
cally mathematical. She pictures intersectional identities as points on a plane plotted
in relation to axes which represent practices of oppression such as sexism and racism.8
Invoking social scientific quantification, this image desubjectivizes gender and race
and accents baleful social realities. But by declaring that the social scientific appara-
tus of variables, equations, and graphs is itself a metaphor, Crenshaw punctures the
objcctivist pretensions of social scientific analysis and resubjectivizes her image. The
axes and plotted points translate into palpable experience.
Crenshaw does not confine herself to a single static, figurative interpretation of
intersectionahty. She also dramatizes it. Sexism and racism are vehicles on a collision
course. The vehicles are hurtling down highways that converge at an intersection,
where a pedestrian who is crossing cannot get out of the way before the two vehi-
cles speed into the intersection and crash.9 Struck simultaneously from two direc-
tions, the pedestrian's injuries are severe. But since it is almost impossible to trace
them to their respective causes, responsibility cannot be apportioned between the
drivers. Consequently, neither is blamed, and both go free. Crenshaw states that sub-
ordinating practices operate synergistically, but plainly she has in mind a sinister
synergy—sinister both because of the devastating harm inflicted and because of the
unwarranted privilege condoned. 10
Crenshaw's crash metaphor vividly symbolizes the victimization brought about
by patterns of oppression. However, it misses a contrary feature of group-based iden-
tity. The targets of oppression seldom experience themselves solely as victims. They
 Intersectional Identity and the Authentic Self 15 5

often defy others' bigotry by valuing their association with the group that others sys-
tematically penalize and by upholding its traditions. The most common vernacular
trope for group-based identity, namely, group membership, expresses this positive
identification. Appropriating and adapting this locution, Kirstie McClure writes
of intersecting group memberships.11 If visualized as lifeless Venn diagrams, her
trope merely represents inert items sorted into various overlapping categories. But it
doesn't take much imagination to animate the trope. Since the items referred to are
people, it is natural enough to construe membership in a group as belonging to a
group and to associate belonging to a group with joining it. Thus, membership im-
agery introduces connotations of feeling at home in a community and of voluntary
participation in it—that is, of willed and willing identification with a social group.
Important as these dimensions of group-based identity are, it might be coun-
tered that group membership should not be portrayed in overly rosy terms. As Cren-
shaw observes, African-American women may find themselves "situated between
categories of race and gender" since these categories are sometimes regarded as mu-
tually exclusive.12 Thus, they may experience ambivalence toward, even alienation
from, their race or their gender. Maria Lugones's cartographic imagery perspic-
uously captures the complex group-based identities that may give rise to this sort
of experience. Intersections of geopolitical entities—that is, borders—symbolize
identity-constituting regimes of stratifying social classifications, and the border
dweller personifies intersectional positioning and experience.13 Although familiar
with and drawn to both sides of the boundary line, border dwellers are not com-
pletely welcome and comfortable in either territory. Despite the awkwardness of this
position, Lugones recommends inhabiting frontiers. In her view, border dwellers oc-
cupy an epistemically favorable vantage point, for the virtues and the defects of each
community are easier to spot from the border.
Lugones assigns border dwellers the role of emissaries who maintain diplo-
matic relations between mutually suspicious (in some cases, hostile) polities. Chan-
tal Mouffe's construal of intersectional identities as articulations elaborates this me-
diational theme.14 On the one hand, her metaphor calls to mind anatomical and
botanical intersections. Articulations are joints or nodes, organic mechanisms of
juncture and mobility. On the other hand, there are articulations that function as so-
cial intersections—utterances that demarcate and communicate positions; that in-
vite others to reply; that may be refined, amended, or renounced. Deemphasizing
the harms wrought by group-based domination and subordination and playing on
this double entendre, Mouffe's image symbolizes both the fixity of group classifica-
tions and their amenability to renegotiation and modification.
This sample of tropes of group-based identity highlights two claims about per-
sonal identity and groups. First, these images put a spectrum of identity determi-
nants on a par. Gender is not portrayed as more fundamental or more pervasive than
race, class, ethnicity, or sexual orientation, nor is gender portrayed as secondary to
any of these other group categories. Insofar as identity is shaped by these social
structures, it is characterized as a commingling of equally robust, equally decisive
identity determinants. Second, the notion of intersectional identity is paradoxical.
Ties to groups are commonly experienced as emotionally gripping and integral to
one's sense of self, yet these ties may be experienced as imposed and confining, even
156 A uttmomy and the Social

wounding. Likewise, the divergent demands entailed by ties to different groups can
lead to estrangement from oneself and from others, yet they endow individuals with
opportunities for agency, both for self-definition and for affiliation with others.

Barriers to the Autonomy of Intersectional Subjects

Intersectional subjects arc not at all promising material for autonomy, for it is diffi-
cult for them to achieve self-knowledge, self-definition, and self-direction. If doing
what one "really wants" is a hallmark of autonomy, one must know who one is —
one's capabilities and weaknesses; one's values, commitments, and goals; and so on
—so that one can make choices that express one's "true self," that is, one's "genuine
desires." Unfortunately, everyday discourses of selfhood conspire against discerning
the ways in which one's identity is codetermined by gender, sexual orientation,
race, class, and ethnicity. Thus, intersectional subjects may be too opaque to them-
selves to be autonomous. Even supposing that this obstacle to self-knowledge could
be overcome, however, intersectional codetermination would itself pose further ob-
stacles to autonomy. When autonomous individuals are dissatisfied with them-
selves, they must be able to make needed changes, and when they have decided
upon a course of action, they must be able to carry it out. But since group ascrip-
tions constrain intersectional subjects and since codetermination unsettles them,
self-definition and self-direction are impeded. Again, intersectional identity jeop-
ardizes autonomy.
In the United States today, intersectionality is not a default conception of per-
sonal identity. Indeed, most people are not even acquainted with this conception.
No doubt, American society's inveterate denial of social stratification and pervasive
injustice combines with its creed of equal opportunity and upward mobility to mil-
itate against admitting relations of domination and subordination as sources of
identity. On those occasions when we need a conception of identity that acknowl-
edges inner heterogeneity, for example, to make excuses, Freud's privatized divided
self can be pressed into service without bringing issues of social justice to the fore.
Likewise, the idealization that is homo economicus or the liberal bearer of rights
valorizes unity, coherence, and constancy, and the ubiquity of this idealization casts
suspicion on the conflicts endemic to intersectional subjectivity. These default con-
ceptions of identity leave little room for alternative conceptions to gain propo-
nents, for the very familiarity of these discourses of selfhood induces us to articu-
late our self-concepts by using these models. Without a discourse that is conducive
to articulating an intersectional identity, few people will view themselves as inter-
sectional subjects.
Current ordinary American English does not furnish language for use in first-
person singular identity avowals that gives equal weight to diverse group determi-
nants of identity and that prompts individuals to consider how these identity deter-
minants interact.15 Colloquial language that demarcates class identities does not
exist- -according to our social mythology, almost everyone is middle class. Racial
categories are conceptually muddled, and affirmations of racial identity are so fraught
with emotional uneasiness that they are often perceived as inflammatory. 16 Yet, the
ethnicities of racial "minorities" are sidelined, so their race remains on center stage.
 Intersectional Identity and the Authentic Self 157

For example, outside the Haitian and Korean communities, these ethnicities are
subsumed by racial classifications. In contrast, for whites, ascriptions of ethnicity
take precedence over race. Since the ethnicities of Euro-Americans are racialized—
to be Swedish American or Italian American is by definition to be white—their race
is ever-present yet pushed into the wings. The conventions of self-ascriptions of
gender and sexual orientation are asymmetrical, too. Equated with humanity itself,
maleness and heterosexuality go unmentioned, whereas womanhood, homosexu-
ality, and bisexuality are mandatory and salient categories of self-description.17 Al-
though people are intersectional subjects, many of them do not know it, for our dis-
course exaggerates the significance of some group memberships while discounting
the significance of others. But to lack self-knowledge is to lack autonomy.
An additional obstacle to autonomy for intersectional subjects stems from the
fact that gender, sexual orientation, race, class, and ethnicity are categories that di-
vide people into dominant and subordinate groups and that prescribe personality
types and behavioral norms for each. In a society that is historically divided along
harsh, unyielding axes of dominance and subordination, individuals cannot escape
the influence of cultural stereotypes and other prescriptive representations of the
groups they belong to, nor can they escape the influence of the social and economic
advantages and disadvantages that institutions confer on these groups. As Anthony
Appiah remarks, "I don't recall ever choosing to identify as a male; but being male
has shaped many of my plans and actions."18 Group ascriptions are not the sole de-
terminants of who we are, nor do members of the same social group have identical
identities or, for that matter, a subset of identical identity components. Neverthe-
less, individuals who belong to different groups have different social experience, and
this targeted and lifelong socialization shapes individualized identities.19 Whether
they know it or not, people have intersectional identities, and these identities influ-
ence what they believe, how they deliberate, and how they conduct their lives.
Many individuals in privileged categories do not experience their group ties as
constraints, but many individuals in disvalued categories experience theirs as subju-
gating. Whether one is assigned to a dominant or a subordinate group, though, one
is inducted into its distinguishing paradigms and norms. Most men never seriously
consider using their right to family leave to spend time with a newborn or newly
adopted child, and a man who does so is likely to have as hard a time climbing the
corporate ladder as a woman who wants to be CEO (chief executive officer). Thus,
to have an intersectional identity is to be culturally defined and directed. But since
self-definition and self-direction are necessary for autonomy, group-targeted cultural
norms put autonomy at risk.
Finally, tensions often arise between the demands of gender, sexual orientation,
race, class, and ethnicity—for example, for an African-American man, the norms of
gender prescribe authority, whereas the norms of race prescribe deference; for a les-
bian, the norms of gender mandate motherhood, whereas the norms of sexual ori-
entation prohibit it. Likewise, the diversity of the individuals assigned to these cat-
egories and their divergent interests may present political conundrums. For African-
American women, commitment to antiracist politics may entail tolerating sexism,
and commitment to feminist politics may entail tolerating racism. As a result, inter-
sectional identities often leave individuals torn by conflicting self-understandings
158 Autonomy and the Social

and conflicting social and political loyalties. But if one cannot decide what one re-
ally wants, one cannot do what one really wants—one cannot be "true to oneself."
To the extent that intersectional identity prevents one from translating one's identity
into action, it thwarts self-direction and hence autonomy.

Authenticity as an Intersectional Subject

Observations along the lines just sketched lead some theorists to jettison the idea of
intersectional identity. Some hold that group ascriptions have profound effects on
individuals but that the impact of these classifications is so variable, so contradic-
tory, and so volatile that no consolidation of identity is possible. In this view,
intcrsectionality provides an argument against the very idea of personal identity and
a compelling reason to join with postmodernists in declaring personal identity an
anachronistic modernist sham.20 Call this the contrarian view of identity. Others
deny that individuals inevitably internalize group paradigms and norms and con-
clude that there are only unique individuals and individual identities. Whether an
individual's identity is constituted by ties to any social group depends on that indi-
vidual's choices. When externalized and lodged in oppressive social structures that
place some individuals in "multiple jeopardy,"21 intersectionality clears the way for
resurrecting a curiously voluntarist, individualist—that is, modernist—conception
of identity. Call this the personalisl view of identity.22
I urge that these two challenges to intersectional identity stem from opposite
misunderstandings of the self and agency, and I take the occasion of examining these
misunderstandings to rethink the concept of the authentic self. For the purposes of
this discussion, I base my conception of authenticity on two ideas that are implicit
in the following autonomy idioms: "Now 1 know what 1 really want" and "Be true to
yourself1." The former expression distinguishes desires that one happens to have from
one's real or genuine desires—that is, the desires of the authentic self. The latter
voices the conviction that it is good for people to act on their authentic desires and
so presumes that doing so falls within the bounds of social acceptability—that is,
that one's authentic desires are morally creditable. One's authentic self points to a
way of living that is both distinctively one's own and socially decent. Thus, I con-
sider the relation between intersectional identity and the authentic self from the
standpoint of three domains of autonomous choice and action—personal, moral,
and political autonomy.23

Self-Knowledge
Proponents of the personalist view of identity—those who do not deny that each
individual has an identity but who do deny that group ascriptions play a key role in
shaping it — overestimate the degree to which individuals are immune to their social
context and overlook a major threat to autonomous choice and action by over-
simplifying the task of self-discovery. Self-ignorance is one of the principal enemies
of autonomy, and self-knowledge requires being honest with oneself about who one
is—owning up to and owning one's identity. In what follows, I urge that under-
standing the impact of group memberships on one's identity is necessary not only
 Intersectional Identity and the Authentic Self 159

for personal autonomy but also for moral and political autonomy. In view of the in-
terconnections between owning up to and owning one's intersectional identity and
leading an autonomous life, intersectional identity should be considered a feature of
the authentic self.
Now, let me forestall the objection that this view of authenticity posits a dan-
gerously static conception of the self with retrograde social consequences by not-
ing that owning up to and owning one's intersectional identity is an accomplish-
ment not merely of self-discovery but also of self-definition (see the following
sub-section).24 Thus, accepting intersectional identity as a feature of one's authen-
tic self does not entail clinging to a community of origin or capitulating to stereo-
typical group norms. Rather, it entails analyzing the social significance of one's com-
munity of origin, disclosing to oneself the ways in which associated norms have
become embedded in one's own cognitive and motivational structure, appreciating
how entrenched they are, and assuming responsibility for the ways in which one
may enact them. Autonomy unfolds in situ, and autonomous individuals must work
with whatever material is at hand. Since part of what is at hand in historically strat-
ified societies is an endowment of intersecting group-identity determinants, indi-
viduals who seek autonomy are well advised to candidly appraise where they stand in
society's hierarchies and to make as complete an inventory as they can of what they
have internalized as a result of being assigned to certain social groups. I stress that
this advice applies differently but with equal force to individuals who identify pri-
marily with dominant social groups and to individuals who identify primarily with
subordinate social groups. Whatever one's social position, intersectional identity
serves as a platform for autonomy.
One reason to regard intersectional identity as a feature of the authentic self is
that it is a starting point for leading an autonomous life. Moreover, as I now argue,
impoverished, mistaken, or deluded views about one's group memberships and their
psychic influence—that is, failure to own up to and to own one's intersectional
identity—undermines autonomy. Self-knowledge of intersectional identity is not
merely a possible point of departure for autonomy, it is a portal autonomous in-
dividuals in invidiously stratified societies cannot avoid passing through. Since the
self-knowledge that is necessary for autonomy in unjust societies includes knowl-
edge of one's intersectional identity, intersectional identity should be regarded as a
feature of the authentic self as long as social injustice persists.
Since intersectional identity is not among our default self-concepts, unacknowl-
edged group memberships commonly structure peoples feelings, thoughts, and ac-
tions. Poor understanding of the implications of gender, sexual orientation, race,
class, and ethnicity ascriptions may have damaging repercussions. Ideologies of so-
cial stratification value or disvalue attributes depending on what "sort" of person
bears them—ambitious white men are admirable, whereas ambitious women are
distasteful; boisterous Irish Americans are lively, whereas boisterous Jews are loud;
and so forth.25 Self-interpretations are filtered through this noxious fog of bigotry.26
As a result, individuals whose self-esteem is enhanced may be emboldened to hope
for more out of life, whereas those whose self-esteem is diminished may be dissuaded
from hoping for much of anything. Plainly, individuals' voluntarily elected, identity-
constituting goals and plans often echo the mandates of hierarchy-enforcing insti-
160 Autonomy and the. Social

tutions and practices. A talented, ambitious woman may reproach herself for im-
modesty, pare down her aspirations, and reconcile herself to being the CEO's wife
and unsalaried business hostess. To the extent that the unconscious influence of in-
tersectional identity leads people to misjudge their potential and pursue plans that
cither stunt their development or exceed their capacities, and to the extent that this
unconscious influence disrupts relationships or sabotages undertakings that matter
to them, self-ignorance undercuts their personal autonomy.
I take it to be axiomatic that desires, attitudes, and the like are enigmatic and
subject to interpretation and reintcrprctation because indefinitely many factors con-
verge in their genesis. The origins of a person's overall moral outlook and specific
political preferences are especially baffling. Fueled by one's social context—the opin-
ions of friends, family, and respected public figures, as well as the opinions of people
whom one detests or dislikes—and by one's prognostications about how social pol-
icy may further or hinder personal objectives and plans, a person's moral views and
politics defy easy, linear explanation. Still, if the intelligibility of desires and atti-
tudes requires that they be autobiographically situated, and if gender, sexual orien-
tation, race, class, and ethnicity are important autobiographical strands, tracing the
relations between one's moral convictions and political stance, on the one hand,
and one's membership in these groups, on the other hand, will be necessary to au-
tonomous moral judgment and autonomous political action.
Crenshaw's crash metaphor draws attention to a dimension of intersectional
identity that is particularly germane to the issue of moral and political autonomy.
The crash metaphor represents those subordinated by group-based injustice as suf-
fering from multiple injuries and those enforcing the subordination as exempted
from harm and exonerated of all responsibility. Although Crenshaw's focus is on the
fate of the victims, the concept of intersectional identity reminds us that many peo-
ple are reckless automobile drivers in some respects and pedestrian fatalities in oth-
ers. Many members of subordinated social groups occupy some dominant social po-
sitions as well. Their identities fuse axes of harm with axes of advantage. Likewise,
it is not unusual for individuals who belong to privileged social groups to belong to
one or more subordinated groups as well. Thus, the trope of intersectional identity
dismantles the stark opposition between dominant and subordinate positions. Very
few people have wholly privileged identities. Very few people have wholly subordi-
nated identities. Most people's identities blend socially disvalued and socially valued
group-identity determinants. To many people, this is news. But, I urge, this self-
ignorance compromises autonomy.
Members of subordinated social groups face a cruel dilemma: they have very
good reason to dwell on the proportions, the frequency, and the irreparability of the
iniquities they suffer, and they also have very good reason to forget all about it. As
long as there is group-based disenfranchisement, many people will find their need
to pursue their own projects and fulfill personal goals on a collision course with a
commitment to social justice. Thus, some highly successful professional women de-
nounce feminism and dismiss (even despise) women who blame their career setbacks
on discrimination or family responsibilities—"If I could make it," goes their refrain,
"why can't you?" Moral equilibrium is elusive when social structures and customary
practices are stacked against you. Still, if individuals routinely framed their self-
 Intersectional Identity and the Authentic Self 161

portraits by using the intersectional model, disavowal of group-identity determinants

would be a less feasible, less tempting defense—less feasible because locating oneself
in gender, sexual orientation, race, class, and ethnicity hierarchies would be regarded
as a minimal condition of self-knowledge; less tempting because avowing member-
ship in a subordinate group would not reduce individuals to victimhood since mem-
bership in subordinate social groups is usually coupled with membership in one or
more privileged social groups. In reconnecting people to all of their group-identity
determinants, intersectional identity works as an antidote to shame, self-contempt,
and self-limitation and therefore as a support for personal autonomy.27 It may enrich
the individual's moral and political outlook as well, for in counteracting the deraci-
nation of denying one's membership in a subordinated social group, it would coun-
teract moral anesthesia and motivate people to overcome cooptation and anomie.28
Apprehending one's intersectional identity—that is, knowing who one is—may
prompt resistance to unjust treatment—that is, moral and political autonomy.
Intersectional subjectivity also broadens people's moral purview in other re-
spects. Members of subordinated social groups sometimes become fixated on the
subset of hardships that they themselves suffer, thereby losing sight of other grievous
wrongs. Middle-class women who ardently object to sexist discrimination are no-
torious for exploiting the poor, uneducated women they employ as nannies. But
greater awareness of how group identities are amalgamated in individual identities
can alleviate this problem, for intersectional identity obliges many individuals to
admit to some social privileges, such as the advantage of affluence in a sexist society
or the advantage of heterosexuality in a homophobic society, at the same time as
they fight the injustices that afflict them as members of subordinated social groups.
The point, of course, is not to pacify people into thankfulness for small blessings,
nor is it to condemn strategic prioritization of one group identity over others.
Rather, the point is to mobilize the resources of intersectional identity to temper the
morally untenable zealotry that may accompany a monolithic group identification.
Acute sensitivity to injustice is no threat to moral insight unless one group-identity
determinant is split off from others or gains superordinate status in relation to oth-
ers. Ignorance of one's intersectional identity—that is, ignorance of who one is—
is a major threat to nuanced social critique, and knowledge of one's intersectional
identity is needed for autonomous moral judgment and autonomous political
agency.
For people who are identified primarily with dominant social groups, moral in-
tegrity may become a casualty when intersectional identity is suppressed and dis-
creditable political views are displaced onto universalistic moral principles. Failing to
make connections between one's socially valued group memberships and one's po-
litical views leads to self-deception, rationalization, and corruption of principles—
in short, to moral heteronomy.29 If people fail to acknowledge their reliance on and
eagerness to keep advantages conferred by maleness, heterosexuality, wealth, white-
ness, or ethnic privilege, impartiality is liable to be perverted into a self-serving ex-
cuse for dismissing morally warranted, urgent political demands.30 In camouflaging
the nefarious instrumental purposes served by steadfast adherence to ostensibly fair
principles and the nature of people's emotional investment in the status quo, the
"high ground" of moral scruple becomes an unassailable refuge for the beneficiaries
162 Autonomy and the Social

of social stratification and elitism.31 In societies in which gender, sexual orientation,

class, race, and ethnicity matter enormously to individuals, it strains credulity to
conduct social life as if venerated moral theories were insulated from these patterns
of domination and subordination. Again, intersectional self-knowledge is indispen-
sable to autonomous moral judgment and autonomous political agency.
In urging people to analyze their intersectional identities, what I am endorsing
is maximum transparency of group-identity determinants together with unsparing
honesty about the ways in which group identities are replicated in people's moral
and political thinking. It is as important to autonomy that self-concepts based on
the trope of intersectionality would expose privileged identity determinants and the
conservative views they undergird as it is that such self-concepts would expose de-
valued identity determinants (or combinations of valued and devalued identity de-
terminants) and the social critiques they spark. Without intersectional identity, self-
knowledge is, at best, incomplete and inaccurate. Without self-knowledge, moral
and political autonomy are imperiled.
Now, according to feminist standpoint cpistemology, the social analyses gener-
ated by oppressed social groups arc more reliable. If the moral psychology I have
sketched is correct, my account of the relation between intersectional identity and
autonomy lends support to this claim. Since it is probably less threatening for a per-
son identified primarily with a subordinate social group (or groups) to acknowledge
previously unnoticed ties to a privileged social group (or groups) than it is for a per-
son identified primarily with a privileged social group (or groups) to acknowledge
dependency on undeserved advantages, autonomous moral judgment and au-
tonomous political agency may be more accessible to oppressed individuals.
Still, it might seem that to accept the view I have presented is to despair of au-
tonomy, for it seems unlikely that anyone who was not already autonomous could
gain the self-knowledge that autonomy presupposes. As 1 pointed out before, the
grammar of identity self-ascription in current American English downplays some
group-identity determinants while spotlighting others. Since vernacular discourse
does not invite people to articulate intersectional identities, it is no accident that
relatively few people see themselves intersectionally. Indeed, owning up to and own-
ing an intersectional identity involves a struggle to overcome these very cultural
constraints—that is, a program of autonomous self-exarnination and interpreta-
tion. 1 turn in the next subsection to the question of how this struggle can be con-
ducted, and I offer in the last section a theory of the nature of authenticity that dis-
solves the paradox that one needs to be autonomous to become autonomous. At this
point, I confine myself to suggesting that if owning up to and owning an intersec-
tional identity is an act of autonomy, we have a further reason to regard intersec-
tional identity as a feature of the authentic self. If authentic attributes are neither
innate nor transcendent, if authentic attributes are those that are autonomously dis-
covered or defined, the codeterminants of intersectional identity are authentic.

Self-Definition
According to proponents of the contrarian view of identity—theorists who repudi-
ate the possibility of personal identity of any sort—individuals are at the mercy of
 Intersectional Identity and the Authentic Self 163

the ebb and flow of incoherent group discourses. Wafted about by these crosscutting
discursive currents, individuals have no stable traits and therefore no personal iden-
tity (never mind authentic identity). Plainly, this view flies in the face of most peo-
ple's experience. People have reflexive experience of their own continuities. Some of
these continuities are conspicuous because they are aggravating, frustrating vices,
faults, weaknesses, or foibles that just won't go away. Others may be less attention
grabbing, perhaps because they are not troublesome, but nonetheless real—for ex-
ample, gratifying emotional bonds, beneficial capabilities, pleasing personality traits,
and rewarding interests. Likewise, people have interpersonal experience of others'
continuities and rely on their associates to react in characteristic and hence fairly
predictable ways.
The experience of personal identity that I am invoking does not require an im-
mutable or a monotonously consistent core self. Decentered, processual subjectivity
is compatible with a sense of personal identity. Indeed, to lose all sense of personal
identity (or to find oneself split into multiple identity fragments) is not to advance
to the postmodernist condition with winning verve. Nor is experiencing oneself as
a tight identity knot cleansed of social taint to regress to the modernist condition in
pitiable trepidation. Both are symptomatic of a dire need for the services of a psy-
chotherapist. There is no doubt that continuities stemming from group membership
have sometimes been portrayed as if they were trait nuggets implanted in individu-
als. I suspect that the metaphor of internalization fosters a proclivity for reading the-
ories that report group-specific patterns of attributes or conduct in this simplistic
way—as if our psyches swallowed social inputs whole and never metabolized them.32
But it is also true that the contrarian view of identity exaggerates the flux of subjec-
tivity and, more important, overlooks faculties of self-definition.
Reprising some of the tropes of intersectionality highlights the intersectional
subject's capacity for self-definition. Crenshaw's crash metaphor polarizes groups
into axes of harm or axes of benefit. But McClure's image of group membership
complicates this picture by suggesting that axes of harm are not exclusively harmful.
Despite long histories of oppression, subordinated social groups have sustained tra-
ditions that are worthy of respect and that many group members willingly uphold.
Thus, belonging to a group is often a source of sustenance and guidance. Still, as
Lugones's trope of border habitation implies, belonging to a group does not entail
total immersion in that group's culture and norms. To have an intersectional identity
is to belong to a number of groups and not to belong wholly to any. Thus, group
members can use their experience of alienation to gain critical insights into the
norms and values of different social groups.33 Responding to the concern that a purely
negative critique condemns people to unchanneled rage or unavailing despair,
Mouffes trope of articulation endows people with a capacity for piecemeal synthe-
sis and provisional reconstruction. Although group-based identity is neither expun-
gible at will nor free of liabilities, persons with intersectional identities are plainly
not helpless captives of social groups.
To clarify the view of self-definition these identity tropes adumbrate, I begin by
sketching an example. Worldwide, almost all women want to have children, and
most of them become mothers. Yet, studies of U.S. women provide abundant evi-
dence that few of these women autonomously choose to become mothers; most re-
164 Autonomy and the Social

gard motherhood as their destiny and proceed automatically to fulfill it.3/l I submit,
however, that reflecting on this undertaking from the standpoint of intersectional
identity would provide one route through which autonomous motherhood might
be achieved.
Consider a young, white, middle-class, heterosexual (let us suppose, recently
married), Italian-American woman. What perspectives on motherhood do her vari-
ous group memberships afford her? Her gender, sexual orientation, and ethnicity
collude in propelling her toward pregnancy and motherhood. "Woman," as Simonc
de Beauvoir reminds us, "is a womb";35 the telos of female heterosexuality is child-
bearing; Catholicism-infused Italian-American culture reveres motherhood. Plainly,
these identity codctcrminants— - the ones she is most likely to be conscious of—
converge on and affirm her becoming a mother.
What about her race and class identities? They, too, support: this choice, but not
quite so unambiguously. The ubiquitous (albeit more and more outdated) image of
U.S. middle-class life—the happy, prosperous suburban family— includes a hetero-
sexual couple and a pair of kids. It is less obvious, though no less sure, that her white
racial identity reinforces the motherhood norm. A century ago, it was still respect-
able for American public figures to urge white women to procreate to maintain
white racial dominance. 36 Such blatantly racist doctrines are no longer proclaimed
publicly, but they contaminate white facial heritage and surface occasionally in the
fulminations of right-wing fanatics. Notice, though, that if my hypothetical woman
attended to her racial identity and exposed the nature of the support: it proffered for
motherhood, it might give her pause. Unconscious racism remains common among
white middle-class Americans, but naked racism is hard for most of them to stom-
ach. Likewise, inspection of the class backing for maternity might raise an alarm.
That every baby bom into the American middle class can be expected to consume a
colossal chunk of the planet's resources might prompt her to consider abstaining
from bringing another resource glutton into the world.
I am not suggesting that attention to intersectional identity would convince
women to stop having children. On the contrary, many women, including quite a
few feminists, see mothering as a practice worthy of celebration and as a legacy that
ought to be self-consciously and joyously perpetuated. What I am suggesting is that
intersectional identity both reinforces conventional gender expectations and affords
opportunities to reassess those norms. Linking oneself to white racial status or to
middle-class economic status exposes reasons militating against conformity. Most
important, perhaps, the tensions between the different dimensions of intersectional
identity introduce a wedge of optionality that authorizes individualized reflection
and choice. In the end, a woman might refuse on principle (antiracist or ecological)
to reenact the feminine maternal norm, but alternatively she might conclude that
the satisfactions of motherhood would probably outweigh the negatives and decide
to have children.
I have not attempted to canvass every conceivable way in which intersectional
identity might figure in an autonomous decision about becoming a mother. How-
ever, the selection of factors 1 have presented suffices to demonstrate that inter-
sectional identity contributes to autonomy by connecting individuals to systemic
social relations and to the social meanings of those structures. The light social inter-
 Intersectional Identity and the Authentic Self 165

pretation sheds on seemingly personal conduct makes room for autonomous self-
definition— thoughtful clarification or reshaping of one's desires, personal traits,
values, interests, and goals—and thus for autonomous self-direction—plotting a
course of action that enacts those attributes as fully as possible.
As members of social groups with worthy traditions, individuals may find in-
spiration and guidance; as border dwellers between social groups, they may find crit-
ical tools in one tradition to apply to another; as articulators linking diverse groups,
they may find ways to synthesize and reconstruct familiar traditions. But how do
they do this? And what does this say about authenticity?
First, it is important to underscore the fact that in the example I sketched the
most promising springboards for critical reflection are the dimensions of intersec-
tional identity that are most suppressed and least likely to be taken into account—in
this case, race and class. Most U.S. whites do not think of themselves as having any
racial identity, and they regard middle-class status as virtually universal and therefore
nothing special. Thus, avowing an intersectional identity with race and class as code-
terminants itself requires an initial process of autonomous self-definition.
Second, embracing such an intersectional identity would not necessarily lead a
woman to the critical levers I have described. She might perceive her race as neutral
and her class as an unalloyed blessing. Thus, gaining access to the critical potential of
group-identity determinants requires scrutinizing and analyzing the history and cur-
rent position of the social groups to which one belongs—another process of au-
tonomous self-definition.
Finally, consider a woman who has engaged in the first two forms of self-
definition and who has decided against motherhood but who is beset by residual
conflicts between that choice and her identity as a woman, a heterosexual, and an
Italian American. Her independence of mind and will notwithstanding, her gender,
sexual orientation, and ethnicity continue to construct her as a mother and thus
wage a psychic guerrilla campaign against her commitment to her chosen course. Let
us suppose that she confronts these challenges to her decision and, defying them, re-
news it. Let us suppose further that she resents the rigidity of the reproductive
norms correlated with her gender, sexual orientation, and ethnicity and that she re-
solves to put a stop to these attacks on her autonomy.37 To do this, she needs to re-
define herself as a woman, a heterosexual, and an Italian American. But, of course,
she lacks the cultural and linguistic authority to redefine these categories as she
pleases. She can vow to ignore well-intentioned associates' efforts to induce her to re-
consider; she can terminate relationships with those who cannot resist menacing her
with disparaging comments about her barrenness; she can boycott the pronatalist
media onslaught; and so on. But since she has already internalized the motherhood
imperative and since extirpating this message is tantamount to repudiating feminin-
ity as it is defined by or for these groups, individualistic strategies of self-definition
have serious drawbacks, and moreover, they are unlikely to be entirely successful. To
address this problem, a nonconformist woman would be well advised to join with
other group members and embark on a project of cultural critique aimed at recon-
figuring the social identity of their group—a project of autonomous collective self-
definition.38
For individuals, self-definition involves addressing questions along the follow-
166 Autonomy and the Social

ing lines: What sort of person am I? What groups do I belong to? What is the sig-
nificance of belonging to these particular groups? What sort of person do 1 aspire to
be? How do the groups I belong to hold me back? How do they help me? What re-
ally matters to me? What desires, personal traits, values, interests, and aims should I
seek to enact?
At the collective level, self-definition focuses on social groups, people's attitudes
toward them, and their expectations of group members: What are the social mean-
ings of being, say, a woman? What is worth preserving from traditional feminine
norms? What should be scrapped? How can we build solidarity among women de-
spite our diversity and disparate needs? What objectives should we pursue as a
group? How do we want our group to be symbolized in the popular imagination?
To accomplish self-definition, individuals must exercise a repertory of skills.39
The competency needed for self-definition includes, but is not limited to, the fol-
lowing:

1. Introspective skills that sensitize individuals to their own feelings and desires,
that enable them to interpret their subjective experience, and that help them
judge how good a likeness a self-portrait is
2. Imaginative skills that enable individuals to envisage a range of self-concepts
they might adopt

3. Memory skills that enable individuals to recall relevant experiences not only
from their own lives but also experiences that associates have recounted or that
they have encountered in literature or other artforms
4. Communication skills that enable individuals to get the benefit of others' per-
ceptions, background knowledge, insights, advice, and support
5. Analytical and reasoning skills that enable individuals to compare different self-
concepts and to assess the relative merits of these alternatives
6. Volitional skills that enable individuals to resist pressure from others to embrace
a conventional self-concept and that enable them to maintain their commitment
to the self-portrait that they consider genuinely their own, that is, authentic
7. Interpersonal skills that enable individuals to join forces to challenge and
change social norms

It is important to recognize that intersectional self-definition poses unusually exigu-

ous problems. A gesture of cultural dissicience, intersectional self-definition involves
seeking out heterodox resources within one's culture, perhaps borrowing from other
cultures, and synthesizing these materials in novel ways. Individuals who define
themselves as intersectional subjects are exceptionally dedicated and intrepid, and
they need proficiencies that are sadly uncommon.
It cannot be emphasized enough how difficult intersectional self-definition is.
Indeed, I would venture that one reason why the contrarian view of intersectional
identity is so attractive is that the obstacles to intersectional self-definition are so for-
midable. As I have said before, there is no convenient terminology—no shorthand,
 Intersectional Identity and the Authentic Self 167

as it were—for articulating intersectional identities, nor, I would add, are there nar-
rative templates that are conducive to intersectional autobiographies. For the most
part, then, people's interpretations of their psychological development, character
formation, and desire arousal sideline intersectional identity. Indeed, stories that fea-
ture intersectional subjects can seem precious and overcomplicated, if not down-
right weird. People cannot be expected to cast their gaze inward, behold their inter-
sectional identity, and intuit its import, for culturally transmitted cognitive schemas
and emotional scripts organize introspection, and these frameworks are not hos-
pitable to intersectional self-definition. To define oneself intersectionally, one must
activate competencies that mesh intellect and feeling in order to seek out and assim-
ilate nonstandard interpretive frameworks. One must be introspectively vigilant, at-
tuned to signs of frustration and dissatisfaction, attentive to baffling subjective
anomalies, and willing to puzzle out gaps in one's self-understanding. One must be
equipped to tap into oppositional intellectual currents. Curiosity about other people
and their cultures is invaluable, and so is a passion for ideas. But it would be disas-
trous to sponge up whatever one comes across, for neo-Nazis are publicizing their
views on the internet alongside benign political movements. Although one cannot
learn anything new unless one approaches unfamiliar material in a receptive and
charitable spirit, one must command critical thinking skills. Not only must one be
alert for errors of fact and fallacies in reasoning, but one must also register emo-
tional cues that signal confusion or danger. Still, extracting what is worthwhile from
newly encountered material is the key to enriching one's self-knowledge and to re-
defining oneself.40 Thus, one must be able to identify such ideas, incorporate them
into one's own cognitive and emotional viewpoint, and apply them as one defines
oneself. Yet, since securing an intersectional self-portrait ultimately requires cul-
tural transformation, one must have a way to displace conventional tropes of the self
and stereotypes of social groups, as well as a way to replace them with intersectional
tropes and emancipatory group images.
Now, individuals with different temperaments, talents, priorities, and personal
styles exhibit autonomy competency in distinctive ways. Some rely on reading or ex-
posure to other media to acquaint themselves with heterodox perspectives; others
rely mainly on social exchanges.41 Neither is superior. Some rely on informal sup-
port networks to help them develop and sustain intersectional self-definitions; some
seek out professional psychological counseling.42 Both can be effective. To catalyze
needed cultural change, some people become journalists, scholars, novelists, or artists
and publish books or articles, create artworks or advertizing, or mount dramatic
productions; others form political groups within which intersectional identities are
reaffirmed and through which social structures that reinforce monistic conceptions
of identity and repressive group identities are opposed; others become lawyers,
politicians, or business leaders and work through the system to bring about such
change.43 All of these strategies are needed. There is no universal formula or step-by-
step procedure for intersectional self-definition.
I have emphasized that intersectional self-definition takes place at both the in-
dividual level and the collective level and that self-definition at one level interacts
with self-definition at the other. The concerns that individuals identify as central to
their lives will shape a group's self-understanding, and a group's interpretation of its
168 Autonomy and the Social

social position together with the political agenda it sets for itself will have an impact
on the options that are psychologically and materially available to (or beyond the
reach of) individuals. These reciprocal effects suggest that self-definition has no ter-
minus. Groups must modify their self-concepts in response to individual needs and
initiatives, and individuals must adjust their self-concepts as collectivities recon-
struct the meaning of group membership and gain expanded opportunities for
members of subordinated groups.44 Self-definition is best viewed, then, as an open-
ended process of reflection, reconsideration, revision, and refinement, and self-
portraits are best viewed as works in progress. Authentic intersectional selves are
never finalized.

The Authentic Self as an Integrated Self

Lately, there has been quite a vogue in commending play as the prime mode of per-
sonal and political agency.''5 The values of insouciance and fun are surely needed as
a counterweight to the image of the autonomous individual as an isolated, plodding
planner, wholly lacking in panache and merriment, and also as a corrective to the
image of leftist activism as a deadening drill of solemnity and stalwartncss. Playful-
ness calls to mind relaxing with friends, as well as casual fooling around that breaks
the ice with strangers or eases strained relationships. Likewise, this mode conjures up
the frisson of dreaming up outre schemes, the allure of trying on outlandish per-
sonas, the kick of bashing the establishment, and the puckish sport of tweaking con-
vention. These gambits connect play with a key strategy for achieving autonomy
— namely, improvisation — and also with an important subjective dimension of
autonomy—that is, the exhilaration that stems from hitting on just the right enact-
ment of one's desires and values, especially when circumstances are not conducive to
expressing them at all.
Playfulness contributes to self-discovery and self-definition when it helps inter-
sectional subjects maintain intimate interpersonal bonds and forge connections
across social barriers. Moreover, unlocking one's imagination and freeing up one's
will are plainly necessary for self-definition—particularly for members of subordi-
nated groups who have been tyrannized by severely restricted opportunities and re-
pressive norms, but also for members of privileged groups who need to divorce their
picture of egalitarian social relations from their fears of loss and fantasies of subju-
gation. Yet, conflating agency with play threatens to reduce agency to the random-
ness and arbitrariness of acting on impulse. Members of subordinated groups can-
not afford to be seen this way (or, for that matter, to see themselves this way), for
being cast as a "playmate" is infantilizing (all too reminiscent of being cast as a play-
thing), and this belitdement allows others to decline to take their grievances seri-
ously. Nor can members of subordinated groups afford to see members of privileged
groups who are mobilizing to defend their dominant status as playmates, for con-
struing their retrograde politics as play would exempt them from accountability.
The trait of playfulness is an asset for intersectional subjects, but play is only part of
autonomous agency.
Members of subordinated groups should not stake their future on serendipity.
The compass of authentic individual identity is needed to guide individuals' choices
 Intersectional Identity and the Authentic Self 169

and to chart a progressive political course. If severing agency from identity is mis-
guided, though, it is necessary to clarify what it is like to have an authentic identity.
I begin by considering Harry Frankfurt's influential account of the authentic self. In
my view, Frankfurt's treatment of personal integration illustrates how the exigencies
of addressing a traditional philosophical problem can supersede awareness of
human psychological frailty and the confounding complexity of leading a life in the
real world to yield an excessively rigid conception of integration. Indeed, Frankfurt's
account shows how very difficult it is to explicate integration without appearing to
recommend personal ossification.
Harkening to the classic problem of freedom and determinism, Frankfurt treats
the authentic self as a psychic state that ensures free will once a person achieves it.
For Frankfurt, one knows what one really wants when one has a "second-order voli-
tion" that a first-order desire be effective—that is, when one identifies oneself "de-
cisively" with a first-order desire.46 Identification is an act of self-constitution, which
is accomplished through two types of decision: (1) demarcating the boundaries of
the self, that is, separating oneself from desires one does not want to satisfy at all;
and (2) internally organizing the self, that is, integrating one's competing desires into
a "single ordering" that establishes one's priorities.47 When one makes such decisions
"without reservation" and thus "wholeheartedly," the first-order desires with which
one identifies are "authoritative for the self."48 To have an authentic self is to be
wholeheartedly committed to a rank ordering of the desires one has decided to sat-
isfy and to be wholeheartedly disassociated from those of one's desires that one has
decided against satisfying. Autonomous individuals act only on their authentic de-
sires; hence they have free will.
On Frankfurt's view, an authentic self is integrated in two respects: (1) one's en-
dorsed first-order desires have been rank ordered so that one knows what is most im-
portant, and (2) one does not feel ambivalent about any of these desires or about the
priorities one has set for oneself. Integration eliminates conflict, and wholehearted-
ness complements and completes integration.
Frankfurt accounts for some key phenomena associated with autonomy. At the
beginning of this article, I listed some of the colloquial expressions people use to af-
firm their autonomy—"I feel solid," "I feel right in my skin," and "I feel at one with
myself." Frankfurt's conception of wholeheartedness nicely captures the flavor of
these expressions and the general sense of resolution, equanimity, and confidence
they evoke. In addition, theories of intersectional identity implicitly endorse one
form of integration, for they stress the urgency of opening lines of communication
between differently situated group members to prevent one segment of the group
from undertaking political initiatives that would be detrimental to another. Like-
wise, they call on group members to find shared interests that can serve as a basis for
solidarity. Thus, these theorists underscore the need for intersectional subjects to
conceptualize their group-identity determinants in inclusive terms and to integrate
this enlarged understanding of the diversity of social groups with their political
views. Frankfurt's emphasis on evaluating the worthiness of one's desires and on
sorting out one's priorities is consonant with this view of the intersectional subject's
political task.
Frankfurt's account of self-constitution through identification also seems prom-
170 Autonomy and the Social

ising for intersectional subjects. As the crash trope suggests, members of subordi-
nated groups are victimized by their group identities, although as the membership
trope suggests, they also find sustenance in their group identities. Now, it might
seem that Frankfurt's idea of identification and integration gives these individuals a
technique for coping with this predicament. They can disidentify with the harmful
attributes they have internalized, for example, self-doubt, servility, or submissive-
ness, and they can identify with the empowering attributes they have internalized,
for example, a mordant sense of humor, persistence in the face of adversity, or a love
of convivial gatherings. 49
Unfortunately, the strategy of identification and disidentification breaks down
when faced with a core feature of subordination, namely, victimization. To disiden-
tify wholeheartedly with one's victimization when one is in fact a victim of systemic
injustice is to deny social reality and to foreclose resistance. Such disidentification
may redouble the individual's vulnerability to injustice, or it may draw individuals
into complicity in their own subordination or the subordination of other group
members. Only if individuals can disavow harmful group-linked attributes, while
identifying with their position as members of a wrongfully subordinated group and
retaining their compassion for group members who have not succeeded in ridding
themselves of disabling and disfiguring group-linked attributes, is the strategy of
disidentification feasible for autonomous intersectional subjects. But since it is
doubtful that one can decisively or wholeheartedly identify with being a victim
without succumbing to self-pity or self-annihilation, it is doubtful that such an
identification could be integrated into an empowering and coherent hierarchy of de-
sires and values.511 Indeed, ambivalence toward one's victimization seems a better at-
titude to strike. Neither wholeheartedly embracing it nor eschewing it, one seems
more likely to preserve one's balance and one's autonomy.
Whereas theorizing agency as play underrates the intersectional subject's ca-
pacities fot self-discovery and self-definition, in my judgment Frankfurt's account
of authenticity and autonomy demands more self-mastery than intersectional sub-
jects can (or should try to) deliver. Intersectional identity resists wholehearted in-
tegration. As the trope of border dwelling suggests, tensions between one's various
group-identity determinants are bound to occur. Moreover, since these tensions
fuel critique and social innovation, it would be dangerous to eradicate them. 51 For
intersectional subjects, then, to attain integration would be to betray one's authen-
tic self. Moreover, I have argued that the intersectional subject's need for ongoing
self-definition precludes finalizing one's authentic self. Again, wholeheartedly set-
tling into an ordering of one's desires seems incompatible with authentic intersec-
tional identity.
The root difficulty here is Frankfurt's conception of integration. Recall that ac-
cording to Frankfurt, conflicting desires with which one identifies must be inte-
grated into a "single ordering."52 At first, this might seem unexceptionable, for it
seems that when people have conflicting desires, either they must set priorities, or
events will decide for them. But setting priorities can short-circuit autonomy. Pre-
sumably, an autonomous individual would brainstorm and try to come up with a
way to satisfy both desires before sacrificing either of them. If this is so, Frankfurt's
emphasis on linear rank ordering deflects attention from the skills that enable people
 Intersectional Identity and the Authentic Self 171

to reconcile apparently competing desires—whether by figuring out how both can

be enacted after all or by working to change the social context that brings them into
irresolvable conflict. Thus, Frankfurt discounts the activity of exercising those skills
and inflates the importance of imposing order on one's identity constituents.
It becomes apparent how grave this preoccupation with order is when one recalls
that Frankfurt's account is not limited to situation-specific autonomy. He speaks of
identifying with characteristics that endure, and he avers that integration "provides
. . . for coherence and unity of purpose over time."53 In this broader context, whole-
heartedly committing oneself to a "single ordering" would be self-defeating. There
are no desires or characteristics that should always take precedence in any individ-
ual's life. Insofar as intersectional subjects are subordinated, they need to be on the
qui vive for unanticipated opportunities. Insofar as intersectional subjects are both
privileged and subordinated, they need to be ready to extend their social critique
and build political coalitions with people whose interests are somewhat at variance
with their own. If integration is understood as a wholehearted and lasting commit-
ment to a self-chosen hierarchy of desires, personal traits, values, interests, and goals,
and if an integrated self is an authentic self, integration and authenticity are often
inimical to autonomy.54
Frankfurt might reply, however, that I am fastening on his phrases "single or-
dering" and "wholehearted commitment" and that I am reading him too literally.
When Frankfurt shifts, as he often does, from propounding an account of free will
to exploring the byways of moral psychology, it is clear that he regards the process
of achieving integration as a gradual one and that he is well aware that intrapersonal
conflict may persist despite one's second-order volitions. Nevertheless, we are left
with the question of how to understand his claims about integration.
Maybe the right geometrical image for his conception of integration is not the
vertical line but rather the pyramid. If integration has a pyramidal structure, an au-
thentic self could have quite a few desires, personal traits, values, interests, and goals
on the same level, but as one ascended toward the top of the pyramid, the levels
would accommodate fewer and fewer identity constituents. This model would give
the autonomous individual the multidimensionality and flexibility needed to re-
spond appropriately to a wide range of situations, for individuals could pick and
choose among identity constituents that they have placed on the same level, de-
pending on circumstances. Indeed, the pyramid could be flat-topped, which would
allow for the possibility that two or more projects could be of paramount impor-
tance. I submit, however, that anyone who tried to decide in advance which sets of
desires, personal traits, values, interests, and goals should always take precedence
over other sets would either give up, reverse him- or herself in the face of real situa-
tions, or live to regret sticking to his or her decisions. Moreover, if part of the point
of making decisions and becoming integrated is to eliminate conflict, as Frankfurt
claims it is,55 this model of integration would fail. Suppose an individual puts her
career and her family on the same level of her pyramid. The chances are slim indeed
—even in a postpatriarchal world—that she will never face conflicts between them.
It seems to me that the geometry of integration is rapidly leveling down to a
planar array of multifarious desires, sundry personal traits, plural values, and mul-
tiple overarching interests and goals. Yet, it is not true that autonomous people have
172 Autonomy and the Social

no idea what is more important to them or that they have no guiding ideals. The
trouble with theories of autonomy like Frankfurt's is that there is no way to codify
their self-knowledge without forcing it into a mold that falsifies it and that turns au-
tonomous individuals into cartoon figures, mechanically executing their previously
elected plans, lo be sure, autonomous people think prospectively about how much
weight to give to different identity constituents, but they are prepared to adjust their
plans, even to alter them radically, in light of new experience or unforeseen circum-
stances. If a nco-Na/i party made an unexpectedly strong showing in the next con-
gressional elections, many reclusive intellectuals who arc deeply committed to schol-
arship would autonomously set aside book manuscripts to join the battle against this
threat.
Frankfurt is right that autonomous people are not incessantly besieged by
wrenching conflicts and that they do have a sense of being integrated. I hesitate to
call it a feeling of wholeness, but that is the sort of thing nonphilosophers (includ-
ing many clinical psychologists) say. I think, however, that this sense derives from
their powers of insight and judgment and their ingenuity at devising enactments of
their desires, personality traits, values, interests, and goals that usually satisfy them;
that their associates by and large respect; and that do not regularly antagonize peo-
ple they need and care about. Although it is important to salvage the idea of per-
sonal integration, the quest for a blueprint for personal integration is futile.

Authenticity and Autonomy Competency

We need to reinterpret authenticity and to rcconceive the relationship between au-
thenticity and integration. It is a mistake to conflate authenticity with personal in-
tegration and to regard integration as a state persons must achieve as a precondition
for autonomy. Instead, we should understand personal integration as the emergent
intelligibility of an individual's autonomous self-discovery and self-definition. More-
over, since authenticity relies on intrapsychic tensions to spark personal develop-
ment and social dissent, since it is compatible with ambivalence, and since it lan-
guishes if an individual cannot tolerate provisionality and rushes to closure, having
an authentic self is best understood as the result of an ongoing activity of persons.
By exercising autonomy skills, such as the ones I enumerated previously, people take
charge of their identity and gain authenticity.
What autonomous people do to understand and define themselves is not aptly
figured by any Euclidean shape or formal reasoning pattern. It is akin to improvisa-
tional orchestration. 56 Different instruments take solos; different instruments join in
as backup; some instruments may fall silent for awhile. A tune sets the themes of the
performance. But the rendition is not preplanned, nor will it ever be repeated in
quite the same way. Still, the qualities of the instruments in the ensemble, the mu-
sicians' idiosyncratic playing styles, and the melody coalesce into a discernible logic
— an integrated dynamic, if you will. Moreover, although the group will never du-
plicate a performance, their sound — their musical identity—will be recognizable
whenever they play together.
I have shifted the accent from "integration" to "dynamic"—that is, from a psy-
chic state, instantiating a prescribed infrastructure, to an intelligible process consti-
 Intersectional Identity and the Authentic Self 173

tuted by skillful but unregimented activity. The capability behind improvisa-

tional orchestration is musicianship; the capability behind self-discovery and self-
definition is autonomy competency. The authentic self is nothing but the evolving
collocation of attributes—analogous to a musical ensemble's sound—that issues
from ongoing exercise of this repertory of skills. And an authentic self-portrait is an
interpretation of that evolving collocation—analogous to an astute band member's
commentary on the groups performance career.
This reorientation makes sense both of the intersectional subject's capacity for
authentic identity and of some major obstacles to intersectional authenticity. It is
evident that most people have acquired some degree of facility with most of the
skills needed for self-discovery and self-definition. Thus, there is hardly anyone who
has no authentic identity whatsoever. However, some individuals' circumstances in
life strongly discourage them from acquiring this competency and virtually prohibit
them from exercising it, whereas a few lucky individuals find themselves in environ-
ments that cultivate their autonomy skills and that shower them with opportunities
to exercise their competency. Consequently, different people are more or less profi-
cient at exercising these skills, more or less inclined to exercise them, and more or
less successful at coordinating their skills into a smoothly functioning competency.
Not surprisingly, then, intersectional subjects occasionally feel somewhat estranged
from themselves or even thoroughly alienated from themselves. Yet, there are also
moments when they are able to summon their skills and make enormous strides in
self-discovery or self-definition.57 Exposure to the concept of intersectional identity
itself might precipitate a consolidation of authentic identity. So might discovering
an unrealized potential or entering a new social milieu. Still, it must be emphasized
that these advances are gains in authenticity, not because some core trait has been
disclosed and not because the individual's priorities have been decisively set, but
rather because exercising autonomy skills has drawn these attributes into the orbit of
the authentic self and "authenticated" this self-portrait.
It might seem that what happens when people exercise these skills well is that
they come to have a special feeling about the attributes they include in their self-
portrait. As I said earlier, wholeheartedness is not an implausible candidate for this
special feeling. But to resort to wholeheartedness would be to revert to a managerial-
clerical view of integration, and as we have seen, intersectional identity rules that
out. Furthermore, given the complexity of authenticity—the many different sorts
of attributes that can be authentic identity constituents, as well as the multitude of
ways any one of them can be enacted—it would be astonishing if authenticity were
correlated with a single emotion. Thus, I recommend a multipronged account of
the subjective sense of authenticity that is linked to exercising autonomy skills.
When one is adept and does not need to struggle to exercise a repertory of skills, the
activity itself is often absorbing, enjoyable, and satisfying in itself. Moreover, dem-
onstrated competence produces reinforcing feelings of confidence and security. We
have seen that self-definition involves self-exploration and self-discovery. Finding an
undeveloped aptitude or a hidden virtue can be heartening; it might even spark ela-
tion and excitement. It is important to remember, too, that autonomy competency
cushions the blow of discovering a limitation or a fault, for it enables people to
change and to minimize the ill effects of what they cannot change. Thus, a disap-
174 Autonomy and the Social

pointing revelation about themselves is not likely to plunge them into frustration or
despondency. Ideally, it will be perceived as a fascinating puzzle or a revitalizing chal-
lenge. Sometimes self-definition reconciles seemingly opposed desires, personal traits,
values, interests, or goals. Such resolutions may bring welcome feelings of relief and
repose, short-lived though they may be. Finally, it is important to bear in mind the
social skills that are enlisted in self-discovery and self-definition—autonomous in-
dividuals are equipped both to benefit from others' input and to recruit others to
their point of view. This interaction is interesting and gratifying, and the resulting
interpersonal backing for self-knowledge is reassuring.
Of course, the emotions 1 have mentioned do not exhaust the subjective expe-
rience of self-discovery and self-definition. Nevertheless, I think I have filled in this
constellation of emotions sufficiently to show that it does justice to the idiomatic af-
firmations of autonomy that I have cited, as well as the sense of wholeness that au-
tonomous individuals report feeling. I would add that since these emotions are tied
to the exercise of particular skills, they seem less arbitrary and hence less mysterious
than Frankfurt's conception of wholeheartedness.
The competency-based view of the authentic self that I am recommending dis-
solves the air of paradox around the conjunction of interscctional identity and au-
tonomy. If there is no pattern that an authentic self must fit into, there is nothing
troubling about an authentic self that harbors ambivalence about some matters and
unresolved tensions between some identity constituents. Inasmuch as unresolved
tensions can spur individuals to exercise their self-definition skills in more sophisti-
cated and creative ways, these tensions may prove to be invaluable resources for au-
thenticity. When circumstances are such that ambivalence is a reasonable response,
as I have argued it is when individuals are victimized as members of a social group,
viewing this attitude as an authentic identity constituent need not defeat autonomy.
On a competency-based view of autonomy, it is not necessary to plot out every de-
tail of one's life in advance, for one's autonomy skills enable one to address situations
on a case-by-case basis. To be sure, mixed feelings about an identity constituent ne-
cessitate skillful deliberation each time it becomes an issue. But that hardly seems a
liability if one commands autonomy competency and if no preprogrammed re-
sponse is appropriate to every occasion. Ambivalence, when coupled with autonomy
competency, may ensure supple and subtle responses.
Neither does the evolving, unfinished status of the authentic intersectional self
pose a problem for this view of autonomy. Once again, if there is no pattern that an
authentic self must instantiate, and if authenticity is not an all-or-nothing matter
but allows for degrees, the loose ends of the intersectional subject—gaps and lags
in self-knowledge, as well as approximate or incomplete self-definition—need not
be problematic. Provided that individuals are able to exercise autonomy competency
and are disposed to do so, they will gradually gain authenticity. Piecemeal authen-
ticity, I would urge, is the best that we murky, fallible human beings can hope for.
Correlatively, the competency view makes sense of the fact that intersectional
subjects need autonomy to gain the self-knowledge that they need to become au-
thentic. If the autonomy they need to gain self-knowledge is autonomy compe-
tency—not a particular configuration of the self achieved through higher-order vo-
litions about the self but rather the repertory of skills through which self-discovery,
 Intersectional Identity and the Authentic Self 175

self-definition, and self-direction are achieved—there is no vicious circularity and

no reason to be baffled. As with other competencies, one learns through practice,
and practice augments proficiency. Children learn a language by hearing it spoken
and by speaking. As they incrementally gain linguistic competency, they express
more and more complex and interesting ideas. So, too, with autonomy competency.
As one gains proficiency, one's authentic self develops, and one's self-portrait be-
comes more and more nuanced and rich. On this competency view of autonomy,
authenticity is a lifelong project, yet it is attainable, as well as desirable, for intersec-
tional subjects—that is, for us.

Notes

1. I am grateful to the editors of this volume, Catriona Mackenzie and Natalie Stoljar,
for their extraordinarily helpful comments on earlier drafts of this article. Also I have pre-
sented different versions of this article at the Conference on Feminist Perspectives on Auton-
omy and Agency at the Australian National University, at a meeting of AMINTAPHIL at the
University of Kentucky, at a colloquium of the Philosophy Department at the University of
Illinois, Chicago, at the 1998 World Congress of Philosophy, and at sessions sponsored by
the Society for Analytic Feminism at the Central Division APA and by the Society for
Women in Philosophy at the Pacific Division APA. I am grateful to the audiences at all of
those presentations for their responses to my work.
2. For skepticism about this feminist critique, see Marilyn Friedman, "Autonomy and
Social Relationships: Rethinking the Feminist Critique," in Feminists Rethink the Self, ed.
Diana Tietjens Meyers (Boulder, Colo.: Westview, 1997).
3. See, for example, Teresa de Lauretis, "Feminist Studies/Critical Studies: Issues, Terms,
and Contexts," in Feminist Studies!Critical Studies, ed. Teresa de Lauretis (Bloomington: In-
diana University Press, 1986), p. 10; Deborah K. King, "Multiple Jeopardy, Multiple Con-
sciousness: The Context of Black Feminist Ideology," Signs 14 (1988): 42-72, 72; Maria Lu-
gones and Elizabeth V. Spelman, "Have We Got a Theory for You! Feminist Theory, Cultural
Imperialism and the Demand for 'The Woman's Voice'" in Women and Values, ed. Marilyn
Pearsall (Belmont, Cal.: Wadsworth, 1986), p. 19.
4. See, for example, Jennifer Nedelsky, "Reconceiving Autonomy: Sources, Thoughts,
and Possibilities," Yale Journal of Law and Feminism 1 (1989): 7-36, 7; Diana Tietjens Mey-
ers, Self, Society, and Personal Choice (New York: Columbia University Press, 1989); Trudy
Govier, "Self-Trust, Autonomy, and Self-Esteem," Hypatia 8 (1993): 99-120, 103-104;
Seyla Benhabib, "Feminism and Postmodernism," in Feminist Contentions, eds. Seyla Ben-
habib et al. (New York: Routledge, 1995), p. 21; and, Alison Weir, "Toward a Model of Self-
Identity: Habermas and Kristeva," in Feminists Read Habermas, ed. Johanna Meehan (New
York: Routledge, 1995), p. 263.
5. For examples, see Meyers, Self, Society, and Personal Choice, pp. 3-8.
6. I should note before proceeding any further that I do not regard gender, race, ethnic-
ity, class, and sexual orientation as an exhaustive and final list of group-identity determinants.
However, there is wide consensus about the importance of these identity determinants, and
these are the ones that the scholars who have developed the concept of intersectionality have
stressed. Therefore, although I draw my examples from these sources, I do not foreclose the
possibility that there are additional group categories that function intersectionally as codeter-
minants of identity.
7. The other major contender is multiplicity. See Naomi Scheman, Engendering!: Con-
176 Autonomy and the Social

structions of Knowledge, Authority, and Privilege (New York: Routlcdge, 1993), 102-104.
However, I agree with Amy Mullin that in representing complex group-based identity as an
internal population of personlike entities, this trope misrepresents complex group-based
identity in serious ways. See Mullin, "Selves, Diverse and Divided: Can Feminists Have Di-
versity without Multiplicity?" Hypatia 10 (1995): 1-31. Moreover, I would urge that its as-
sociations with the pathology of multiple personality disorder make it an unfortunate choice.
See my " The Family Romance: A Fin-de-Siecle Tragedy," in Feminism and Families, ed. Hilde
Nelson (New York: Routledge, 1996).
8. Kimbcrle Crenshaw, "Demarginalizing the Intersection of Race and Sex: A Black
Feminist Critique of Antidiscrimination Doctrine, Feminist Theory, and Antiracist Politics,"
in Feminist Legal Theory, ed. Katharine T. Bartlett and Rosanne Kennedy (Boulder, CO:
Wcstview Press, 1991), pp. 57-58.
9. Ibid., p. 63.
10. Ibid., p. 58. For a related discussion of multiple jeopardy, see King, "Multiple Jeop-
ardy," p. 47.
11. Kirstie McClure, "On the Subject of Rights: Pluralism, Plurality, and Political
Identity," in Dimensions of Radical Democracy, ed. Chantal Mouffc (London: Verso, 1992),
p. 115.
12. Kimberle Crenshaw, "Beyond Racism and Misogyny: Black Feminism and 2 Live
Crew," in Words that Wound, ed. Mari J. Matsuda et al. (Boulder, Colo.: Westview Press,
1993), p. 114.
13. Maria Lugones, "On Borderlands/la frontera: An Interpretative Essay," Hypatia 7
(1992): 31-37, 34.
14. Chantal Mouffc, "Feminism, Citizenship, and Radical Politics," in Feminists Theo-
rize the Political, ed. Judith Butler and Joan W. Scott (New York: Routledge, 1992), pp.
373-374,381.
15. I confine my observations to current American English. Obviously, other linguistic
communities, including other English-speaking communities, may furnish richer resources
for articulating intcrsectional identities
16. Anticipating the next U.S. census, some antirace theorists are stressing the fre-
quency of mixed-race ancestry and are challenging government reliance on familiar racial cat-
egories.
17. This is only one of many reasons why the U.S. military's "don't ask; don't tell" pol-
icy is bizarre, as well as unfair.
18. K. Anthony Appiah and Amy Guttman, Color Consciousness (Princeton, N.J.:
Princeton University Press, 1996), p. 80.
19. For discussion of individualized gender identities, see Nancy Chodorow, "Gender
as a Personal and Cultural Construction," Signs2(}, no. 3 (1995): 516-544.
20. See Judith Butler, Gender Trouble (New York: Routledge, 1990), and Patricia Mann,
Micro-Politics: Agency in a Postfeminist Era (Minneapolis: University of Minnesota Press, 1994).
21. I borrow this expression from King, "Multiple Jeopardy."
22. See, for example, Naomi Zack, Race and Mixed Race (Philadelphia: Temple Uni-
versity Press, 1993), and Iris Marion Young, "Gender as Seriality: Thinking about Women
as a Social Collective," Signs 19 (1994): 713-738.
23. For discussion of different varieties of autonomy, see my Self, Society, and Personal
Choice, pp. 8-19. As will become evident, it is impossible to sharply distinguish one kind of
autonomy from another because different domains of autonomous choice and action overlap.
24. It is ironic that some advocates of the intersectional trope hold that just the reverse
is true — that an intersectional self is too fluid, too indeterminate to have an identity, let
alone an authentic identity. I take up this view in a later section.
 Intersectional Identity and the Authentic Self 177

25. See Adrian M. S. Piper, "Higher Order Discrimination," in Identity, Character, and
Morality, ed. Owen Flanagan and Amelie Oksenberg Rorty (Cambridge, Mass.: MIT Press,
1990).
26. If this is so, it blurs Anthony Appiah's distinction between the collective dimension
of individual identity and its personal dimension. See Appiah and Guttman, Color Con-
sciousness, p. 93. Intelligence, charm, wit, and cupidity—these are Appiah's examples of per-
sonal attributes—take on different meanings according to what social groups an individual
belongs to. In an anti-Semitic culture, "cupidity" has distinctive connotations when a Jew is
being described; in a sexist culture, "charm" has distinctive connotations when a woman is
being described; in a racist culture, "intelligence" has distinctive connotations when an
African American is being described; and so forth. Thus, stereotypes infect the personal di-
mension of individual identity.
27. I should stress here that membership in subordinated social groups—indeed, mem-
bership in every major subordinated social group — does not necessarily lead to shame and
self-contempt. I have more to say about subordinated social groups as sources of heterodox
values and personal pride in my discussion of self-definition below. Also, although I do not
have space to pursue this related point, it is worth mentioning that intersectional identity
could play a key role in an account of transformative experience and retrospective ratification
of political conversions. For discussion of this important and difficult topic, see Susan Bab-
bitt, "Feminism and Objective Interests: The Role of Transformation Experiences in Rational
Deliberation," in Feminist Epistemologies, ed. Linda Alcoff and Elizabeth Potter (New York:
Routledge, 1993), and Claudia Card, The Unnatural Lottery (Philadelphia, Penn.: Temple
University Press, 1996), chap. 2.
28. I do not mean to insist that every member of a subordinated social group is obli-
gated to dedicate himself or herself to resisting injustice. Some people are drawn to politics,
and others are not. Still, to ignore injustice may be to dissociate from oneself or to debase
oneself. Moreover, ignoring injustice can betray other group members. While there is noth-
ing wrong with choosing not to devote oneself to politics, denying or falsifying one's own
background and social position is problematic, and there is a great deal wrong with thwarting
others' efforts to overcome group-based subordination.
29. Piper, "Higher Order Discrimination," pp. 296, 305; for related discussion of cul-
turally normative prejudice, see Diana Tietjens Meyers, Subjection and Subjectivity: Psycho-
analytic Feminism and Moral Philosophy (New York: Routledge, 1994), pp. 51-56.
30. Martha Minow, Making All the Difference (Ithaca, N.Y.: Cornell University Press,
1990).
31. Implementing liberal principles without the insights furnished by intersectional
identity often amounts to feigning impartiality. Since impartiality is the centerpiece of the
liberal democratic ideal, proponents of traditional liberal precepts have good reason to en-
dorse intersectional identity.
32. For good measure, I have used the plural "continuities" to stress that awareness of
personal identity is not awareness of an enduring, invariant unit (who would think it is?). I
also want to stress that continuities are not uniformities and that continuities can exist along
with discontinuities.
33. If members of subordinate social groups are more likely to experience alienation
from some of their group identities than members of privileged social groups, they may be
more disposed to critically assess the values and practices of different social groups. Although
other factors may counteract this disposition and neutralize this advantage, alienation may
give individuals from subordinated social groups both a psychological incentive and an epi-
stemic boost in seeking autonomy.
34. See, for example, J. E. Veevers, Childless by Choice (Toronto: Butterworths, 1980),
178 Autonomy and the Social

pp. 15, 40-42; Lisa Kay Rogers and Jeffrey H. Larson, "Voluntary Childlessness: A Re-
view of the Literature and a Model of the Childlessness Decision," Family Perspective 22,
no. 1 (1988): 43-58, 48; and Judith N. Laskcr and Susan Borg, In Search of Parenthood:
Coping with Infertility and High-Tech Conception (Philadelphia: Temple University Press,
1994), p. 11.
35. Simone de Beauvoir, The Second Sex, trans. H. M. Parshley (New York: Vintage,
1989).
36. Actually, at that time her Italian immigrant forebears were deemed as much a threat
to white racial purity as the recently "emancipated" African-American slaves. In 1999, Ital-
ian Americans are fully assimilated whites.
37. It is worth noting that even if she herself capitulates to these reproductive norms
and starts trying to get pregnant, she may resent this impingement on her autonomy, and she
may dedicate herself to working to overturn these norms for the sake of future generations of
women.
38. Nancy Fraser sketches a complementary conception of autonomy as being "a mem-
ber of a group or groups which have achieved a degree of collective control over the means
of interpretation and communication sufficient to enable one to participate on a par with
members of other groups in moral and political deliberation; that is, to speak and be heard,
to tell one's own life-story, to press one's claims and point of view in one's own voice." See
Fraser, "Toward a Discourse Ethic of Solidarity," Praxis International 5 (1986): 425-429,
428. In a discussion of group memberships as affinities, Iris Young stresses the emancipatory
potential of asserting group difference when doing so "reclaims the definition of the group
by the group." See Iris Marion Young, Justice and the Politics of Difference (Princeton, N.J.:
Princeton University Press, 1990), p. 172. Likewise, Linda Alcoff locates agentic possibilities
of collective self-definition in viewing gender as a "construct, formalizable in a nonarbitrary
way through a matrix of habits, practices, and discourses." See Alcoff, "Cultural Feminism
versus Post-Structuralism: The Identity Crisis in Feminist Theory," in Culture/Power/History,
ed. Nicholas Dirks, Geoff Eley, and Sherry Ortner (Princeton, N.J.: Princeton University
Press, 1994), p. 114. In a similar vein, I have defended a discursive politics of collective self-
definition through self-figuration in Subjection and Subjectivity, pp. 103-115.
39. For a detailed discussion of these skills and the ways in which they must be coordi-
nated, see Meyers, Self, Society, and Personal Choice, pp. 76-91. I would like to refer readers
to an intriguing psychological discussion of the experience of control that lends support to
my competency-based approach to autonomy. Ellen J. Langer and Justin Pugh Brown ob-
serve that psychologists have generally identified experiences of control with the ability to
dictate or predict an outcome, and they argue that this conception is misguided. Reflecting
on the problematics of control and self-blame in the psychology of victims of sexual vio-
lence, they maintain instead that one experiences control when one is "mindful of the choice
one was making," that is, when one regards oneself as an able decision maker and has made
one's decision in a thoughtful way. See Langer and Brown, "Control from the Actor's Per-
spective," Canadian Journal of Behavioral Science 24 (1992): 267-275, 269, 273. Presum-
ably, individuals who have autonomy competency are more likely to view themselves as good
decision makers, more likely to exercise those skills when confronted with choices, and there-
fore more likely to feel in control of their lives.
40. I can't resist putting in a plug for liberal education here, for it seems to me that ed-
ucation designed to ensure knowledgeability about history and contemporary society and to
nurture subtle readings of fiction and biography has a pivotal role to play in strengthening
self-definition skills and thus in securing autonomy. I would stress, moreover, that liberal ed-
ucation should not be confined to colleges and universities; grade schools and high schools
should also see liberal education as their mission. Neither can I resist lamenting the vacuity of
 Intersectional Identity and the Authentic Self 179

popular entertainment. The media could be doing a far better job of broadcasting informa-
tion and innovative ideas for people to reflect on and possibly to appropriate.
41. I owe the inception of my feminist consciousness and feminist sense of gender
identity to a friend who would not let me deny the gender bias in a Fellini movie we had just
seen, and I owe a great deal of my limited grasp of my racial identity to the work of critical
race theorists.
42. I regard myself as lucky to have been associated with women's studies programs
throughout most of my professional career, for in addition to their academic contributions,
these programs constitute dissident cultural communities in which nonconformist views of
personal identity flourish.
43. Despite the relatively small audience for scholarly articles like this one, I cannot
help hoping that my writing will have an impact, however small.
44. It is perhaps advisable to issue a caveat at this point. Intersectional identity does
not, in my view, suffice to account for social critique and emancipatory politics. Intersec-
tionality is an invaluable trope, but no single figuration of the self can symbolize all identity
determinants together with the extensive repertory of intellectual, emotional, and social ca-
pabilities that contribute to social critique and emancipatory politics. Since there are different
ways in which views of the self can be monistic, monistic views of the self can be misguided
for different reasons. Not only do we need to recognize each individual's multiple group
memberships, but we also need to draw on multiple figurations of the self (for extended dis-
cussion of the virtues of the tropes of heterogeneity and relationality, see my Subjection and
Subjectivity). Although I endorse the intersectional view of the relation between individual
identity and social groups, I do not believe that this relation exhausts individual identity, and
I do not favor a monotropic view of individual identity.
45. For example, Butler, Gender Trouble; Appiah and Guttman, Color Consciousness;
Susan Rubin Suleiman, "Playing and Motherhood: Or, How to Get the Most Out of the
Avant-Garde," in Representations of Motherhood, ed. Donna Bassin, Margaret Honey, and
Meryle Mahrer Kaplin (New Haven, Conn.: Yale University Press, 1994); and Maria Lu-
gones, "Playfulness, "World"-Travelling, and Loving Perception," Hypatia 2 (1987): 3-19.
46. Harry Frankfurt, "Freedom of the Will and the Concept of a Person," The Journal
of Philosophy^ (1971): 5-20, 10-13, 16.
47. Harry Frankfurt, The Importance of What We Care About (Cambridge: Cambridge
University Press, 1988), p. 170.
48. Ibid., pp. 170, 174-175.
49. Members of privileged social groups might disidentify with attributes that rein-
force their dominant social position and identify with attributes that foster egalitarian so-
cial relations.
50. This is not to say that one should never temporarily and strategically identify or
disidentify with one's own victimization. Such identification can strengthen bonds of soli-
darity with other victims, and such disidentification can make psychic space for renewing
one's sense of self after a bruising experience. However, such tactical identification and
disidentification is not decisive, wholehearted, or identity defining, and it is not the sort of
thing Frankfurt is talking about.
51. It is sometimes argued that thoroughgoing integration—the convergence of one's
group identities into a smooth blend—is a privilege of the totally privileged. See Naomi
Scheman, "Queering the Center by Centering the Queer," in Meyers, Feminists Rethink the
Self, p. 125. Although I agree that social structures and cultural norms confer "easy" integrity
on some privileged individuals, I suspect that this outsider's picture of privilege is not alto-
gether accurate. There are sources of conflict within any group identity; I noted a case in
point with respect to white racial identity in an earlier section. The difference between so-
180 Autonomy and the Social

daily esteemed group identities and socially ostracized group identities is that the conflicts are
well camouflaged and therefore are difficult to discern in the former. Also, members of priv-
ileged social groups have little incentive to uncover them. Ironically, then, in this respect priv-
ilege makes it harder to be autonomous. Pure privilege is no panacea tor heteronomy.
52. Frankfurt, Importance of What We Care About, p. 170.
53. Ibid., pp. 171-172, 175.
54. For related discussion of Frankfurt's account of integration and the virtue of in-
tegrity, see Cheshire Calhoun, "Standing for Something," The journal of Philosophy 92
(1995): 235-260, 236-241.
55. Frankfurt, Importance of What We Care About, p. 175.
56. In a somewhat different context, Martha Nussbaum invokes jazz improvisation to
figure deliberation. See Nussbaum, Love's Knowledge (New York: Oxford University Press,
1990). I would also like to mention that the view of personal integration I propose in an ear-
lier work depicts integration as a very loose weave—"a personality marked by characterolog-
ical strands that are amenable to combination and recombination both amongst themselves
and also with various evanescent traits." Sec my Self, Society, and Personal (Choice, p. 73. Con-
strued in this way, the concept of integration functions as a beacon for self-definition. Since
integration enables autonomous spontaneity and minimizes self-reproach, exercising auton-
omy skills moves self-definition in this direction (pp. 59-75). But notice that getting carried
away and trying to achieve a tight, sharply patterned weave is counterproductive, for people
with rigid personalities cannot adjust to changing circumstances and consequently their ef-
forts to realize their values and fulfill their goals are likely to be thwarted. In short, hyperin-
tcgration impedes self-direction. Finally, I want to emphasize that, in this view, integration
does not dictate the outcome of self-definition. Integrated personalities are unique, individ-
ualized personalities.
57. Elsewhere I have discussed the possibility that there might be threads and pockets of
autonomy in a life that is not globally autonomous. See Self, Society, and Personal Choice, pp.
160-162.
 8

THE PERVERSION OF AUTONOMY

AND THE SUBJECTION OF WOMEN
Discourses of Social Advocacy at Century's End1

Lorraine Code

What have I to say about freedom and the self?

I wondered whether I should not point out that all the components of the
philosophy of the self in the West have . . . had a liberating effect . . . but
point out, too, that this philosophy was undermined by aspects unfore-
seen and at the time unforeseeable, repressive aspects having to do with
phallocentric and colonial patterns of speech. . . . Might it not be neces-
sary to do two things at once: to emphasize both the permanent value of
the philosophy of rights, and, simultaneously, the inadequacy, the limits
of the breakthrough it represented?
(Helene Cixous)2

Autonomy
Autonomy has figured prominently as an emancipatory ideal in the feminist, an-
tiracist, and postcolonial analyses of oppression, subjection, subjectivity, and agency
that are the inheritors of the philosophy of rights, as well as in the activism in-
formed by these theories. Yet while recognizing autonomy's effectiveness and its per-
sistent appeal, my entry into this discussion originates in a profound unease with
the autonomy-saturated theories and rhetorics that infuse affluent Western social-
political spaces in the late twentieth century, holding in place a regulative autonomy
ideal that, paradoxically, underpins patterns of oppression and subjection. At the
same time, and aware that in so doing it produces a contradiction, my engagement
with these issues acknowledges the point it challenges: namely, autonomy's libera-
tory promise for women and other Others, whose marginality is marked and per-
petuated by a lack of the very autonomy that many members of advantaged social

181
182 Autonomy and the Social

groups take unthinkingly for granted and that some feminists — I among them—
represent as a contestable value.3 This discussion thus works within a tension, which
it does not seek to resolve. Its aim is less to reconfigure autonomy "relationally" than
to think through some enactments of autonomy within relations that are as com-
plicit with larger oppressive structures as they are subversive of them, as constraining
of autonomy-realizing potential as enabling of it. Here I focus on advocacy relations.
But my analysis does not suppose that autonomy can easily be disencumbered of the
individualistic baggage it now carries, any more than "care" can be extracted from
its feminine associations with a private realm of intimacy to become a value neu-
trally appropriate for all.4 Rather it engages with autonomy's ongoing aspirational-
inspirational appeal while attempting to disentangle its residual promise from some
of the "perversions" of its capitalist patriarchal effects to which my title alludes.
The tension is generated, in part, out of a rift between sedimented Kantian-
derived conceptions of unified subjectivity and late twentieth-century "decenter-
ings" of the human subject, which destabilize many of the founding assumptions of
Enlightenment-liberal autonomy.'' Modern, often hyperbolized, variations on (Kant-
ian) self-transparency and self-determination frame the picture of human selves that
operates regulatively within the dominant social-political imaginary of liberal dem-
ocratic societies. That picture is remarkably impervious to contestations variously
supported in social scientific inquiry and in twentieth-century social-political theory
of the very conceptions of subjectivity and agency on which it tacitly relies.6
Such contestations suggest that the self with its attendant-constitutive responses
and responsibilities, both epistemic and moral, can (historically) no longer be the
self and (anthropologically and geographically) never was uniformly the same self
for whom classical autonomy ideals were imagined. Twentieth-century psychologi-
cal, linguistic-philosophical, and historical-material evidence radically unsettles the
ideal of unified, self-determining subjectivity. Informed by Freudian theory, Niet-
zschean genealogy, and postcolonial, post-World War II, and post-1968 political
upheavals, it contests the very idea that the subject can be as self-consciously trans-
parent to herself or himself as classical autonomy assumes, thus casting in doubt the
"controllability of our own doing" that makes self-determination a plausible goal.
Conjoined with a post-Wittgensteinian and post-Saussurean linguistic turn that
puts in question the human subject's presumed capacities to "constitute or exhaust
meaning" to the extent: that it can claim authorship of its own being, 7 these chal-
lenges to humanistic ontologies of the self destabilize beliefs in the subject's "mas-
tery" of herself or himself that classical liberal theories require. Intersecting wirh his-
torical-material refutations of classical autonomy's founding belief in circumstantial
homogeneity, whose untenability feminist, antiracist, and postcolonial theorists
have amply demonstrated, these lines of thought culminate in a presumption against
the plausibility of the conception of the self that orthodox autonomy theories pre-
suppose. They delineate a tension between, on one side, an entrenched and indeed
escalating adherence to the core assumptions of the classical ideal in the social
rhetorics of affluent liberal societies and, on the other side, theoretical and practical-
empirical demonstrations, particularly from postpositivistic social science, that the
ideal cannot hold.
Mainstream Anglo-American philosophy is not impervious to these upheavals;
 The Perversion of Autonomy & the Subjection of Women 183

conceptions of self, society, subjectivity, and agency are hotly debated issues at cen-
tury's end. To mention just two loci of contest that are germane to this inquiry:
communitarians take issue with the atomistic picture of the moral and political sub-
ject they find in John Rawls's theory of justice, arguing that it neglects the signifi-
cance of communal values in shaping moral-political lives;8 and many feminists (es-
pecially in the United States), indebted to Nancy Chodorow's work in object-relations
psychology, have adopted her terminology of "relational individualism" to elaborate
a conception of selves embedded in and defined from within intersubjective re-
lations.9 These critical projects notwithstanding, a hyperbolized version of that
presumably supplanted ideal tenaciously dominates the social imaginary—the
common sense—of white, Western societies. It descriptively configures and pre-
scriptively animates the discourses of self-sufficient individualism in which "au-
tonomous man" retains his place as an iconic figure, emblematic of an unrealistic
imperative toward self-reliant self-making.10 Exhortations about "being one's own
person" and having a coherent "life plan," which infuse twentieth-century social-
political discourse, continue to promote self-determination both descriptively and
normatively, as though its possibility and desirability were hors de question.11 This
rift between theoretical interrogations of autonomy and social practices that assume
and uphold it as an old-style universal entitlement perpetuates a discontinuity be-
tween philosophical analyses and the politics of autonomy, in an era when many
philosophers are as fully committed to addressing the exigencies of everyday lives as
to achieving analytic clarity. It maintains a division between theory and practice that
amounts, for most feminists, to a false dichotomy.
In its Kantian origins, autonomy is an achievement of Enlightenment, under-
stood as man's "emergence from . , . self-incurred immaturity" where the immaturity
endemic to heteronomy manifests itself in an "inability to use one's own understand-
ing without the guidance of another." Emancipation is a task "for each separate indi-
vidual," as the Kantian motto "sapere aude!—have the courage to use your own un-
derstanding!"—signals. It requires people to cultivate "their own minds," making it
the "duty of all men to think for themselves"12 in order to escape the shackles of het-
eronomy. Yet in his 1984 reading of Kant's essay, Michel Foucault urges twentieth-
century readers to notice the limited scope of Kant's "all." He asks: "Are we to under-
stand that the entire human race is caught up in the process of Enlightenment?"13
The negative answer Foucault expects is anticipated in Kant's assertion that the ma-
jority of people "including the entirefair sex' will find the move to maturity too diffi-
cult and dangerous.14 Kant's emphasis on the "freedom to make public use of one's
reason in all matters"15 confirms the circumscribed scope of this emancipation in so-
cieties marked by hierarchical divisions that determine whose rational utterances
merit public acknowledgement, preserving areas of heteronomy even where auton-
omy is heralded as a universal possibility. Such exclusions are differently configured in
diverse cultural-historical circumstances; they run as much along class, race, and
other lines of power and privilege as along lines drawn by membership in "the fair
sex." Yet autonomy's defenders tend to read these exclusions as inconsequential to a
"universal" release from thraldom, which nurtures a stark individualism fueled by the
silent assumption that autonomous man is free to sidestep the constraints of materi-
ality and the power of social-political structures in his projects of radical self-making.
184 Autonomy and the Social

In The Unnatural Lottery, Claudia Card articulates feminist hesitations about

endorsing this Kantian—and latterly, Rawlsian—conviction that "we can all act as
duty requires, come what may," that "our goodness (or badness) is entirely up to us,"
a conviction that minimizes the extent to which the ordinary lives of many ordinary
people are lived within circumstances so oppressive and damaging as to block all
routes to autonomous agency. She is "sceptical of Kant's apparent assumption that
the same basic character development is accessible to everyone."16 Guiding her proj-
ect, Card explains, is a belief that
it is riot enough to confront the inequities of the "natural lottery" from which we
inherit various physical and psychological assets and liabilities. It is important also
to reflect on the unnatural lottery created by networks of unjust institutions and
histories that bequeath to us further inequities in our starting positions and that
violate principles that would have addressed, if not redressed, inequities of
nature. 17

The "perversion of autonomy" I name in my title renders these circumstances not

merely descriptive but prescriptively coercive in societies permeated by the impera-
tives of a hyperbolized autonomy conjoined with individualistic conceptions of sub-
jectivity and agency. In what follows, 1 begin to think through the immobilizing,
ultimately autonomy-contesting effects of some of these injustices and coercions.
Autonomy in its hegemonic articulations, I am suggesting, is a locally specific value
and ideal that is prone to exceed its reach. As Margaret Walker aptly puts it: "A cer-
tain kind of society holds out to us, and gives some of us, to varying degrees and at
different costs, the gift of (roughly) autonomous lives."18 Because the scope of its
easy realization seems to encompass most of the social spaces inhabited by the up-
holders of individualistic autonomy—those Sabina Lovibond aptly dubs "worship-
pers of the individual"19— its local specificity often disappears from view.
My discussion therefore traces a curious rhetorical genealogy by which a hyper-
bolized conception of autonomy has become a controlling, even a distopian, ideal,
particularly in the United States and, derivatively, in Canada. This powerful (re-
gional) confluence of autonomy with individualism has issued in a cluster of regu-
lative principles whose effects are as negative as they are positive for people hitherto
marginalized. It has generated a social-political-epistemic imaginary visibly peopled
by the self-reliant rational maximizer, the autonomous moral agent, the disinterested
abstract knowet, and the rational economic man of late capitalism, to mention the
most familiar tropes; and invisibly peopled by less consequential, because less per-
fectly autonomous, Others. Thus my autonomy skepticism finds echoes in the title
of Willard Graylin and Bruce Jennings's book, The Perversion of Autonomy (which
supplies part of my title). Hete Graylin and Jennings contend that "autonomy's suc-
cess in the struggle for the American moral imagination has made it overbearing and
overweening"; that "the imperialism and arrogance of autonomy must be bridled."20
This "imperialism" fuels the tension between women's (and other Others') legitimate
struggles to claim an autonomy that has never been theirs, as these struggles con-
front perversions of autonomy in the affluent parts of the Western world.
The tetms of the tension ate writ large in an opposition between stark auton-
omy and a cluster of cpistemic and moral practices I call "advocacy," which include
 The Perversion of Autonomy & the Subjection of Women 185

representing, arguing for, recommending, acting, and engaging in projects of in-

quiry in support or on behalf of someone or some group of people, whether episte-
mologically or morally. My interest is in how advocacy plays into the politics of
knowledge, where "advocacy research" becomes a denigratory slogan with which to
dismiss research that appears to serve "special" as opposed to universal (read: "dom-
inant") interests and where credibility is conferred in unevenly distributed practices
of acknowledgement that claim the power to recognize the experiences of would-be
knowers as knowledgeable, grant their testimony fair hearing, and enable their par-
ticipation in authoritative and emancipatory epistemic practices. Nonstandard (non-
male, nonaffluent, nonwhite) members of epistemic communities often occupy
"starting positions" from which they cannot achieve those outcomes "without the
guidance of another."
Advocacy and autonomy emerge as oppositional in this "speaking for" sense in
the individualistic discourse that shapes the social welfare debates of the 1990s, in
which the mythologized "individual" achievements of autonomous man produce as
their foil and negative counterexample a reinstituted subjection of women. Women
come to be judged as unreliable knowers, women on welfare as failing to meet a
standard of civic self-sufficiency, women seeking child care as inadequately autono-
mous in assuming responsibility for their "own" choices, all represented as "entirely
up to them." Structurally, the process recapitulates the subjection to paternalistic
tutelage John Stuart Mill sought to counter more than a century ago.21 Against this
autonomy-advocacy opposition, I am working to construct an ecologically modeled
conceptual apparatus that will counter autonomy's hyperbolic excesses. My project is
both diagnostic and prescriptive. Locating itself "down on the ground," it is in-
formed as much by the rhetorical effects of autonomy imperatives in people's lives
and situations—what I call their "habitat"22—as by philosophical-conceptual an-
alyses of autonomy "as such."

Epistemology
In Anglo-American mainstream theories of knowledge, this hyperbolized autonomy
ideal permeates and legitimates the discourses of impersonal mastery that trade
on an image of autonomous man as a ubiquitous, invisible expert-authority, who
stands above the fray to view "from nowhere" the truths the world reveals to a mind
prepared. Epistemic autonomy is a multifaceted ideal, which I paint here with broad
strokes. I discern it also where "autonomy" was not the term of art that the philoso-
pher^) used, yet where the effects of the ideals that animate the theory justify the
translation. The autonomous knower escapes the governance of the body, tran-
scends reliance on the senses, to cultivate reason freed from every distracting influ-
ence. Neither for a Cartesian disembodied reasoner nor for a merely incidentally
embodied empiricist knower do knowledge-productive effects flow from the speci-
ficities of embodiment. Epistemic autonomy celebrates an escape from the influ-
ences of Bacons "Idols,"23 and latterly, too, from the influences—the governance
—of location and the particularities of experience and identity. Hyperbolized, this
facet of autonomy becomes the stark objectivism that is the keystone of twentieth-
century scientistic positivism, in which it is abstract individuals who know, each a
186 Autonomy and the Social

potential surrogate for every other because they are never individuated. Continuous
with Kant's injunctions, autonomy is also about thinking for oneself, having the
courage to use ones own understanding. Epistemic self-reliance is its watchword:
freedom from dogma, opinion, or hearsay and from subjection to the heteronomy of
higher authority, be it sacred or secular.
Hie appeal of epistemic autonomy is clear. Transcending the confusion of sen-
sory, social, emotional, and locational particularity is the Holy Grail that inspires the
epistemological project, promising the certain knowledge that only objective de-
tachment can yield. And within controlled circumstances, success has seemed to be
possible. As physical and technological science extend "man's" knowledgeable con-
trol of his environment, they fuel this autonomous dream, perpetuating the social
imaginary that sustains it and disqualifying more frankly situated contenders from
recognition as knowers.24 Epistemic autonomy legitimates mastery over the "exter-
nal" world, generating structures of authority and expertise, as the power to predict,
manipulate, and control objects of knowledge — both human and nonhuman—
informs and guides inquiry. It is but a short step to the place where autonomous
man in his epistemic robes claims a responsibility to think and know for Others too
immature to escape the constraints of heteronomy—to know "their own" interests
or to understand their experiences—even as he separates responsibility for the uses
to which his knowledge is put from a quest for knowledge that speaks for itself and
is good for its own sake.
Yet the tension holds. Successor epistemology projects committed to destabiliz-
ing power-infused patterns of epistemic authority often incorporate campaigns for
epistemic autonomy by or on behalf of those who continue to dwell in (enforced)
tutelage: autonomy as the "freedom to make public use of one's reason," the "courage
to use [one's] own understanding." Epistemic heteronomy has consistently been
women's lot under patriarchy, manifested at its most outrageous in a conviction that
women cannot even know the truth of their own experiences. Small wonder, then,
that autonomy persists as a feminist goal, despite the often antifeminist conse-
quences of its hyperbolic invocations.
Testimony's uneven successes in claiming acknowledgement are at the heart of
the issue. Of the classical sources of empirical knowledge—perception, memory,
and testimony—testimony ranks a distant third on a scale of epistemic respectabil-
ity. Greater esteem accrues to the putatively more direct, more self-reliant processes
of perception and memory. Testimony's shuatcdness, its inescapable positioning as
someone's speech act, locates it closer to opinion, to hearsay rather than to sanitized
— thus presumably more trustworthy—sources of knowledge in controlled obser-
vation conditions. Its source in the specificities of experience puts testimony's de-
tachment and replicability in question. Hence denigrations of testimony persist in
epistemology and the politics of knowledge, insisting on and reinforcing the illusion
of cognitive autonomy. Testimony challenges that illusion; it stands as a constant re-
minder of how little of anyone's knowledge, apart from occurrent sensory input, is
or could be acquired independently, without reliance on others. Testimony—in the
elaborated sense of learning from other people; from the cultural wisdom embedded
in everyday language; and from books, media, conversation, journals, and standard
academic and secular sources of the information—makes knowledgeable living pos-
 The Perversion of Autonomy & the Subjection of Women 187

sible. Thus it is curious to discredit testimony because it offers evidence not achieved
"independently" and "at first hand." Most epistemic negotiations and many justifi-
catory strategies take testimony as their starting point, as they engage questions
about who knows and how. Their resolutions have as much to do with testimony's
variable credibility as with perception and/or memory singly conceived. Yet despite
overwhelming everyday evidence, showing that people are fundamentally reliant on
testimony and advocacy, the image of the self-reliant knower confronting the world
directly remains integral to and constitutive of the dominant model of knowledge
in mainstream epistemology. Indeed, if that image could not be held intact, the
basic tenets of the system would also fail to hold. With other feminist epistemolo-
gists, my purpose is to contribute to the evidence that they cannot.
In standard epistemologies, it is in its "flattened" moments, when it reports on
—replicates—simple (Sknows that/)) knowledge claims and can be assimilated to
aural perceptual-observational knowledge, that testimony is most amenable to for-
mal analysis: "Jane said that she knew Dick had fed Spot."25 Here it claims its great-
est credibility, at scant remove from perception and memory. But when it moves
closer to concrete and frankly personal reporting, testimony no longer claims the
replicability required of orthodox empirical knowledge, and thus its credibility
diminishes.
Yet testimony is seldom concerned with generalities, with universals. It is
about particularity, about concrete things that "merely happen," a feature that ex-
acerbates its uneasy relations with objectivist theories of knowledge. Thomas Nagel
puts the problem well: "Something in the ordinary idea of knowledge must explain
why it seems to be undermined by any influences on belief not within control of
the subject—so that knowledge seems impossible without an impossible foundation
in autonomous reason."26 Because a morally sensitive epistemology of particularity
that pays responsible attention to testimony cannot claim such a foundation, it ap-
pears to face two choices. Either it must relinquish its claims to count as epistemo-
logical inquiry, or it must push against the boundaries of the orthodoxy to demon-
strate its effectiveness in addressing issues on which moral and epistemic agents
require guidance. Here I opt for the second alternative: for an epistemology-morality
that (borrowing Margaret Walker's words) "bears a far greater descriptive and empir-
ical burden, in pursuing details of actual moral [and epistemic] arrangements, than
it is commonly thought to entail."27 In so doing, I challenge a philosophical assump-
tion that beginning in particularity entails remaining stuck there, enmeshed in the
minutiae of the concrete and unable to escape the merely experiential.
Anne Seller's claim that "as an isolated individual, I often do not know what my
experiences are" and Ludwig Wittgenstein's claim that "knowledge is in the end
based on acknowledgement" capture the limitations of the autonomy-of-knowledge
credo as it discredits the testimony that makes experiential knowledge possible.28
Gaining acknowledgement requires more than the courage "to use one's own under-
standing," more than an individual resolve to make free "public use of [one's] rea-
son," in a hegemonic social imaginary that discounts testimony because of the tes-
tifier's failure of self-reliance or of her or his disenfranchised position in the social
order. Membership in "the fair sex" is but one such position: a sex whose very "fair-
ness" annuls expectations of fair acknowledgment.
188 Autonomy and the. Social

Advocacy
Modern science and orthodox epistemologies established themselves in a founding
gesture in which, as Lynette Hunter observes,

neutral logic and pure language replaced "rhetoric." [And science assumed] a
rhetorical stance that denied that the rhetoric of social persuasion in historical
context was needed, even denied that it existed . . . built upon the concept of the
"universal autonomous man," able to communicate infinitely replicable experi-
ence [and] isolation from the social world, wbich might contaminate the purity or
challenge the totality of explanation. . . . Its rhetoric claims that there is no need for
rhetoric^

With testimony at its epistemic core, advocacy defies these assumptions. Hence,
transgressing the boundaries that claim to separate objective knowledge claims from
rhetoric, it moves to the margins of epistemic legitimacy.
Like testimony, advocacy bears the taint of the Kantian claim that emergence
from beteronomous immaturity requires using "one's own understanding without
the guidance of another." To the media-informed eye, legal contests, whete the aim
is to ensure that advocacy triumphs, confirm the mutually antithetical character of
advocacy and truth. Despite legal limitations on misrepresentation, lawyers' well-
known commitments to advocating for their clients reinforce the belief that advo-
cacy, by definition, abdicates respect for truth. Analogously, in medical discourse a
sharp distinction between measurable, diagnosable symptoms and context-providing
narratives proclaims a connection with advocacy that disqualifies narratives from
contributing truth-producing evidence to a physician's medical knowledge.30 In
moral-political discourse, advocacy is cast as the villain behind conflicts of interest
and ideologically driven lobbying, evoking the "interest group" label that invites
condemnation in the social rhetorics of equal (impartial) treatment for all, in which
impartiality is objectivity's analogue.
Empiricist epistcmologists, for whom facts are self-presenting—they need no
one to speak (advocate) for them — confirm that advocacy obstructs objectivity.
There is no need to see facts through anyone else's eyes, and an autonomous knower
will refuse such (heteronomous) dependency. Such assumptions fuel charges that
represent "advocacy research" as inquiry in the service of "special interests," which
sacrifices objectivity to those interests. Peter Novick, writing of historical knowl-
edge, summarizes the position well: "The objective historian's role is that of a neu-
tral, or disinterested, judge: it must never degenerate into that of an advocate or,
even worse, propagandist."31 The juxtapositions he invokes display advocacy's pre-
sumed proximity, for epistemologists, to the worst rhetorical excesses.
Nor are such worries unfounded. Advocacy operates variously; sometimes, un-
doubtedly, it "shapes the truth" to serve specific ends, truth becomes subservient to
those ends (which are themselves often nefarious), and "propaganda" is the appro-
priate label; exposure and censure the required responses. Yet this is neither advo-
cacy's only nor even its principal function. In the examples I address, advocacy
practices endeavor to get at truths that operate imperceptibly and implicitly, below
the surface of the assumed self-transparency of evidence. They are strategically ef-
 The Perversion of Autonomy & the Subjection of Women 189

fective in claiming discursive space for "subjugated knowledges," putting knowl-

edge into circulation where it can claim acknowledgment and working to ensure
informed, emancipatory moral-political effects.32 Because advocacy's epistemic and
moral-political aspects are intertwined, responsibility is central to its respectable func-
tioning. Advocacy is at once dependent on the quality of knowledge that informs it
and on the level of epistemic responsibility and the climate of trust in which it cir-
culates. Just and fair interpretive judgment of testimonial claims is at issue. Advo-
cacy practices rely for their "goodness" on the expertise of the advocate(s) and their
responsiveness to testifiers, who in their isolation, often require the guidance of
others to know what their experiences are. It makes truths available for negotiation
with, by, and for people diversely positioned in an epistemic hierarchy that erro-
neously assumes universal, autonomous access to the evidence and the acknowl-
edgement on which knowledge depends. At its best, then, advocacy can counter-
balance the patterns of incredulity into which the testimony of marginalized knowers
tends to fall.
My interest, specifically, is in women's and other Others' positionings within
advocacy relations that promote or thwart autonomous epistemic stances. But this is
a collective, collaborative autonomy; it is not "relational" in the more enclosed sense
that derives from object-relations theory. It is enacted in interpretive, "advocative"
communities of which one-one dyadic relations are only one minor example. De-
mocratically engaged, advocacy relations foster negotiations aimed at countering
the isolation that inhibits the experiential knowledge of which Anne Seller writes
and promoting the acknowledgement that Wittgenstein declares the sine qua non
for establishing knowledge.
At the confluence of ideal autonomy and individualism, the epistemic dichot-
omy between autonomy and advocacy is sustained and runs parallel to two further
regulative dichotomies: the fact-value dichotomy and the justification-discovery di-
chotomy. Within their confines, bona fide knowers stand judiciously and impartially
outside processes of discovery, which are represented as temporally prior to and sub-
stantively distinct from knowledge and knowing proper. These knowers distance
themselves from the guesses, trials and errors, persuasions, and negotiations integral
to "contexts of discovery." Discovery is conceived as a messy, value-infected process
whose planned obsolescence ensures its invisibility in "contexts of justification": the
evaluation, adjudication, verification, and/or falsification that are the proper busi-
ness of epistemology. Here values and human interests are rigorously expunged to
allow facts to stand forth in their crystalline purity. Thus even though values indeed
permeate processes of discovery, knowledge as product can be cleansed of their dis-
torting influence in autonomous, impartial justification.
Feminist epistemologists, together with naturalistic and other successor episte-
mologists, however, have reclaimed the epistemic centrality of practices of discovery,
both institutional and everyday, moving them back into focus as places where situ-
ational specificities shape knowledge-producing projects at every level. Demonstrat-
ing that politically infused assumptions structure most research projects from the
bottom up, if often tacitly, feminists expose the elusiveness of value neutrality, even
as they show that mainstream epistemology is itself as value-laden as the "special in-
terest" projects it denigrates for their explicit commitment to values and interests. As
190 Autonomy and the Social

Donna Haraway, with characteristic irony, notes: "the messy political does not go
away because we think we arc cleanly in the zone of the technical."33
These feminist arguments contest the emotivism that represents values and in-
terests as merely noncognitivc expressions of feeling. Values, background assump-
tions, and locational specificities that inevitably inspire and inform inquiry are as
open to critical evaluation as are its "products." Indeed, as Sandra Harding puts it,
the "strong objectivity" that characterizes many feminist projects refuses to allow in-
quirers "to be unconcerned with the origins or consequences of their problematics
and practices or with the social values and interests that these problematics and prac-
tices support."3'' Strong objectivity demands a breadth of scrutiny that encompasses
both the specificities of the standpoint of the knower and the "nature" of the known
in order to fulfill its claims that avowedly engaged, politically committed investiga-
tions can yield well-warranted conclusions.
"Naturalized" epistemologists—even those who claim no political investment
—are comparably committed to abandoning the quest for a priori necessary and
sufficient conditions for knowledge in general in their studies of how epistemic
agents actually produce knowledge, variously, within the scope and limits of human
cognitive powers as these are revealed in the same projects of inquiry. 3 '' Naturalism
(recalling Walker's words) takes on "a far greater descriptive and empirical burden, in
pursuing details of actual [epistemic] arrangements" than standard Anglo-American
epistemologies have tended to do. Contrary to some of its critics, it is not merely de-
scriptive, for it aims to achieve good descriptions in order to bring its normative re-
quirements within the range of human (natural) cognitive abilities. 36 Indebted to
the work of W. V. O. Quine, many naturalists examine knowledge production
within cognitive science laboratories to establish their conclusions. In my work, by
contrast, I loosen the connections between naturalism and science (especially in its
scientistic aspects) to propose that secular knowledge making supplies "more natu-
ral" sites for studying knowledge production in its local specificity and its global ef-
fects. Here, advocacy relations are those sites: exemplary of places and processes
where knowledge is collaboratively negotiated.
I examine advocacy in two locations: first, in medicine, where philosophers
often discuss it as a moral issue that either compromises or enhances autonomy, yet
where I read it also as an epistemological issue; and second, in the discourses of so-
cial welfare that work from a presumption of individualistic self-reliance that gen-
erates coercive, regulative autonomy imperatives. These examples are about the
functioning of advocacy in a relational nexus in which lines of epistemic power and
privilege structure the constitutive relations.

Medicine
Advocacy tends to appear as an ethical issue in the literature of theoretical medicine,
where the commonly named advocates are nurses, physicians, patient or lay advo-
cacy groups, lawyers, and pharmacists.37 Sometimes they—especially nurses—act
as intermediaries, from patient to doctor or to the medical establishment, repre-
senting the patient's interests, which for structural reasons the patient is unable to
represent autonomously (for herself or himself). Often (though less so in patient
 The Perversion of Autonomy & the Subjection of Women 191

advocacy groups), the relationship is structured around assumptions of a patient's

passivity, with autonomy as the preferred contrast and advocacy as the route toward
achieving it. The language of "representing interests" may appear to contain the is-
sues within a moral-political frame. But closer reading shows that they are as signif-
icantly about knowledge—experiential knowledge—about facilitating its passage
through structures of public visibility and scientificity en route to claiming the ac-
knowledgment that makes informed moral deliberation possible.
Of interest to feminists has been a path depicted as running from inarticu-
late patient to appropriately "experienced" nurse to scientifically knowledgeable
doctor, following an ascending line of legitimation that culminates in and con-
firms the commonsense belief that "doctor knows best." Here the (presump-
tively) female nurse acts as an intermediary, advocating from patient to (presump-
tively) male doctor. Assuming that the nurse's "special understanding" enables
her to promote the patient's "autonomy, dignity and best interests," this picture at
once confirms the nurse's position within a care-giving sphere and contains her
within gender-specific patterns of deference to the doctor's acknowledged ex-
pertise.38 Peta Bowden's nuanced analysis shows how this capacity of sensitive
nursing practice can perform what I call the advocacy that enables patients to "in-
terpret their dependency as harmonious with an established context, yet free of
conflict with celebrations of autonomy and independence in the wider culture."39
From a feminist point of view, the politics of these discussions — of which Bow-
den is well aware, yet which I cannot discuss fully here—are contestable in their
confirmation of the power structures of traditionally gendered divisions of labor.
But the issue of how testimony-based knowledge can cross the threshold of ac-
knowledgement is salient; clearly the autonomous knower—even of his or her
"own" experiences—is a rare bird. Speaking from her experiences of physical dis-
ability, Susan Wendell observes: "My subjective descriptions of my bodily expe-
rience need the confirmation of medical descriptions to be accepted as accurate
and truthful." 40 The question is how such experiential knowledge could achieve
acknowledgment; the answer is not to be found in the autonomy-of-knowledge
credo.
Wendell's observation thus invites the stronger claim that even with pain and
suffering—experiences that are quintessentially "my own" and that form the cor-
nerstone of epistemic "privileged access"—I cannot, as an autonomous knower,
know what my experiences are. For people who are traditionally silenced, marginal-
ized, and discredited—women, blacks, the poor, the disabled, the elderly—these
problems are acute. Thus Elaine Scarry eloquently confirms the isolation of experi-
ential certainty:
For the person in pain, so incontestably and unnegotiably present is it that "hav-
ing pain" may come to be thought of as the most vibrant example of what it is to
"have certainty," while for the other person it is so elusive that "hearing about
pain" may exist as the primary model of what it is "to have doubt."41

Acknowledgement is far more elusive than arguments from first-person privileged

epistemic access can allow; here epistemic autonomy recedes to a vanishing point.
Scarry observes:
192 Autonomy and the Social

The success of the physician's work will often depend on the acuity with which he
or she can hear the fragmentary language of pain, coax it into clarity, and inter-
pret it ... many people's experience . . . would hear out the . . . conclusion that
physicians do not trust (hence, hear) the human voice, that they in effect perceive
the voice of the patient as an "unreliable narrator" of bodily events, a voice which
must be bypassed . . . so that they can get around and behind it to the physical
events themselves. /l2

When that voice speaks from within stereotypically sustained patterns of incred-
ulity that in the dominant social imaginary tell against the veracity of its tellings
because of the marginal status of the teller, it is no wonder that autonomously
knowing even one's own experiences can neither be taken for granted nor achieved
without advocacy.
In short, I am claiming that practices of advocacy often make knowledge possible
within the hierarchical distributions of autonomy and authority in Western soci-
eties. The point comes across clearly in a cardiac condition known as syndrome X
and in chronic neurological-muscular pain known as fibromyalgia: two afflictions
more common for women than for men.43 Because neither pathology is readily de-
tectable in standard diagnostic procedures, patients who present their symptoms to
a physician—because they know that there is something wrong with them—arc fre-
quently dismissed as complaining of a pain that is "all in their minds," and therefore
not real.'1'1
Advocacy is the episternic (the politics-of-knowledge) issue. The point is not
that practitioners of "normal science" (normal medicine), where male experiences
silently establish the norm, should know that women's experiences might manifest
differently from men's, and thus that women should not merely receive instructions
to live with their symptoms. These are hollow "shoulds" when the authority of tex-
tual interpretation is confirmed by the best scientific standards, according to which
symptoms that do not "fit" count as aberrant and anomalous/' 5 The force of the par-
adigms that govern medical knowledge is such that individual dissenting voices,
whether of patients or physicians, have scant hope of claiming a hearing/ 16 Only a
substantial and authoritative chorus of interrogating voices can initiate research that
could dislodge an established procedure or check the inertia of established practice.
It is not enough for people autonomously (separately) to maintain the veracity of
"their own" experiences; nor is women's failure to gain autonomous acknowledg-
ment a sign of their immaturity as members of the fair sex, dwelling still in tutelage.
Thus the stripped-down fiction of uniformly perceptive, autonomous knowers
confronting self-announcing facts has to yield to a picture of knowledge construc-
tion as a social-communal-political process, in which items of would-be knowledge
are embedded in discourses informed by interests, hierarchies of power and privi-
lege, uneven credulity, and the pragmatics of giving or withholding trust. Such an
epistemic terrain becomes a fertile ground for advocacy to demonstrate its effective-
ness in opening out rhetorical spaces prepared for the insurrection of subjugated
knowledges.^1
Consider the following example. Speaking of syndrome X, Robin Henig notes
that in response to persistent and growing patient activism, once "a handful of car-
 The Perversion of Autonomy & the Subjection of Women 193

diologists" began to hear the hitherto muted voices of these "frustrated heart pa-
tients," a massive clinical trial was launched in the United States, all of whose 140,000
subjects are women.48 Analogously, when Harvard Women's Health Watch corrobo-
rates these women's experiential reports, the credibility index of the syndrome X
story rises.49 The conclusion again suggests itself that advocacy research, empiri-
cally grounded, scientifically conducted, and politically enacted—but advocacy
nonetheless—is necessary if such breakthroughs are to disrupt the sedimented
knowledge that informs institutional power structures.
For feminism's detractors, however, any hint that research is advocacy-driven
condemns it as deficient in objectivity and, therefore, in credibility. To cite just one
example, Christina Hoff Sommers criticizes a study of rape in contemporary U.S.
society with the claim that "High rape numbers serve gender feminists by promot-
ing the belief that American Culture is sexist and misogynist." She contends: "We
need the truth for policy to be fair and effective. If the feminist advocates would
stop muddying the waters we could probably get at it."50 The tacit contrast here is
instructive: Sommers assumes that research that is not explicitly feminist operates
from no preconceived agenda, hence, apparently, that objective inquiry amounts to
random fact finding designed to serve no specific ends. Yet this assumption sits
oddly with her belief that projects designed to show that the rape statistics are un-
realistically high will, by contrast, be innocent of political commitments.51 The con-
trast is an odd one, commonly drawn to distinguish research that implicitly con-
firms the status quo (and thus claims apolitical status) from research that contests
received truths from an explicitly declared standpoint or situation. The rhetoric of
"interest group advocacy" preserves the presumption of disinterestedness on the part
of unmarked, dominant inquirers.
The issue is complex. Politically informed theorists are well aware of the advo-
cacy research that promotes the interests of drug companies, governments, the arms
industry, manufacturers, and multinational corporations with research designed to
demonstrate the harmlessness of toxicity-producing practices too numerous to men-
tion. On the basis of these examples alone, it is no wonder that advocacy research
has a bad name. Yet advocacy research is multifaceted in its aims and agendas, as fre-
quently benign as it is malign in its findings. Think, by way of contrast, of the ad-
vocacy research devoted to developing safer contraception, fighting to eradicate pol-
lutants and toxic substances, studying environmental and workplace harms, and
investigating scientifically worthy remedies for disease. The issue, then, is not about
condemning advocacy simpliciter but about establishing deliberative communities
where inquiry is subject to public scrutiny that evaluates agendas according to larger
criteria of responsible epistemic practice.52 Without advocacy, phenomena such as
syndrome X could not achieve recognition as pathologies that require major recon-
figurations in "normal medicine."
Advocacy is thus both an epistemic and a political issue. Without advocacy, few
of "us" will become sufficiently expert to challenge the combined force of scientific
paradigms and hegemonic social imaginaries that sustain the politics of knowledge.
Nor is the problem about expertise and authoritative practice tout courtbut rather
about specific instances when authoritarian expertise operates behind a screen of au-
tonomous objectivity to paper over the gaps, the breaks in the putatively seamless
194 Autonomy and the Social

imaginaries that legitimate it. Ignorance unacknowledged becomes the thematic

issue; it takes an expert advocate to expose the gaps in the knowledge of the experts.
What was long seen as women's and other "underclass" knowers' failure to have
the courage to use their own understanding (sapere aude) emerges as a universally
unrealizable, even impossible demand. And (some, usually white) men's apparent
successes recall Claudia Card's "unnatural lottery." The ease with which they fit text-
book descriptions, with which their testimony receives acknowledgment and their
experiential utterances are taken on faith, can only be explained in terms of the
starting positions they occupy in "unjust institutions" such as those Card names.

Social Welfare
Graylin and Jcnnings's book referred to in my title does not engage with advocacy.
It locates itself, still, within the autonomy versus coercion debates that originated in
1960s protests against paternalism, in medicine and elsewhere. Autonomy run wild,
they suggest,53 denies coercion's necessary function in promoting human well-being.
Showing people what is best for them requires reclaiming a "community-oriented"
version of paternalism, which they represent as a "softened" and therefore justifiable
form of coercion.'54 The authors, however, offer no analysis of the power and privi-
lege asymmetries in societies where patterns of incredulity prevail at a level well
below the place at which their deliberations begin. In consequence, their analysis
leaves open no space between "the perversion of autonomy"—about whose features
I think they are often right—and "the proper uses of coercion"—where I think they
misrepresent the social-political situation.
A false dichotomy between autonomy hyperbolized and reclaimed practices of
social control that promote "the good of the individual" structures the analysis.''5
The authors are less concerned with the space between, where the interpretive, ne-
gotiative, and testimonial work of advocacy occurs, collectively and critically, in
communities that resist the impossible demands that an autonomy obsession exerts.
This is the space where open, responsible, and creatively critical democratic deliber-
ation, both personal and social-political, can occur. By contrast, Graylin and Jen-
nings's quest for the proper uses of coercion appears to start from a place where a
stark condemnation of advocacy reinstates the iconic figure of autonomous man as
a regulative exemplar. He reenters as the central figure in a discourse in which stereo-
types short-circuit any need to know people and their circumstances well, posing as
knowledge-achieved, that renders unnecessary the very negotiations that could make
more responsible knowing possible.
Examples abound in the politics and rhetorics of social welfare, in which an as-
sumed equality of access to social goods, that requires no advocacy, underwrites the
belief that failure to achieve autonomy is a social sin. Thus "public" measures are en-
acted to enforce self-sufficient autonomy after all, while reliance on social services
slides rhetorically into a weakness, a dependence on social advocacy that, paradoxi-
cally, invites—and receives—judgments of moral turpitude. For example, in her
poignantly titled essay, "The Unbearable Autonomy of Being," Patricia Williams
shows how the entrenched stereotype of "the black single mother" in the United
States is often invoked to disguise "the class problems of our supposedly classless so-
 The Perversion of Autonomy dr the Subjection of Women 195

ciety . . . by filtering them through certain kinds of discussion about race and the
shiftless, undeserving, unemployable black 'underclass.'"56 In consequence, single
mothers (who are not, she notes, predominantly black or "teenaged") are left res-
olutely to their own devices, as the advocacy that might elicit more responsible re-
sponses from an affluent society is condemned as special pleading for members of
"interest groups" who have not made the right choices, not exercised self-restraint
and self-reliance, and not contributed "appropriately" to the society that demonizes
them. Even arguments that ought to awaken the self-interest of the affluent in re-
sponse to evidence that single mothers "now bear a greater responsibility than at any
time in our history for raising the children of this society," and hence that everyone
has a stake in the well-being of these children, are reconstructed as truth-ignoring
advocacy for the "shiftless welfare queen who always gets more than she de-
serves."57 The presumed autonomy of the affluent sex, class, and race that allows
them to "know" her thus produces and condemns a new heteronomy of others who
fail by its unexamined coercive standards.
Cornel West offers an analogous example:
The new black conservatives claim that transfer payments to the black needy en-
gender a mentality of dependence which undercuts the value of self-reliance and
of the solidity of the black poor family. . . . [Yet] in the face of high black unem-
ployment, these cutbacks will not promote self-reliance or strong black families
but will only produce even more black cultural disorientation and more devas-
tated black households.58

An ever vigilant individualistic and coercive imperative takes on a censorious de-

meanor that suppresses the possibility that such responsibilities are more collective
than individual. A set of intransigent assumptions keeps in circulation the prescrip-
tive consequences of hyperbolized autonomy, bypassing empirical evidence that
demonstrates how the effects of "the unnatural lottery" count among the conditions
that make knowledge possible or block its passage altogether.
Such rhetorical shifts reinforce the coercive power of an old-style autonomy
that discredits the very advocacy practices that are integral to the idea of a society as
a locus of mutual responsibility. Yet social welfare, in the United States and Canada
as elsewhere, came into being because of advocacy (often by women) to privileged
people who were themselves able to expect social benefits, as a matter of course, that
in less affluent positions earned the label "welfare."59 The rhetoric of self-reliance,
sanctimoniously manipulated, enforces a stoic version of autonomy, separately and
singly, on people thwarted in their efforts to achieve it by the very free-market soci-
ety whose "just" distributions of power and privilege produce their deplored "in-
ability." As Wendell notes, "When governments want to make cutbacks in social
services, they may 'release' people from institutions without adequate support to en-
able them to live on their own . . . describing the changes they make as creating
more 'independence.' "60 Yet these perversions of autonomy enforce renewed pat-
terns of subjection after all.
The rhetoric that encourages dismantling the welfare state reiterates the rhet-
orics of opposition to its creation. Martha Minow records that "to many observers,
state-supported social welfare jeopardizes individual autonomy, independence and
196 Autonomy and the Social
61
freedom." Opposition to the infant, and maternal health legislation (1927) de-
ployed "communist" and "socialist" as terms of invective, inveighing against it be-
cause it "interfered . . , with the free-market economy" and voicing fears that
"women would reject their traditional roles, given the acts encouragement of women to
seek information about their own bodies and health"^2 Sapere elude indeed. The voice
of autonomy hyperbolized is as audible here as in the 1994 California repeal of wel-
fare for single mothers, with arguments that welfare thwarts their autonomous self-
reliance, and in Ontario's 1995- 1996 conservative attacks on social welfare that "in-
terferes" with personal autonomy. Welfare as a vehicle for a "free ride" is the guiding
metaphor, an image that masks the extent to which welfare humiliates and coerces
even as it "helps." Minow notes that "public assistance and social work interventions
too often . . . controlled and restricted . . , people [in need] and subjected them to
governmental regulation without protection against power imbalances."63 Linda
Gordon observes that welfare "humiliates its recipients by subjecting them to de-
meaning supervision and invasions of privacy."''4 What the discourse of autonomy
initiates as an entitlement mutates into a disempowering gift that imposes obliga-
tions its recipient cannot honor, for the very reasons that led her to seek assistance
in the first place.
Welfare issues, then, are about relations to and within an intransigent system in-
formed by stereotyped "knowledge" of subjectivity and agency that continues to
hold autonomous man in place as a regulative fiction while subjecting women and
others who cannot measure up to the dictates of an impossible imperative. This pu-
tative knowledge legitimates policies, actions, and treatments modeled on his expe-
riences and options and glosses over the extent to which autonomous man himself is
dependent on patterns of invisible advocacy for "his own" putatively autonomous
achievements.
In both of these relational complexes, it may appear that the point of advocacy
is to secure the very autonomy that is lacking. Indeed, the empowering practices en-
joined and engaged often name autonomy realization among their goals. But this is
autonomy entirely remodeled: disindividuated, deccntered, and reconfigured along
patterns of mutual and crisscrossing recognitions and responsibilities. It does not
define itself in opposition to a caricatured paternalism, as in Graylin and Jennings's
account; it is multiply, relationally structured all the way down, eschewing the indi-
vidualism around which twentieth-century hyperbolized autonomy defines itself.
Jane Flax rightly observes that "the success of feminist projects does not depend on
the acquisition of an autonomous agency some white men can pretend to exercise"
noting that "the appearance of stability, autonomy, and identity is generated and sus-
tained by relations of domination, denial of aspects of subjectivity, and projection
of contradictory material onto marked others."65
Feminists are well aware that within the insider-outsider structures that frame
the politics of public knowledge and the prestige of scientific knowledge, "ordi-
nary" women's voices—like those of other disenfranchised knowers—often go un-
heard and rarely achieve autonomous acknowledgment. Their reports of violence,
sexual assault, racism, and sexism in the workplace are as often discredited as acted
on. Such patterns of incredulity prompt the uneasy suggestion I have elaborated in
this section, that first-person experiential reports often require "expert" advocacy to
 The Perversion of Autonomy & the Subjection of Women 197

claim a serious hearing. The suggestion is "uneasy" because of the delicacy of issues
about who can or should speak for whom, in a long history of propertied white
men speaking for, thinking for, voting for, and making decisions for "their" women
and their alleged inferiors, while claiming to know women and other Others bet-
ter than they can know themselves. Feminists have to be as wary of women speak-
ing for one another as of men's presumptions to speak on women's behalf. And ad-
vocacy seems to erase women's hard-won capacity to speak in their own voices; it
threatens a renewed silencing that would replicate oppressive patterns of hierarchi-
cal expertise.
Thus, as I note in the previous section, the politics of advocacy are among the
most contested issues in present-day activism and research. The task, then, is to de-
vise ways of distinguishing responsible advocacy from appropriation, acknowledg-
ing that we cannot always speak for ourselves, yet that people who speak for us, on
our behalf or about us, are as often underinformed, self-interested, and imperialis-
tic as they are supportive and empowering. Although I cannot spell out a finite set of
rules for responsible advocacy, clear violations of the trust, sensitivity, and integrity
on which it relies become visible in feminist-informed deliberation, indicating, by
contrast, how responsibilities might differently be assumed. Thus, for example, the
failures of "one policy, one diagnosis fits all" that are vividly apparent in my welfare
and medical examples become catalysts for transformative advocacy. In guiding that
effort, I propose that ecological thinking offers a more responsible approach than
those mandated by the discourses of disinterested autonomy.

Ecology
The decentering of the subject, gained by the labors of structuralism and
poststructuralism, levelled hierarchies and shifted a vertical vision of the
world toward a more horizontal one, so important for feminism, that
places on the same surface both multiculturalism and ecology. (Verena
Conley)66

When I do not see plurality stressed in the very structure of a theory, I

know that I will have to do lots of acrobatics—like a contortionist or
tight-rope walker—to have this theory speak to me without allowing the
theory to distort me in my complexity.
(Maria Lugones)67

Revisionary projects of thinking through the renewed subjections consequent on

"perversion [s] of autonomy" can only be minimally transformative while they re-
main framed within the confines of individualistic conceptions of knowledge, sub-
jectivity, and agency, where, as Val Plumwood persuasively observes, the subject
"loses a sense of itself not only as an organic but as a social being, as an agent in and
chooser of political, economic and technological frameworks."68 Softening the co-
ercive force of the social imperatives that individualistic autonomy generates by
elaborating individualistic selfhood into intersubjective multiples of the Same, be-
gins to contest the hegemony of the framework. Yet the discourses of mastery I refer
198 Autonomy and the Social

to earlier, which are spawned and thrive at this confluence of individualism and hy-
perbolic autonomy, permeate the instituted social imaginary of the affluent West-
ern world so thoroughly that internal tinkering with their detail cannot achieve the
transformative effects that revisionary politics require.
Ethical self-mastery, political mastery over unruly and aberrant Others, and
epistemic mastery over the "external" world pose as the still-attainable goals of the
Enlightenment legacy that maintains the tensions within the discourses of auton-
omy and social advocacy I have discussed. They sustain hierarchical, generally ex-
ploitative relations between members of privileged cultures and of nonaffluent,
non-Western societies and cultures at home and abroad. These relations are config-
ured into top-down (vertical) patterns of authority and expertise that legitimate
domination in the name of an achieved mastery of the human and natural world
and mastery over the personal idiosyncrasies of the masters.''9 They endorse exag-
gerated proclamations of the "permanent value" (recall Cixous) of the abstract au-
tonomous agency integral to the philosophy of rights, which reduces "plurality" to
variations on the Same; substitutes instrumental policy decisions for open practices
of deliberative democracy; and distorts the complexities of specific positionings
and allegiances in the name of universal, impartial rights, obligations, and duties.
These discourses underwrite an equivalently reductive picture of epistemic agents
as isolated units on an indifferent landscape, to which their relation is one of dis-
engaged indifference.
In contradistinction to discourses of mastery and domination, I propose to en-
list the ethicopolitical potential of ecological thinking to reconfigure knowledge, so-
ciality, and subjectivity. 70 This proposal does not recommend interjecting an alter-
native social imaginary into the old-but-not-yet-tired hegemonic imaginary, thus
merely offering a "choice." Ecological thinking reconceives human locations and re-
lations all the way down. It infiltrates the interstices of the social order, where it ex-
pands to undermine its intransigent structures, as the ice of the Canadian winter ex-
pands to produce upheavals in the pavements and roads, working within and against
these seemingly solid structures to disrupt their smooth surfaces.
The "decentering" that ecological thinking entails is enacted in its refusal to
continue philosophy's tacit derivation from the white, affluent, Western, male ex-
periences that also generate hyperbolic autonomy ideals. It displaces "man" from
his central position in the world and "in" himself and disturbs the (often narcis-
sistic) inwardness of autonomy. Ecological thinking's revolutionary potential
recalls Kant's Copernican revolution, which moved "man" to the center of the
philosophical-conceptual universe and whose effects were constitutive of human-
istic, post-Enlightenment thought. Its picture of the physical and human world de-
rived from and served the interests of a small class and race of people powerful
enough to claim dominion over the rational and moral territory. Yet Kantian phi-
losophy was parochial in the generic conception of "man" on which it turned—a
recognition central to feminist, Marxist, postcolonial, and critical race theory. As
humanism vied with theism in the seventeenth and eighteenth centuries, ecological
thinking now vies with capitalism and its attendant discourses of autonomous mas-
tery. Yet ecological thinking is not unilinear, for it emerges from and addresses so
many interwoven, crisscrossing, often contradictory issues—feminist, classist, post-
 The Perversion of Autonomy dr the Subjection of Women 199

colonial, antiracist, environmental—that it requires multifaceted, multitextured chart-

ings. It redirects the focus of theoretical analyses toward situated knowledges and
situated ethicopolitics, in which situation is constitutive of, not just the "context"
for, the backdrop against which enactments of subjectivity occur. (Donna Haraway
comments: "I have a body and mind as much constructed by the post-Second
World War arms race and cold war as by the women's movements."71)
Ecological thinking—a peculiarly late twentieth-century phenomenon—
examines the potential of epistemic and ethicopolitical practices to produce habi-
tats where people can live well together, locally and globally, and respectfully within
the physical and natural world.72 This refocusing links ecological thinking's decen-
tered aspects to its horizontal patternings. Both require conceptual and geographi-
cal remappings of terrains where practices are engaged and theoretical conclusions
constructed—mappings sensitive to detail and diversity and resistant to hasty trans-
lation from situation to situation.
In these remappings, deductive (vertical) reasoning from overarching premise,
principle, or governing law to conclusions systematically deduced and applied no
longer claims solitary dominion over the epistemic terrain. Often it yields to and
modifies its conclusions in concert with reasoning that traces evidential analogies
and disanalogies horizontally, from situation to situation—"transversally," as Verena
Conley puts it, "making a link between ecosystems that international capitalism
seeks to polarize and ecological issues that pertain to local cultures."73 Aware of the
force of disanalogy, analogical reasoning is skeptical of universalizing conclusions
and solutions: such skepticism is activated by syndrome X in my medical example.
At its best, analogical thinking is interpretive, deliberative, and collaboratively ne-
gotiated; it is usually more inductive than deductive, thus more modestly conjec-
tural than arrogantly conclusive. Whereas it is often discredited for being subjective,
"particularist," and situational rather than objective and universal, it can just as often
be more fully empirical and responsible to its subject matter than conventionally re-
spectable deductive-nomological reasoning.74 Nor, in my rendition, is it only hori-
zontal, for it also derives its understandings from genealogical analyses of the power
that permeates and legitimates the instituted social imaginary.
The discourses of mastery enlist ready-made, easily applied categories to sum
up and contain parts of the personal, social, and physical-natural world within a
neatly manageable array of "kinds." Ecological thinking is more finely attuned to
differences, as it resists the desire for unity that drives these practices—the desire to
assemble the confusion of the human and natural world into maximally homoge-
neous patterns. As the examples I have discussed show, such control legitimates ex-
trapolation from the situations, symptoms, "needs," and goals of the powerful into
diagnoses and policy decisions enacted on people who "must," if they are sufficiently
enlightened, want to be just like them. By contrast, ecological thinking, in its com-
mitment to complexity, urges sensitivity to detail, to minutiae, to what precisely—
however apparently small—distinguishes this patient, this welfare recipient, from
that; this practice, this locality, from that; as biologically based ecologists would dis-
tinguish this plant, this species, from that. It investigates the meaning of these dif-
ferences, and studies how they promote or thwart possibilities of living well.
Ecological thinking thus resists the epistemic practice common in science-
200 Autonomy and the Social

venerating cultures of superimposing a grid on events and situations it needs to

know, tucking in the bits that spill over the edges, and letting aberrations fall
through the cracks. It is not an easy project, for it combines the careful attention
to evidence that it finds in traditional versions of empiricism with investigations
that locate symptoms, social issues, problems within wider patterns of power and
privilege, oppression, and victimization. Thus, for example, in an analysis that dis-
plays many of the features 1 am describing as ecological, Lucy Candib argues
against readings of medical symptoms that gloss over narrative issues derived from
what I am calling the patient's "habitat" but are integral to responsible diagnosis.
Connecting bclievability to advocacy, she writes: "Patients subject to abuses of
power bring their stories of vulnerability to clinicians every day. . . . Advocacy
means arguing for the believability of those stories."75 She notes that women can
experience chest pain, like many other symptoms, as a result of abusive practices
that permeate their lives. Mapping the effects of these oppressions for this pa-
tient— abdicating the "disinterested posture" expected of family physicians, a
posture that "corrupts] the idea of advocacy"76— a physician is able to cooperate
with the patient in making sense of her symptoms, refusing to hold her au-
tonomously responsible for their occurrence or to treat problems as merely me-
chanical failures, and doctors and patients are able to think differently about
causes and cures. This is just one example of how ecological thinking can work.
In its respect for particularity, ecological thinking is committed to a method-
ological pluralism. Thus advocacy becomes neither a new paradigm nor a universally
valid method simply to be "applied" across the places where ecological thinking can
make a difference. Nor are all advocacy relations alike or all advocacy practices ho-
mogeneous. They produce multiple studies of (often linked) oppressions and gen-
erate multiple coalitions and forms of activism, where the trick is to understand the
connections that grant the myriad oppressions of white, patriarchal, capitalist soci-
eties their mutually reinforcing power and to develop liberatory strategies.77
Bcing-in-the-world guided by ecological thinking differs radically from the
masterful way of autonomous man, whose assumption that he can be master of all
he surveys allows surveying to substitute for engaged participation arid mastery to
suppress diversity for the sake of instrumental simplicity. For Verena Conley (fol-
lowing Michel de Certeau), ecological thinking is "a, way of inhabiting the world" "a
social and natural concern aimed at measuring habitability."78 These structural fea-
tures inform ecological thinking through and through.
The sources in feminist and postcolonial theory of these features of ecological
thinking will be apparent. Ecofeminism is a principal source, as are deep ecology
and the "green" movement. All are contested and contestable positions, neither po-
litically innocent nor able single-handedly to remake the world. All contain the seeds
of "perversions" as inimical to "habitability" as the philosophy of rights, as capable
of sowing renewed "subjections" as autonomy in its hyperbolized modes. Yet their
intellectual-activist promise justifies the risks involved in drawing on them.
One such risk is in an alignment—indeed, an identification—in some ecofem-
inist writings of an essential "woman" with an equally essential "Nature": a practice
which threatens to reconfirm the biological determinism that has kept women "in
their place."79 Thus, too, some ecofeminists affirm the naturalness of women's tra-
 The Perversion of Autonomy dr the Subjection of Women 201

ditional care giving, extended now to include nature, while attending too little to
how oppressive social structures exploit women's "natural" caring capacities and per-
petuate their subjection. Equally conducive to wariness, for me, are the rehearsals of
hypermasculinity enacted in romantic quests for a "return to nature," often as a place
for male self-discovery; the spiritualism of some facets of environmentalism; and a
tendency Plumwood finds in deep ecology "to focus exclusively on identification, in-
terconnectedness, sameness and the overcoming of separation, treating nature as a
dimension of self."80
Nonetheless, my position takes very seriously, and derives its guiding metaphor-
ical apparatus from, the creative projects of ecological thinkers who are rethinking
and reimagining social, epistemological, political, and economic philosophies of
mastery by opening deliberative, discursive spaces in which to enter, critically and
constructively, into a program such as Donna Haraway names in her injunction:
"We must find another relationship to nature besides reification, possession, appro-
priation, and nostalgia."81 As Haraway shows throughout her writings, reworking
that relationship is vital to human survival at century's end.
The "situated knowledges" of Haraway's "modest witness" count as provocative
models of ecological thinking. This modest witness
insists on its situatedness, where location is itself always a complex construction
as well as inheritance, and . . . casts its lot with the projects and needs of those
who could not or would not inhabit the subject positions of the self-invisible and
the discursive sites, the "laboratories," of the credible, civil man of science.82

Sandra Harding's "strong objectivity" figures prominently in the picture in its insis-
tence on counting the circumstances of inquiring subjects as integral to the evidence
from which knowledge is derived, as does Linda Alcoff's "real knowing" as enacted
in the interpretive dimensions of coherence theory, and my "epistemic responsibil-
ity" as it dismantles boundaries between ethics and politics and between ethicopoli-
tics and epistemology, reconfiguring patterns of accountability.83
Ecological thinking finds its point of departure in the natural dependence of
knowledge production on interactive practices, such as the advocacy practices I have
discussed, throughout human epistemic lives. Drawing critically on empirical evi-
dence to determine how survival could be ensured and enhanced, not just quan-
titatively but qualitatively, it is wary of the power-infused racial, gender, and class
stereotypes and essentialized conceptions of "science" and "nature" that become both
self-fulfilling and self-perpetuating. Ecology (literally) looks to state-of-the-art nat-
ural science to supply some of the substance of its deliberations, yet it does not as-
sume that science has a direct line to "the truth" nor merits uncontested licence to
intervene where it pleases.84 Ecology (metaphorically) draws the conclusions of sit-
uated inquiries together and maps their interrelations, and their consonances and
contrasts, and their impoverishing and mutually sustaining effects.85 Yet ecological
thinking is not innocently benign, promising an unimpeded unfolding of epistemic
and moral potential. Ecosystems are as competitive and unsentimentally destructive
of their less viable members as they are cooperative and mutually sustaining; nor
could ecological thinking in its metaphorical frame pose as a univocal good, free
from critical and self-critical responsibilities. So if work within it is to avoid repli-
202 Autonomy and the Social

eating the exclusions endemic to traditional epistemologies, it will often require

advocacy engaged out of the vigilant monitoring on which transformative social
movements always depend to preserve their fragile gains and counter the threat of
renewed oppressions.
In the epistemically responsible and sensitive negotiations on which good advo-
cacy depends, hyperbolized autonomy will have to yield space to a fuller recognition
of integrity, a principle less individualistically based and more at home on this hori-
zontally spread-out landscape. Claudia Card notes its connotations of "consistency,
coherence, and commitment"; Margaret Walker connects it to "the sturdiness of
structures people [especially socially subordinated people] have built, the property of
holding up dependably under the weights and stresses these structures are apt to en-
counter,"86 showing how remarkably even people wholly deprived of autonomy can
yet preserve integrity. While refusing to romanticize nature, ecological thinking also
assumes an integrity in natural processes and biodiversity that is worthy of respect
if not of such fanatical preservation as principled refusals to swat a mosquito, in-
noculate against diphtheria, or restrain a dangerous offender might display.
In an analysis that links ecology and autonomy within an epistemological-
ethical frame and is consonant with my discussion here, Cornelius Castoriadis notes
how representations of scientific knowledge as disinterestedly neutral shape "the
dominant social imaginary (imaginaire social) of our epoch," for which "the central
aim of social life is the unlimited expansion of rational mastery."87 This imaginary
claims the assent of the social majority, as citizens are nurtured from infancy to in-
gest: an unquestioned relationship to an "ensemble of needs" whose satisfaction be-
comes a life-long project. The discourse of needs "naturalizes" its referents and ob-
scures their artifactuality.
For Castoriadis, genealogies of the creation of the "individual," complete with
his needs, within this imaginary expose the contingency of the process and the in-
terests it tacitly serves, thus making it possible to turn it against itself by utilizing
that same scientific, technological, and social-psychological knowledge to recon-
struct human relations to the social and natural world.88 The ecological movement
becomes a resource for addressing the "political problems of the reconstruction of
autonomous society," which has posed as many problems for the women's move-
ment and the youth movement as for the worker's movement. It offers a creatively
inspired emancipatory conceptual apparatus within which to refuse ongoing colo-
nization by an imaginary that requires relations of dominance and submission to
sustain it, relies on concealed linkages between power and knowledge, and promotes
the individual autonomy of a few through the heteroriomy of the many.
Here I have discussed advocacy relations as micro practices that work within
and for a commitment to ecological thinking by setting creative ethicopolitical
changes in motion to disrupt the social imaginary governed by the rhetoric of au-
tonomous individualism. They initiate disruption not for its own sake but for the
sake of creating more habitable habitats in which to recognize, acknowledge, and
thus make space (rhetorical, physical, and agenda space) for such subjugated knowl-
edges as women's inexplicable medical disorders and for the kind of social being that
affirms an ideal of citizenship in which the distortions that fuel the perversions of
autonomy that animate renewed subjections of women and other socially margin-
 The Perversion of Autonomy & the Subjection of Women 203

alized Others can no longer claim a place. Such work opens spaces for a new, cre-
atively interrogating, instituting social imaginary to undermine and displace the in-
stituted imaginary that claims to represent the natural way of being and knowing.89
Whereas a relational autonomy drawn from the polite relations of ordinary lib-
eral societies starts well above the bottom line of struggles for autonomous control
even over one's body that sickness, poverty, malnutrition, and oppression erase or
render impossible, in ecological thinking places of excess in the negative effects of
the "unnatural lottery" become starting places for investigation, not for blame or for
subjection within an illusory autonomous sameness. With its conception of mate-
rially situated subjectivity for which embodied locatedness and deliberative interde-
pendence are constitutive of the very possibility of knowledge and action, ecological
thinking opens the way to a renewed conception of responsible citizenship, as re-
sponsible in its knowing as in its doing.

Notes

1. I presented earlier versions of this chapter to the departments of philosophy at La

Trobe University and SUNY Buffalo, at the Centre for Gender Studies at the University of
Adelaide, and at the conference on Feminist Perspectives on Autonomy and Agency at the
Australian National University. I have benefitted from the discussions on those occasions and
from comments by Murray Code, Helen Connole, Pablo DeGrieff, Catriona Mackenzie, and
Richard Schmitt.
2. Helene Cixous, "We Who Are Free, Are We Free?" trans. Chris Miller, in Freedom
and Interpretation: The Oxford Amnesty Lectures 1992, ed. Barbara Johnson (New York: Harper-
Collins, 1993), p. 18 (emphasis in the original).
3. See, for example, Lorraine Code, What Can She Know?: Feminist Theory and the Con-
struction of Knowledge (Ithaca, N.Y.: Cornell University Press, 1991), chaps. 3 and 4, and
Jean Grimshaw, Philosophy and Feminist Thinking (Minneapolis: University of Minnesota
Press, 1986), chap. 5.
4. For an analysis of care's associations with feminine values that stay too close to home,
see Sarah Lucia Hoagland, "SomeThoughts about 'Caring,'" and for a discussion of its op-
pressively feminine associations, see Michele Moody-Adams, "Gender and the Complexity of
Moral Voices," both in Feminist Ethics, ed. Claudia Card (Lawrence: University Press of
Kansas, 1991). JoanTronto analyzes care's "underclass" associations in "Beyond Gender Dif-
ference to a Theory of Care," Signs: Journal of Women in Culture and Society 12 (Summer
1987): 644-663.
5. Axel Honneth offers a succinct analysis in "Decentered Autonomy: The Subject after
the Fall," trans. John Farrell, in Axel Honneth, The Fragmented World of the Social: Essays in
Social and Political Philosophy, ed. Charles W. Wright (Albany: SUNY Press, 1995). David R.
Hiley discusses analogous tensions between autonomy and community in Philosophy in Ques-
tion: Essays on a Pyrrhonian Theme (Chicago: University of Chicago Press, 1988), especially
chap. 2, "Progress against Ignorance." He writes: 'We [in the twentieth century] have retained
eighteenth-century conceptions of freedom which have as their central feature the idea of in-
dividual self-determination, yet we seem no longer able . . . to support both our autonomy
and genuine community" p. 63.
6. For a sampling of recent texts, see M. Jacqui Alexander and Chandra Talpade Mo-
hanty, eds., Feminist Genealogies, Colonial Legacies, Democratic Futures (New York: Routledge,
204 Autonomy and the Social

1997); Diana T. Meyers, ed., Feminists Rethink the Self (Boulder, Colo.: Westview Press,
1997); Lewis R. Gordon, Her Majesty's Other Children: Sketches of Racism from a Neocolo-
nialAge (Lanham, Md.: Rowman & Littleficld, 1997); Erica Burman, Deconstructing Devel-
opmental Psychology (London: Routledge, 1994); Debbora Battaglia, ed., Rhetorics of' Self-
MafeVzg-(Berkeley: University of California Press, 1995); Marilyn Waring, Three Masquerades:
Essays on Equality, Work and Human Rights (Toronto: University of Toronto Press, 1996);
Kirs ten Hastrup and Peter Hervik, eds., Social Experience and Anthropological Knowledge
(London: Routledge, 1994); and Jennifer Radden, Divided Minds and Successive Selves: Eth-
ical Issues in Disorders of Identity and Personality (Cambridge, Mass.: MIT Press, 1996).
7. These phrases are Honneth's, in "Dccentercd Autonomy," p. 261, to whom I am in-
debted in framing these issues.
8. See especially Michael Sandel's Introduction in Liberalism and Its Critics, cd. Michael
Sandel (Oxford: Blackwell, 1984), p. 5; and for a critical analysis, see Honneth, "The Lim-
its of Liberalism: On the Political-Ethical Discussion Concerning Communitarianism," in
Honneth, Fragmented World of the Social.
9. See especially Nancy Chodorow, "Toward a Relational Individualism: The Media-
tion of Self through Psychoanalysis," in Reconstructing Individualism: Autonomy, Individual-
ity, and the Self in Western Thought, ed. T. C. Heller, M. Sosna, and D. E. Wellberry (Stan-
ford, Cal.: Stanford University Press, 1986).
10. Note that individualMft'c subjectivity and agency (and thus also individual MOT) con-
trast with North American vernacular talk of "individuals" to mean "people." They contrast
also with individualz/y, which is not under threat in these contestations of its boundary
claims.
11. Richard Sdhmitt discusses how "The Autonomy of the Philosophers" culminates in
the importance of having a "rational life plan," in Beyond Separateness: The Social Nature of
Human Beings—Their Autonomy, Knowledge, and Power (Boulder, Colo.: Westview Press,
1995), pp. 4 — 6. I discuss the class, race, and gender specificity of such "plans" in "Rational
Imaginings, Responsible Knowings: How Far Can You See From Here?" in EnGenderingRa-
tionalities, ed. Nancy Tuana and Sandi Morgen (forthcoming). See also Margaret Walker,
"Career Selves," in Moral Understandings: A Feminist Study in Fjhics (New York: Routledge,
1998).
12. Immanuel Kant, "An Answer to the Question: 'What is Enlightenment?'" (1784),
trans. H. B. Nisbct, in Kant's Political Writings, ed. Hans Reiss (Cambridge: Cambridge Uni-
versity Press, 1970), pp. 54-55 (emphasis in the original).
13. Michel Foucault, "What Is Enlightenment?" trans. Catherine Porter, in The Fou-
cault Reader, ed. Paul Rabinow (New York: Pantheon Books, 1984), p. 35. I discuss Fou-
cault's reading of Kant's essay at greater length in "Critiques oi Pure Reason," in Lorraine
Code, Rhetorical Spaces: Essays on (Gendered) Locations (New York: Routledge, 1995).
14. Kant, "Answer to the Question," p. 54 (my emphasis).
15. Ibid., pp. 55 (emphasis in the original).
16. Claudia Card, The Unnatural Lottery: Character and Moral Luck (Philadelphia:
Temple University Press, 1996), pp. 3, 4. The title plays on John Rawls's claim that the dis-
tribution of natural human assets is decided by a "natural lottery," which is "arbitrary from a
moral perspective." See Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University
Press, 1971), p. 74. The members of a society who are "satisfying the principles of justice as
fairness... are autonomous and the obligations they recognize self-imposed." Ibid., p. 13 (my
emphasis).
17. Card, Unnatural Lottery, p. 20 (emphasis in the original).
18. Walker, Career Selves, p. 151.
 The Perversion of Autonomy & the Subjection of Women 205

19. Sabina Lovibond, Realism and Imagination in Ethics (Minneapolis: University of

Minnesota Press, 1986), p. 169.
20. Willard Gaylin and Bruce Jennings, The Perversion of Autonomy: The Proper Uses of
Coercion and Constraints in a Liberal Society (New York: Free Press, 1996), pp. 5, 24.
21. John Stuart Mill, The Subjection of Women (1869), in John Stuart Mill and Harriet
Taylor Mill, Essays on Sex Equality, ed. Alice Rossi (Chicago: University of Chicago Press,
1970).
22. I use the term to capture the social-material-historical dimensions of situatedness.
See also note 71.
23. Mary Tiles and Jim Tiles note that Bacon's four "Idols"—the idol of the tribe, the
theater, the marketplace, and the cave—classify "the various false ideas which obscure and
interfere with our grasping the true objects of knowledge," in Tiles and Tiles, An Introduc-
tion to Historical Epistemology: The Authority of Knowledge (Oxford: Blackwell, 1993), p. 37.
24. The reference is to Donna Haraway, "Situated Knowledges: The Science Question
in Feminism and the Privilege of Partial Perspective," in Simians, Cyborgs, and Women: The
Reinvention of Nature (New York: Routledge, 1991). Haraway denies that knowers can per-
form "the god trick" of extracting themselves from the world to perceive and know as if
from nowhere. She refines and elaborates the conception of situated knowledges that es-
chew illusions of epistemic autonomy mModest_Witness@Second_Millennium.FemaleMan'e>_
Meets_OncoMouse™: Feminism and Technoscience (New York: Routledge, 1997).
25. For such analyses, see Elizabeth Fricker, "The Epistemology of Testimony," Aris-
totelian Society Supplementary Volume 61 (1987): 57-83, and C. A. J. Coady, "Testimony
and Observation," American Philosophical Quarterly 10 (April 1973): 149-155.
26. Thomas Nagel, "Moral Luck," in Mortal Questions (Cambridge: Cambridge Uni-
versity Press, 1979), p. 36.
27. Walker, Career Selves, p. 13 (emphasis in the original).
28. Anne Seller, "Realism versus Relativism: Towards a Politically Adequate Epistemol-
ogy," in Feminist Perspectives in Philosophy, ed. Morwenna Griffiths and Margaret Whitford
(Bloomington: Indiana University Press, 1988), p. 180; Ludwig Wittgenstein, On Certainty,
ed. G. E. M. Anscombe and G. H. von Wright, trans. Denis Paul and G. E. M. Anscombe
(New York: Harper Torchbooks, 1971), § 378. See also Code, What Can She Know?chap. 5,
and "Incredulity, Experientialism and the Politics of Knowledge," in Rhetorical Spaces.
29. Lynette Hunter, Critiques of Knowing: Situated Textualities in Science, Computing
and the Arts (London: Routledge, 1999), pp. 30-31.
30. For instructive discussions of these issues, see Hilde Lindemann Nelson, ed., Sto-
ries and Their Limits: Narrative Approaches to Bioethics (New York: Routledge, 1997).
31. Peter Novick, That Noble Dream: The '"Objectivity Question" and the American His-
torical Profession (Cambridge: Cambridge University Press, 1988), p. 2.
32. "Subjugated knowledges" is Michel Foucaults term in Power/Knowledge: Selected In-
terviews and Other Writings, 1972-1977, ed. Colin Gordon, trans. Colin Gordon, Leo Mar-
shall, John Mepham, and Kate Soper (New York: Pantheon, 1980). I appeal here to Foucaults
second sense of the term, in which it refers to "a whole set of knowledges that have been dis-
qualified as inadequate to their task or insufficiently elaborated: naive knowledges located low
down on the hierarchy, beneath the required level of cognition or scientificity." For Foucault,
"It is through these . . . disqualified knowledges, that criticism performs its work." Ibid., p. 82.
33. Haraway, Modest_Witness, p. 68.
34. Sandra Harding, "Rethinking Standpoint Epistemology: What Is 'Strong Objectiv-
ity'?" in Feminist Epistemologies, ed. Linda Alcoff and Elizabeth Potter (New York: Routledge,
1993), p. 71. See also Nancy Hartsock, Money, Sex and Power: Toward a Feminist Historical
206 Autonomy and the Social

Materialism (Boston: Northeastern University Press, 1985); Helen Longino, Science as Social
Knowledge (Princeton, N.J.: Princeton University Press, 1990); and Lorraine Code, "Taking
Subjectivity into Account," in Feminist Epistemologies.
35. In Lorraine Code, "What Is Natural about Epistemology Naturalized?" American
Philosophical Quarterly^, no. 1 (January 1996): 1-22, I develop a critical analysis of these
connections.
36. Prominent among these critics is Jaegwon Kim, "What Is 'Naturalized Epistemol-
ogy'?" in Naturalizing Epistemology, ed. Hilary Kornblith (Cambridge, Mass.: MIT Press,
1994).
37. See, for example, Alice Herb, "The Hospital-based Attorney as Patient Advocate,"
Hastings Center Report I1), no. 2 (1995); American Medical Association Council, "Ethical Is-
sues in Managed Care," Journal of the American Medical Association 273, no. 4 (1995); Leo
T Rosenberg, "Delaying Approval of a Critical Drug," Journal of Medical. Humanities 15,
no. 4 (1994); William May, "On Ethics and Advocacy," Journal of the American Medical As-
sociation 256, no. 13 (1986); Fredrick R. Abrams, "Patient Advocate or Secret Agent?" jour-
nal of the American Medical Association 256 no. 13 (1986); Nancy S. Jecker, "Integrating
Medical Ethics with Normative Theory: Patient Advocacy and Social Responsibility," Theo-
retical Medicine 11, no. 2 (1990).
38. The quoted phrases come from E. Charlotte Theis, "Ethical issues: A Nursing Per-
spective," New England Journal of'Medicine'315, no. 19 (1986). See also Ellen M. Bernal,
"The Nurse as Patient Advocate," Hastings Center Report 22, no. 4 (1992).
39. Peta Bowden, Caring: Gender Sensitive Ethics (London; Routlcdge, 1997), p. 124.
40. Susan Wendell, The Rejected Body: Feminist Philosophical Reflections on Disability
(New York: Routledge, 1996), p. 122.
41. Elaine Scarry, The Body in Pain: The Making and Unmaking of the World (Oxford:
Oxford University Press, 1985), p. 4.
42. Ibid., p. 6 (my emphasis).
43. I discuss the politics of knowledge of syndrome X in "How Do We Know? Ques-
tions of Method in Feminist Practice" in Changing Methods: Feminists Transforming Practice,
ed. Sandra Burt and Lorraine Code (Peterborough, Ontario: Broadview Press, 1995).
44. Of syndrome X, Kathleen King notes: "Men often experience . . . 'textbook' cases
of angina and other heart-disease symptoms because the textbooks are written to describe
men's symptoms. . . . You're not going to think of heart disease unless the symptoms fit the
classic picture. And . . . we don't know what the classic picture for women is." Quoted in
Robin Marantz Henig, "Kind Hearts and Coronaries," The Globe and Mail (Toronto), 20
November 1993, D 8.
45. Sec Kirsti Maltcrud, "Women's Undefined Disorders — A Callengc for Clinical
Communication," Family Practice*) (1992): 299-303; "Strategies for Empowering Women's
Voices in the Medical Culture," Health Care for Women International 14 (1993): 365-373;
"The (Gendered) Construction of Medical Diagnosis. Interpretation of Symptoms and
Signs in Female Patients," Theoretical Medicine andBioethics, 1999 (in press).
46. I am referring to Thomas Kuhn's claims about the intransigence of paradigms. See
Kulm, The Structure of Scientific Revolutions, 2nd ed. (Chicago: University of Chicago Press,
1970).
47. Foucault, Power/Knowledge, p. 81 (emphasis in the original).
48. Henig, "Kind Hearts and Coronaries."
49. Harvard Women's Health Watch 1, no. 6 (February 1994).
50. Christina Hoff Sommers, Who Stole Feminism? (New York: Simon & Schuster,
1994), p. 103.
 The Perversion of Autonomy dr the Subjection of Women 207

51. I am indebted in thinking about this example to Victoria Davion in "Listening to

Women's Voices: Rape, Epistemic Privilege, and Objectivity," in Daring to Be Good: Essays in
Feminist Ethico-Politics, ed. Bat-Ami Bar On and Ann Ferguson (New York: Routledge,
1998). The report under discussion is Mary Koss, "Hidden Rape: Sexual Aggression and Vic-
timization in a National Sample in Higher Education," in Rape and Sexual Assault, Vol. 2, ed.
Ann Wolpert Burgess (New York: Garland Press, 1988).
52. Donna Haraway's account of how Danish panels of citizen deliberation contribute
to shaping democratic science and technology policies offers an impressive example of how
such a process could work. It encourages a remarkable "degree of scientific and technical lit-
eracy . . . in ordinary people." Modest_Witness, pp. 95-96.
53. They call it "autonomy gone bonkers." Gaylin and Jennings, Perversion of Auton-
omy, chap. 10.
54. Ibid., chap. 9, "In Defense of Social Control: The Ethics of Coercion."
55. Ibid., p. 182.
56. Patricia J. Williams, "The Unbearable Autonomy of Being," in The Rooster's Egg:
On the Persistence of Prejudice (Cambridge, Mass.: Harvard University Press, 1995), p. 177.
57. Ibid., pp. 174, 175.
58. Cornel West, Race Matters (New York: Vintage Books, 1994), p. 86.
59. Linda Gordon notes that "the meaning of 'welfare' has reversed itself. What once
meant well-being now means ill-being." Gordon, Pitied but Not Entitled: Single Mothers and
the History of Welfare 1890-1935 (New York: Free Press, 1994),p. 1.
60. Wendell, Rejected Body, p. 148.
61. Martha Minow, Making All the Difference: Inclusion, Exclusion and American Law
(Ithaca, N.Y.: Cornell University Press, 1990), p. 268.
62. Ibid., pp. 248, 249 (my emphasis).
63. Ibid., p. 267. See also Kathy Ferguson, The Feminist Case against Bureaucracy
(Philadelphia: Temple University Press, 1984).
64. Gordon, Pitied but Not Entitled, p. 2.
65. Jane Flax, "Displacing Woman: Toward an Ethics of Multiplicity," in Bar On and
Ferguson, Daring to Be Good, p. 152 (emphasis in the original).
66. Verena Andermatt Conley, Ecopolitics: The Environment in Poststructuralist Thought
(New York: Routledge, 1997), p. 26. (emphasis in the original).
67. Maria C. Lugones, "On the Logic of Pluralist Feminism," in Card, Feminist Ethics,
p. 43 (emphasis in the original).
68. Val Plumwood, Feminism a nd the Mastery of Nature (London: Routledge, 1993), p.
119.
69. Plumwood writes of "variations on the reason/nature story which develop in tan-
dem with the emerging [capitalist] systems of individual appropriation and distribution
[and] turn on a concept of ... individual rationality, which denies both human and social
others and earth others in its concept of the rational egoist subject of social and economic
life" (ibid., p. 141).
70. I owe the term "ethicopolitical" to Bar On and Ferguson. In the introduction to
Daring to Be Good, p. xiii, they pose "the central ethico-political question of feminism" thus:
"What is wrong with our current state, government, economy from a feminist perspective, and
what ethical and political alternative values, visions, and strategies should feminists stand for and
engage in?" (emphasis in the original).
71. Donna Haraway, "A Cyborg Manifesto: Science, Technology, and Socialist-Feminism
in the Late Twentieth Century," in Simians, Cyborgs, and Women, p. 173.
72. Raymond Williams connects "ecology" to "habitat" understood as "a characteristic
208 Autonomy and the Social

living place," in Keywords: A Vocabulary of Culture and Society, 2nd ed. (London: Harper-
Collins, 1983), p. I l l , noting that the term enters everyday use in English in the mid-
twentieth century, "though its scientific use . . . dates from the 1870s." It evolves into "the
study of the relations of plants and animals with each other and with their habitat." Post-
19605 ecology reinterprets economics, polities, and social theory "from a central concern
with human relations to the physical world as the necessary basis for social and economic pol-
icy." I am claiming that in such reinterpretations, questions about the politics of knowledge
claim an equally significant place.
73. Conley, Ecopolitics, p. 110.
74. For an analysis of analogy's low epistemic status, see Alison Wylie, "The Reaction
against Analogy," Advances in Archaeological Method and Theory 8 (1985): 63- 111.
75. Lucy Candib, Medicine and the Family: A Feminist Perspective (New York: Basic
Books, 1995), p. 159.
76. Ibid., p. 76.
77. See also Chris J. Cuomo, Feminism and Ecological Communities: An Ethic of Flour-
ishing (London: Routledge, 1998), especially chap. 7, "Activism That Is Not One."
78. Conley, Ecopolitics, pp. 110, 114 (emphasis in the original).
79. See, for example, Susan Griffin, Woman and Nature: The Roaring inside Her (New
York: Harper & Row, 1978); Carolyn Merchant, The Death of Nature: Women, Ecology, and
the Scientific Revolution (San Francisco: Harper & Row, 1980); Judith Plant, ed., Healing the
Wounds: The Promise of Ecofeminism (Philadelphia: New Society Publishers, 1989); Irene Di-
amond and Gloria Feman Orenstein, eds., Reweaving the World: The Emergence of Ecofemi-
nism (San Francisco: Sierra Club Books, 1990); and Janet Biehl, RethinkingEcofeminist Pol-
itics (Boston: South End Press, 1991).
80. Plumwood, Feminism and the Mastery of Nature, p. 174. For discussions of deep
ecology, see Bill Devall, "The Deep Ecology Movement," Natural Resources journal'20 (April
1980); and Arnc Naess, "The Deep Ideological Movement: Some Philosophical Aspects,"
Philosophical Inquiry?, (1986): 10-31.
81. Donna Haraway, "Otherworldly Conversations, lerran Topics, Local Terms," in
Biopolitics: A Feminist and Ecological Reader on Biotechnology, ed. Vandana Shiva and Ingunn
Moser (London: Zed Books, 1995), p. 70.
82. Flaraway, Modest^Witness, p. 270.
83. Harding, "Rethinking Standpoint Epistemology"; Linda Alcoff, Real Knowing:
New Versions of the Coherence Theory (Ithaca, N.Y.: Cornell University Press, 1996); Lorraine
Code, Epistemic Responsibility (Hanover, N.H.: University Press of New England, 1987).
84. Catriona Sandilands is eloquently critical of claims that the "voice of ... science
can 'see' nature for what it really is, and translate it into a form that can be readily perceived."
See Sandilands, "From Natural Identity to Radical Democracy," Environmental Ethics 17
(Spring 1995): 75-91, 76.
85. For an example of such a mapping, see Chandra Talpade Mohanty, "Women
Workers and Capitalist Scripts: Ideologies of Domination, Common Interests, and the Poli-
tics of Solidarity," in Feminist Genealogies. 1 discuss Mohanty's essay as a sample of ecological
thinking in "How to Think Globally: Stretching the Limits of Imagination," Hypatia: A Jour-
nal of Feminist Philosophy 13, no. 3 (Spring 1998): 73-85.
86. Card, Unnatural Lottery, p. 32; Walker, Career Selves, pp. 115, 124.
87. Cornelius Castoriadis, "From Ecology to Autonomy," Thesis Eleven, no. 3 (1981):
8-22, 8. See also his Philosophy, Politics, Autonomy: Essays in Political Philosophy, ed. David
Ames Curtis (New York: Oxford University Press, 1991).
88. Castoriadis is referring to technology, but his argument permits this expansion. He
 The Perversion of Autonomy dr the Subjection of Women 209

writes: "The transformation of present technology . . . will have to seize part of what exists
at present as technology and utilize it to create another technology" ("From Ecology to Au-
tonomy," p. 20).
89. Castoriadis discusses the creative power of the "instituting imaginary" in "Radical
imagination and the social instituting imaginary," in Rethinking Imagination: Culture and
Creativity, ed. Gillian Robinson and John Rundell (London: Routledge, 1994).
This page intentionally left blank
 II
RELATIONAL AUTONOMY
IN CONTEXT
This page intentionally left blank
 9

CHOICE AND CONTROL IN

FEMINIST BIOETHICS1

Susan Dodds

Within bioethics, respect for—and protection of—personal autonomy has been

linked historically to the requirement that prior to invasive treatment, physicians
obtain the informed consent of their patients. In placing a duty on physicians to
inform patients adequately about their condition and the proposed treatment
and to seek their patients' consent, the paternalism previously associated with
medicine was thought to have been overturned in favor of respect for patients'
autonomy. Over the past three decades, the effect has been to change the image
of the doctor-patient relationship from the presumed beneficent paternalism of
the doctor, acting on the best (medical) interests of the compliant patient, to a con-
tract between patient-consumer and doctor-service provider. In this approach
the physician has the role of expert adviser, providing information to the con-
sumer, who then makes her or his health-care decisions free from paternalistic
medical interference.
The narrow focus on consent as the sole locus for autonomy considerations in
medicine has worked to constrain debate about autonomy in health care. This effect
is similar to the way in which the protection of freedom of speech as the paramount
expression of personal autonomy in social life constrains examination of the ways in
which hate speech and other protected public expressions limit the capacity for some
members of a society to develop and exercise personal autonomy.2 Identifying au-
tonomy with informed consent in health care makes it easy to overlook other limi-

213
214 Relational Autonomy in Context

rations on autonomy in health care, just as the identification of autonomy with free
speech in civil law facilitates a very narrow understanding of the effects of social in-
stitutions and practices on the development and exercise of personal autonomy. For
example, the identification of respect for autonomy with informed consent presup-
poses that in the absence of pathology or extreme youth, all patients can be assumed
to be fully autonomous agents; thus, those patients found not to be autonomous are
presumed to be pathological or infantile and are treated appropriately in a paternal-
istic manner. Furthermore, identifying respect for autonomy with informed consent
presupposes that ethical concern should be directed to the actions of the physician in
obtaining consent: (whether the physician gives full and adequate information,
whether the physician has unduly influenced the decision of the patient, whether
the physician explains the information clearly, etc.) and not to the decision-making
process of the patient. Finally, the focus on informed consent ignores the ways in
which health-care practices influence the development and demise of the capacity
for personal autonomy.
Feminists who arc working in the area of biocthics have written about their am-
bivalent attitude toward the ideal of autonomy as it is understood in that field.3
Feminists are not alone in questioning die role of autonomy in bioethics, yet their
concerns more readily draw out questions about the nature of autonomous agents
and the social conditions conducive to the development of autonomy overlooked by
this narrow "informed consent" approach. On the one hand, the dominant concep-
tion of autonomy used in the bioethical literature—autonomy understood as in-
formed, rational, free choice—ill fits women's experience of medical intervention,
especially in the area where feminist critiques have contributed most effectively, re-
productive decision making. On the other hand, if women see themselves as lack-
ing autonomy in these choices, it appears that the only alternative supported by
bioethics is to treat women with beneficent medical paternalism, thereby denying
women's agency.
Given the cost of relinquishing the ideal of autonomy, there are good grounds
for reexamining the concept as it is employed in bioethics, asking how feminists
have responded to it and whether an alternative conception of autonomy that is
more applicable to health-care contexts could be developed. In this article I chal-
lenge the link, which is prevalent in the bioethical literature, between respect for
autonomy and procedures for obtaining informed consent to medical treatment. I
argue that an appropriate, relational approach will attend not just to health-care
choicesbut also to the ways in which health-care practices can contribute to the de-
velopment and shaping of people's capacity for autonomy. It will also attend to the
role that medicine plays in socialization and the institutional factors that can impede
or enhance both the development and continued exercise of autonomy.
The structure of this article is as follows. In the first section, I briefly recount the
sources of the identification of personal autonomy with informed consent in
bioethics. In the second section, I review the ways in which feminist bioethicists have
problematiy.ed the focus on respect for autonomy that is understood as free, informed
consent. In the third, I examine Susan Sherwin's approach to relational autonomy
and her critique of the conception of autonomy employed in mainstream bioethical
literature, and I identify limitations with that account of relational autonomy. In the
 Choice and Control in Feminist Bioethics 215

fourth, I use Diana Meyers's procedural approach to personal autonomy to show that
an adequate approach to respect, protection, and promotion of personal autonomy
in health care is concerned with an array of health-care practices that move well be-
yond the narrow focus on choice in medical decision making. Whereas rational
choice in health-care decision making may reflect personal autonomy, so that re-
specting those choices would constitute respect for that person's autonomy, auton-
omy and choice ought not be conflated.4 In making a rational choice in a pressing
health-care context, an individual would usually consider the currently available al-
ternatives; her current understanding of her values, desires, and goals; and what ac-
tion or choices she believes, based on the information she has, would be most likely
to bring about their realization. Her rational choice, however, may reflect hetero-
nomously acquired values, desires, and goals; her medical condition may affect her
understanding of herself and her priorities; the available information and alternatives
may be the result of autonomy-limiting policies and practices; and so on. Within
health care there is a range of autonomy-influencing practices and policies to which
those who wish to promote, protect, and respect autonomy ought to attend.

Bioethics
The discipline of bioethics grew out of a change in public attitude toward health
care, which coincided with rapid changes in medical technology that affect both the
beginning and end of life. Although these medical advances were hailed as break-
throughs, in many cases they failed to be cures. Rather, they made life possible when
it might not otherwise have been so. Critics, including some in the medical profes-
sion and the families of those whose lives were extended, began to question whether
the life thus saved was worth living, whether the opportunity offered was worth tak-
ing. In particular, people started questioning whether physicians ought to be left
with the responsibility for making decisions of this kind.
As bioethics emerged as a discipline, Thomas Beauchamp and James Childress
developed their very influential, principlist approach to bioethics. The approach is
based on the application of four general moral principles to particular ethical prob-
lems. The four principles—autonomy, justice, nonmaleficence, and beneficence5 —
are grounded in Kantian deontology, Rawls's theory of justice, Mill's utilitarianism,
Judeo-Christian morality and even a vestige of the Hippocratic oath. Beauchamp
and Childress's metaethic shared some similarities with W D. Ross's intuitionism,
particularly his understanding of prima facie moral considerations.6
According to Beauchamp and Childress, the tetrad of principles are to be un-
derstood in the following way:

• Autonomy is to be understood as "personal rule of the self that is free from both
controlling interferences by others and from personal limitations that prevent
meaningful choice, such as inadequate understanding." The principle of respect
for autonomy requires respecting those choices made by individuals whose deci-
sions are free from external interference or personal limitations.7
• The principle of nonmaleficence is that one ought not to inflict evil or harm,
and in general this principle has priority over the principle of beneficence.8
216 Relational Autonomy in Context

• ( he principle of beneficence covers three kinds of duties, all of which concern

promoting the good of others: one ought to prevent evil or harm, one ought to
remove evil or harm, and one ought to do or promote good.9
• The principle of justice requires fair or equitable distribution of benefits and
burdens within a society.10

'I hese principles are to be used as prima facie considerations in responding to an

ethical problem. Beauchamp and Childress argue that the principles are sufficiently
universalizable that any two careful ethical thinkers would apply them in the same
way to similar cases, as well as flexible enough to respond to changing contexts, tech-
nology, and social concerns. 11 They are applicable to concrete incidents, avoiding
the difficulty of applying a full-blown moral theory to specific cases. Despite Beau-
champ and Childress's insistence that all four principles have a role to play in
bioethics, the principle of autonomy has become elevated so that it is now, surely,
the first among equals.12
1 he focus on autonomy has a good pedigree in both Kantian deontology and
Millian utilitarianism. The conception of autonomy found in much of the literature
in bioethics is more closely linked with Mill's understanding of personal liberty and
the appropriate limitations of state authority 13 than with moral psychological ac-
counts that question how and whether a self can be autonomous. Because bioethics
is primarily concerned with the proper moral conduct of health-care providers, ex-
amination of the requirements of autonomy is directed at the actions that such
providers ought to take to avoid interference with or constraint of patients' auton-
omy. Reflecting liberal political theory, respect for freedom (autonomy) is under-
stood to require that the state, or health-care providers, do nothing that interferes
with the self-regarding choices of rational adults. Thus within mainstream bioethics,
the requirement to respect autonomy is understood as the duty of health-care
providers to identify those patients who have the capacity for choice, in terms of de-
velopmental criteria and the absence of psychological pathology; to offer those pa-
tients choices concerning their health care and the information necessary to make an
informed choice; and to respect the free, informed choices they make. As such, au-
tonomy can be quite easily tied to the requirement of informed consent to medical
procedures.1''1
I argue that the conception of autonomy that has emerged in bioethics is nar-
row and underthcorized. If we first think of autonomy in terms of the duties of
health-care providers to autonomous agents and then link it with choice and con-
sent, its scope is seriously constrained in a number of ways. Like the liberal political
and theoretic conceptions of legal and political autonomy that have been roundly
criticized by feminist political theorists, the conception of autonomy used in
bioethics is rationalistic, atomistic, and individualistic. It assumes something like an
atomistic individual, making a choice wholly for herself or himself. It assumes that,
paradigmatically, individuals are equally rational and able to reflect on complicated
choices once given adequate information. It ignores the social circumstances and
power relations that affect choice contexts. Finally, it rules as out of order questions
about any effects of the choice other than those selected by the chooser as signifi-
cant. Because consent processes emphasize the responsibilities of health-care work-
 Choice and Control in Feminist Bioethics 217

ers to offer information and avoid undue influence, informed consent can be un-
derstood as, at best, voluntary choice. Medical choices made in the absence of igno-
rance, coercion, or impediment to decision making capacity can be understood as
voluntary. Whether those choices reflect what a person truly values, wants, or be-
lieves, however, is not something that can be determined simply by ensuring the ab-
sence of these impediments to choice.
In many ways it is astounding that this conception of autonomy, shorn from
the self who chooses, has become the guiding light for so much of bioethics. It may
be true that respect for autonomy can generate answers to bioethical problems, but
this conception of autonomy seems particularly unsuited to the kind of decision
making that goes on in health care. Many of the important, but by no means un-
usual, health-care decisions that individuals, friends, and families make are far re-
moved from the cool, reflective, clear-headed decision making that is the paradigm
of this view of autonomy. For many people, health-care decisions are made in a state
of confusion, and the chooser is influenced by a number of internal and external
pressures, including pain, discomfort, worry, and concern for others. In making de-
cisions, patients may be primarily concerned with their relationships with others in
several ways. Furthermore, their choices may affect people they have never met; and
particular patients may, for the first time or for the umpteenth time, feel powerless
and in a position of subordination to the health-care providers.
This conception of autonomy seriously constrains how we may conceptualize
those who are not fully autonomous and how they are treated as a result. First, in
this conception of autonomy, it is unclear how health-care workers ought to treat
those who have some degree of autonomy but luck fall autonomy. In practice those
people whose capacity for choice might not be thought fully autonomous, for ex-
ample, people in the early stages of Alzheimer's disease (who may have difficulty re-
membering some recent events but who may at the same time have a clear under-
standing of their settled preferences and central values), may be treated in one of
two unacceptable ways. On the one hand, because autonomy is identified with in-
formed consent, if a person in this situation is given adequate information and
makes a choice that appears to reflect her or his stable preferences, that choice will
be treated as autonomous, even if the person has failed to understand or retain
salient features of the information required for a genuinely autonomous choice.
Alternatively, the person might display some signs of incompetence, be diag-
nosed as suffering a mental illness, and thus have her or his authority to make
self-determining decisions removed. In this latter case, even if the person has the
capacity to make a specific health-care decision, her or his claim to make au-
tonomous choices is undermined by the diagnosis of mental incapacity. At best,
then, the person will be treated in a way that protects her or his "best interests," as
understood by health-care professionals.
This lack of a middle ground, of an awareness that the capacity to make health-
care decisions may admit of degrees, is one effect of the identification of autonomy
with informed consent, which can be particularly harmful to women's interests.
Given the effects of patriarchy, the kinds of medical choices women are asked to
make, and the cultural association between "femininity" and "irrationality," women's
options in health care are frequently constructed in ways that limit their autonomy.
218 Relational Autonomy in Context

feminists attempting to challenge those conceptualizations and practices that inad-

equately respond to -women's lives and experience often find themselves stuck with
two unacceptable alternatives. They may choose to accept the limitations of the
identification of personal autonomy with informed consent, and so preserve for
women some, imperfect, control over their bodies; or they can attempt to reject
those limitations, but in so doing risk having women conceptualized as, in general,
incompetent to make health-care choices.

Feminist Bioethics
Given this unpalatable set of options, feminists working on bioethical issues have at-
tempted to reshape the conceptual terrain on which women, and others subject to
oppressive social conditions, are expected to make health-care decisions. Not sur-
prisingly, there has been an array of different responses to these tensions.
Rosemarie Tong writes that there are three key approaches in feminist bioethics:
(1) liberal feminism, focusing on autonomous choice; (2) radical feminism, which
illuminates women's limited control over the institutions of health care, reproduc-
tion, and sexuality; and, (3) cultural feminism, sometimes identified with a feminine
ethic of care.15 long selects these three approaches to bioethics for special attention
because "their different perspectives on the nature, form (embodied or disembod-
ied), and value of the self's relationship to others explain liberal feminists' emphasis
on issues of choice, radical feminists' emphasis on issues of control, and cultural
feminists' emphasis on caring."16
I think that each of these three approaches contributes to a general critique of
the conception of autonomy found in bioethics, and it can contribute to the devel-
opment of an alternative conception of autonomy that better reflects our embodied
existence and the concrete social context of health-care decision making. At the
same time, all three contain significant limitations because they fail to adequately
challenge the equation of autonomy with informed consent, thus failing to draw
women out of the tension between accepting a conception of autonomy understood
as choice, independent of the etiology of the choice, or rejecting the value and sig-
nificance of personal autonomy for women.
Feminist writings on reproductive technology have highlighted the ways in
which limited knowledge, power differences, interpersonal relationships, and the
social meanings given to activities influence both autonomous decision making
and the ethical landscape of health-care practice in certain contexts. The light
shed by feminists on medical and social practices surrounding reproduction can
be used to illuminate the limitations of the conception of autonomy in bioethics
more generally.

Choice
Liberal feminists who are confronting ethical questions about reproduction argue
that there is nothing special about women's reproductive capacity that justifies lim-
iting their freedom to choose reproductive alternatives that are not inherently harm-
ful to others. Laura Purdy's arguments concerning surrogate pregnancy, for example,
 Choice and Control in Feminist Bioethics 219

use the language of contract and reproductive liberty to argue that women ought to
be free to be or to use the services of surrogates.17 Liberal feminists deny that "mere"
biological difference is morally significant and emphasize the importance of ex-
pressing autonomy through choice. Purdy rejects the paternalism involved in pro-
tecting women from their own choices and questions whether the risk of harm that
may follow choices about pregnancy or child rearing are so very different from other
kinds of choice.
For liberal feminists, the fact that women have a different biological relation to
procreation than men is not grounds for undermining women's autonomy, either
through exclusive focus on the welfare of the fetus or future person or by attribut-
ing to women an inability to think clearly about reproduction because of their bio-
logical or social ties to gestation and motherhood. Liberal feminists have drawn at-
tention to the ways in which women's autonomy has been trammeled by medical
practice and research in reproduction, leaving women, on the one hand, resorting
to dangerous treatments to avoid or achieve pregnancy because of poor information
or limited alternatives and, on the other hand, subjecting women to massive inter-
vention into and monitoring of pregnancy and birth.18 Furthermore, liberal femi-
nists show how choices are restricted by health-care policy, for example, through leg-
islative or policy restrictions on access to donor insemination or in vitro fertilization
(IVF) or by surrogacy practices that limit potential gestational mothers' access to in-
formation or legal advice.
Liberal feminists thus argue that the scope of personal choice ought not to be
arbitrarily limited in a way that unjustifiably limits women's health-care alternatives,
and that respect for personal choice in health care requires the provision of accurate,
clear, and appropriate information. At the same time, liberal feminist bioethicists
challenge assumptions about women's supposed inadequate competence to choose
autonomously because of their gender. The liberal feminist approach is consistent
with the bioethical focus on respecting autonomy through recognition of informed
choice. Also, it forces the bioethicist to challenge social perceptions or prejudices
that would limit the scope of women's choices and demands that bioethicists attend
to the obstacles to effective communication of medical information.
Two kinds of criticisms can be raised about the adequacy of the liberal feminist
conceptions of autonomy, women's embodiment, and decision making. First, the
liberal feminist view assumes that health-care decision making occurs in a social vac-
uum. On the one hand, it assumes that the social factors that influence the avail-
ability of alternatives are neutral and, on the other hand, that all people are equally
well placed to make health-care decisions because no competent person is unduly
constrained by factors like oppressive social conditioning.
Second, the liberal feminist approach does not recognize the various ways in
which women's embodiment can affect women's control in their decisions, particu-
larly in regard to reproductive and child-rearing decisions. Menstruation, pregnancy,
childbirth, and breast-feeding, for example, are not activities in which participation
can be chosen or rejected in the same way that, for example, purchasing a book, de-
ciding to practice the piano, or building a bookshelf are chosen or rejected. An ad-
equate account of autonomy in health care must recognize the significant differ-
ences among kinds of choices. Different kinds of choices are affected by the degree
220 Relational Autonomy in Context

or extent to which bodily processes are involved and by the personal or social signif-
icance attributed to such processes. Although it is important to avoid equating
women's possibilities with their reproductive capacities, the differences between re-
productive choices and other kinds of choices must be recognized. The liberal fem-
inist approach, by failing to recognize adequately the role of human embodiment in
health-care decision making is unable to account for these differences.' 9

C Control
Radical feminist approaches to bioethics emphasize the role that patriarchy has
played in the institutions of medical science, marriage, heterosexuality, and the fam-
ily, arguing that women's reproductive capacity itself has been the source of their op-
pression and that the apparent choices they make about reproduction are not really
autonomous choices because the contexts are constructed by patriarchy. 20 The iden-
tity of woman, under patriarchy, is tied to her reproductive capacity, and this ren-
ders woman connected, dependent, and related to others in a manner that is both
damaging to her and that subordinates her to men. In this view, women lack con-
trol over reproductive technology and medical practice and thus should be wary of
being exploited by them.
In the radical feminist approach, the expression of a choice in a medical con-
text should not be understood simply as an expression of autonomous choice be-
cause the social context within which women make reproductive decisions is both
exploitative and oppressive. Until women gain control over reproductive technolo-
gies and until the link is broken between being a woman and having an inferior so-
cial status, because of one's reproductive capacities, women ought to reject forms of
medical intervention that exploit this connection. For radical feminists like Robyn
Rowland and Janice Raymond, then, women's reproductive capacity, as understood
by patriarchy, is the source of women's oppression. For women to be free they must
not simply choose to use, or to be used by, the tools available to reinforce their op-
pressed state; rather they must seize the controls. 21
Radical feminist contributions to bioethics demand an awareness of the effects
of race, class, gender, ability, and sexual orientation in the distribution of power in
a context of medical care, rather than a narrowly construed understanding of au-
tonomous choice. The imbalance in power between physicians and patients, be-
tween husbands and wives, and between rich and poor become significant features
for understanding the ethical context of health-care decision making. A view that
focuses on purely formal criteria for autonomous choice and that overlooks the so-
cial and political context of health-care choices will be found inadequate by radical
feminism.
The radical feminist approach to bioethics challenges the presumption that the
rational choices of adults are autonomous in the absence of certain kinds of pathol-
ogy. In this view, social conditions affect the development of personal autonomy
and the desirability of the alternatives among which a person rationally chooses. Al-
though a physician may not have unduly influenced a patient's choice to seek a spe-
cific kind of treatment, that patient may not have sufficiently developed skills of
critical self-understanding to challenge, for example, the socialization of women
 Choice and Control in Feminist Bioethics 221

as mothers and wives, which leads her to seek medical reproductive assistance. The
radical feminist approach queries the identification of respect for rational, well-
informed, uncoerced choice with autonomous choice, in the absence of an exami-
nation of the significant social factors that shape decision making contexts.
The radical feminist approach to bioethics, however, contains some inconsis-
tencies. For example, the account of patriarchy that supports the radical feminist
mistrust of medical technology is not similarly applied to governmental institu-
tions that are called on to protect women. In this account, women need to be pro-
tected from the harms that result from making nonautonomous decisions when
they lack control over the context of the decision. In one version of the argument,
for example, as put forward by Robyn Rowland, all interaction between women
and medical professionals is regarded as necessarily exploitative, but state interven-
tion seems to be viewed as nonproblematic.22 Surely state institutions have some-
times been used to serve patriarchal interests, and some medical interactions have
been nonexploitative.
An adequate account of the effects of patriarchal oppression on women's au-
tonomy would need to explain the different ways in which the social contexts of
health-care decision making influence women's choices, (including the reasons why
these influences affect women as they do), and it would need to be sufficiently de-
veloped to distinguish among more and less exploitative medical interventions and
practices. There is an inconsistency, however, between the view that women as a
group may be exploited by unregulated access to reproductive technologies and the
lack of a demand for similar regulation of other potentially exploitative practices.
Even the weaker claim that women's choices in reproduction are likely to be influ-
enced by internal and external factors that can cloud their judgment needs to be
made consistent with the ways in which other significant health choices can be sim-
ilarly clouded; reproductive decision making ought not be singled out for special
treatment.

Care
Cultural feminism draws on the work of Carol Gilligan and focuses on caring rela-
tions among persons, rejecting the isolated atomism of justice-based ethical theories.
Cultural feminism rejects the centrality of autonomy in bioethical principlism. Un-
like those radical feminists who view women's interconnection with others as the
source of oppression, cultural feminists positively value the activity of caring and re-
sponsibility for care. For cultural feminists, the proper focus of ethical concern is in-
terrelationships, the connections between people.23 Policies and practices are to be
evaluated in terms of the degree to which they foster or dissolve caring relationships.
Some versions of cultural feminism are not particularly feminist in approach; for ex-
ample, Nel Noddings argues that ethical behavior involves putting oneself at the
service of others and valuing the quality of relationships,24 an approach that might
well reinforce a traditional view of women as subservient, other-directed, and ap-
propriately relegated to the role of carers rather than agents in the world.25
Some writers working within the care perspective, however, have articulated ex-
plicitly feminist versions of the approach, highlighting the need for a critical ac-
222 Relational Autonomy m (Context

count of oppression to supplement any account of care, so that social practices that
use care roles to exploit or oppress can be criticized. JoanTronto, for example, does
not simply argue that women's approach to ethical decision-making is better cap-
tured by the ideas of care and interconnection, but she also claims that a well-ar-
ticulated ethical theory in which care has a central role to play can best challenge all
forms of oppression.26 Susan Sherwin, in No Longer Patient, shows how this kind of
care approach, informed by an account of oppression, can be applied to bioethical
issues.27
The care approach to bioethics highlights the many significant ways in which
those facing decisions in health- care contexts do not make choices as isolated, ra-
tional atoms. Rather, their choices are very often affected by concern for a range of
relationships that they value, relationships that tend to be discounted by the under-
standing of autonomous choice in mainstream bioethics. The care approach makes
sense of the apparent altruism involved in many health-care choices, of parents,' pa-
tients,' and children's concern for the effects of different kinds of care on others, and
of the responsibilities that go with certain kinds of choices. The care focus gives us
an awareness of the relationships between people and the ways these relationships
are affected by health care.
Both cultural feminist and radical feminist interventions in bioethics challenge
the assumption that medical decision making is apolitical and asocial and that per-
sons are best understood as independent, self-interested, atomistic choosers. The
cultural feminist critique can be seen as the flip side to the radical feminist critique.
The former account reminds us that social relationships are not wholly negative in
their influence on people's lives. Relationships with others, including relationships
founded on chance rather than choice, are central to human existence and can be
valuable contributions to people's lives. Autonomy ought not to be conceived as in-
dependence or isolation from others; it ought to be conceived as a way to foster
nonoppressivc relationships of care. Health-care providers ought, therefore, to at-
tend to the relationships involved in health care, that is, the significant relationships
of patients that are affected by illness, health care, and health-care decision making.
At the same time, bioethics ought not to presume that people will all share the same
form of personal relationships; some relationships reflect and enhance the auton-
omy of those engaged in them, whereas others may not.
Although each of the three feminist approaches identified by'long contributes
to the critique of the narrow understanding of autonomy as rational choice within
mainstream bioethics, each has been found to lack an adequate positive account of
autonomy in health care. Liberal feminism's narrow attention to choice, independ-
ence, and freedom lacks a sufficiently well-developed understanding of the ways in
which socialization and cultural contexts shape choices. Radical feminist accounts of
social control over the circumstances of health-care choice are insufficient to offer
any guide to developing a less oppressive health-care ethics. Finally, cultural feminist
accounts have failed to provide a critical analysis of the relationships of care and de-
pendence that could guide the identification of which relationships ought to be fos-
tered in health-care contexts and which arc inherently oppressive. Any adequate un-
derstanding of autonomy in health care will need to offer an alternative account of
autonomy, one that incorporates the liberal, radical, and cultural feminist critiques.
 Choice and Control in Feminist Bioethics 223

Relational Autonomy
In her recent writings, Susan Sherwin presents an argument for a feminist perspec-
tive on bioethics and autonomy that develops an account of autonomy informed
by these critiques. She argues for a richer, more complex understanding of auton-
omy than the one described in the mainstream bioethics literature.28 Sherwin rec-
ognizes that respect for autonomy has an important role to play in health-care
ethics, yet she challenges the standard assumptions made about the criteria for au-
tonomous choice in health care. Her approach avoids the problems of those criti-
cized in the previous section, as she does not overlook or oversimplify social rela-
tions; she provides a coherent account of oppression, and she recognizes that
human dependency and relationships can be a double-edged sword in a patriarchal
culture.
The account of autonomous choice in health care that Sherwin offers contex-
tualizes health-care decision making. She gives due recognition to the ways in which
health-care choices are shaped by a range of social values. For example, on the one
hand, the scope of the decision is marked out by the institutional framework of
health-care provision, including health-care resources, medical education, and com-
munity care. On the other hand, what people seek from health care is moderated by
the social influences that shape their self-understanding to varying degrees. For ex-
ample, women's demand for reproductive technology and cosmetic surgery are, at
least in part, effects of the patriarchal association of femininity with childbearing
and beauty understood as attractiveness to men. Through her contextualization of
health-care decision making Sherwin begins to articulate an alternative understand-
ing of autonomy in health-care ethics.
In Sherwin's view, feminist bioethics must be understood as an ethics of the op-
pressed:
Feminists share a recognition that women are oppressed in our society and an un-
derstanding that their oppression takes many different forms, compounded often
by other forms of oppression based on features such as race, ethnicity, sexual ori-
entation, and economic class. Because feminists believe that oppression is objec-
tionable on both moral and political grounds, most are committed to transform-
ing society in ways that will ensure the elimination of oppression in all its
forms.2?

Feminist bioethics is focused, then, on identifying those features of health care that
exacerbate or ameliorate oppression. Although feminists have appealed to autonomy
to reject paternalistic treatment by the medical profession, there are good grounds
for critically examining the ways in which the dominant conception of autonomy
used in bioethics may contribute to systems of oppression.
According to Sherwin, autonomy is usually understood, in bioethics, as apply-
ing to particular decisions in particular contexts. A patient is thought of as au-
tonomous with regard ro a specific choice if the following criteria are met:

1. The patient is deemed to be rational with regard to the decision at hand.

2. The patient makes a choice from a set of available options.
224 Relational Autonomy in Context

3. The patient has adequate information and understanding.

4. The patient is free from explicit coercion toward or away from one of these
options. 30

Sherwin argues that each of these criteria raises problems for feminists concerned
with the eradication of oppression.
1'irst, if oppressed people are less likely to be thought of as autonomous agents
— indeed, if the core of autonomy, rationality, lias been conceptualized as anti-
thetical to the characteristics identified with women—there is good reason to ques-
tion the gender neutrality of autonomy as an ideal.31 The rational competence of
women and other oppressed groups is frequently questioned, insofar as they arc
thought to lack sufficient emotional distance and objectivity to act rationally. 32
Members of oppressed groups may lack sufficient autonomy for their choices to be
accorded adequate respect by health-care providers — leaving them dependent on
medical paternalism.
Second, the set of options a patient chooses from may already be constructed
in a way that seriously limits the patient's autonomy. Decisions about medical re-
search priorities and funding of the health-care system affect what alternatives are
available for the physician to offer to the patient. These decisions may reflect dis-
criminatory or biased practices that: affect the particular patient's autonomy, and yet
in most bioethical discussions these alternatives are thought of as a given set that
does not require ethical scrutiny. 33
Third, information made available to patients is inevitably that information
deemed relevant by the health professionals who care for them; but the large gap be-
tween the life experience of health professionals, who arc relatively privileged, and
their sometimes seriously disadvantaged patients makes the likelihood that the former
will provide information that meets the specific needs of their patients rather slim.34
Fourth, the forces that contribute to the coercion of the oppressed may be very
difficult to identify. Drawing on Sandra Bartky's characterization of psychological
oppression, Sherwin argue.s that those who experience this form of oppression may
identify with their oppressed condition. 35 It is distorting, then, to describe as au-
tonomous the choices made under circumstances of oppression. Echoing the views
of radical feminists, Sherwin says that the absence of explicit coercion at the time of
making a decision does not make that choice free. 3f>
1 agree with each of Sherwin's criticisms of the dominant bioethical conception
of autonomy. However, I would take the criticisms further. In drawing out their im-
plications, I think it is important to note that the inadequacies of the dominant
conception of autonomy apply to any person who is in need of health-care and who
must come to understand one's condition, make choices about one's care, and adapt
to one's changed situation. While the dominant bioethical conception of auton-
omy serves to reinforce discriminatory attitudes toward women and other oppressed
groups and their agency, 1 think that the problem with the dominant conception of
autonomy is not simply one of oppression or gender.
The conception of autonomy that dominates bioethics, which identifies auton-
omy with informed consent, is inadequate for anyone in serious medical need, al-
 Choice and Control in Feminist Bioetbics 225

though the effects of its inadequacy may often be greater for those who have expe-
rienced oppression. The voluntary choice model is just as inadequate a tool for gen-
uinely protecting and enhancing the autonomy of male patients in health-care situ-
ations as it is for women.
My first criticism of Sherwin, then, is that although she correctly identifies the
kinds of features of the medical environment that limit people's decision-making ca-
pacity, she is wrong to assume that these limitations affect only those who have been
subjected to oppression. A range of features about health-care crises and health-care
institutions pose a threat to the capacity of anyone facing a health-care crisis, not
just those oppressed by gender or social disadvantage.
Let us consider an example. A man (for the purpose of the argument, a white,
able-bodied, tertiary-educated professional) who finds himself in a hospital, having
had a heart attack, or a man who must make a decision about a prostate operation,
after being diagnosed with prostate cancer, suddenly finds himself in a circumstance
in which he is not fully in control of his life; he can no longer assume that his choice
and consent will largely determine his future possibilities. Faced with his own mor-
tality, his ability to engage in abstract reasoning may be clouded. The choices open
to him are likely to be restricted by factors over which he has no control but which
affect the quality of his care. Is he in a teaching hospital? Then he may be offered
trial medication. Does his physician lean heavily toward leaving the prostate alone?
The information given by his physician may not reflect his own set of priorities, and
his socialization may make asking questions difficult. Will the bypass operation ren-
der him impotent? How can he avoid incontinence after the prostate operation? Fi-
nally, the factors that influence his decision, rendering it less free than the autonomy
ideal presupposes, may be ones that have been inculcated in him as part of his en-
culturation. He may be pushed to accept invasive medical procedures because he be-
lieves that it would be weak or unmanly to accept his condition passively. He may be
pushed to accept risky experimental treatment because he believes that it is the only
way to avoid becoming dependent on others for care.
This example illuminates the ways in which even a bastion of patriarchy may
find his autonomy threatened or undermined if his physician exclusively attends to
information and choice in negotiating his health care, rather than understanding au-
tonomy relationally. Even medical specialists find health care threatening and intim-
idating when they face medical emergencies as patients. Those who have been pa-
tients claim that it has changed "completely and forever" the way they practice
medicine and has forced them to rethink the approach to informed consent by
which the physician stands as dispassionate information provider rather than as a
partner in medical decision making.37 The point of the criticism is that the factors
that Sherwin identifies as limiting health-care decision making affect most patients,
not just those who are members of oppressed groups. It is quite true that our imag-
ined white, male professional is likely to have readier access to wider alternatives in
health care than relatively disadvantaged members of the community, but that does
not mean that his autonomous health-care decision making is not open to limitation
in a manner similar to theirs.
My second and more substantive criticism of Sherwin is that although her ac-
count challenges the connection between autonomy and informed consent familiar
226 Relational Autonomy in Context

in the bioethical literature, it still assumes that personal autonomy is correctly iden-
tified solely with medical choices. Whereas Shcrwin wants health-care workers and
policymakers to examine more carefully the features of health-care decision making
that can constrain personal autonomy, her focus is still on autonomy understood as
choice. It may be that if health care practices were revised to acknowledge the effect
of the array of influences on choice that Sherwin identifies, patients would be able
to make not just voluntary choices but also rational choices. That is, their choices
would be made in accordance with principles of rational choice and deliberative ra-
tionality.38 Sherwin's account encourages review of the processes of medical decision
making but does not require any reexamination of the process whereby a person de-
velops the capacity for rational decision making. Her references to the effects of so-
cialization and oppression indicate that she is aware that people's capacities for au-
tonomy are influenced by the environment within which those capacities develop,
and yet she still identifies respect for autonomy in health-care with respect for ra-
tional medical choices.
As I argue below, an adequate understanding of autonomy in health care must
not be restricted to an examination of the exercise of autonomy through choice but
must also encompass an understanding of the ways in which autonomy is developed
or, in Diana Meyers's terms, the ways in which the array of "autonomy competen-
cies"39 are fostered, shaped, and potentially thwarted. In the next section 1 look be-
yond patient decision making to other areas of health care in which autonomy can
be enhanced or limited by health-care providers or health-care systems.

Autonomy Competency and Health-Care Contexts

As I argued in the first section, the identification of personal autonomy with in-
formed consent has meant that health-care workers concerned to respect autonomy
have understood their task as the improvement of processes for obtaining informed
consent and acting in accordance with the consent so obtained. Sherwin's argument
raises questions about the narrow focus on the relationship between the physician
and the patient in the consent process because factors external to that process, for
example, oppression, inadequate socialization, and limited health-care funding, af-
fect the choices available to patients. These factors also influence the patient's (and
physician's) subjective understanding of those choices and shape the capacities of
patients to resist, inform, or control the process of obtaining information and mak-
ing a decision. Sherwin thus pushes against an identification of autonomy with vol-
untary choice and toward an understanding of autonomy in health care as aligned
with contextualized rational choice.
In this section I argue that an adequate understanding of respect for autonomy
in health care must extend to an understanding of the development and exercise of
the capacity for autonomous decision making, rather than focusing solely on in-
formed consent or even rational choice of the kind that Sherwin supports. In argu-
ing that the conception of autonomy appropriate to biocthics must include an un-
derstanding of the development of the capacity for autonomy, I am not arguing that
respect for the informed consent (voluntary choice) or rational choice of patients is
irrelevant to respect for autonomy. Rather, I am arguing that respect for autonomy is
 Choice and Control in Feminist Bioethics 227

not restricted to respect for choices of a certain kind but also requires promotion of
the development of autonomous selves.
Procedural approaches to autonomy, such as those offered by Dworkin, Young,
and Meyers,40 require that autonomy be understood not merely in terms of the
choices people make but also in terms of the relationships among the choices peo-
ple make, their values, and the capacities they engage in making choices.41 Accord-
ing to Meyers, "Autonomous people must be able to pose and answer the question
'What do I really want, need, care about, believe, value, etcetera?'; they must be able
to act on the answer; and they must be able to correct themselves when they get the
answer wrong."42 To conduct this procedure, the autonomous person employs an
array of skills that constitute "autonomy competency." An autonomous person asks
and answers questions about her real wants, needs, concerns, and so on so that she
can structure her life plan to secure personal integration.43
Autonomous self-direction occurs, according to Meyers, both episodically and
programmatically:
Autonomous episodic self-direction occurs when a person confronts a situation,
asks what he or she can do with respect to i t . . . and what he or she really wants
to do with respect to it, and then executes the decision this deliberation yields.
Autonomous programmatic self-direction has a broad sweep. Instead of posing
the question "What do I really want to do now?" this form of autonomy ad-
dresses a question like "How do I really want to live my life?" To answer this latter
question, people must consider what qualities they want to have, what sorts of
interpersonal relations they want to be involved in, what talents they want to
develop.44

Respecting the rational choices of a patient may also respect that person's auton-
omous episodic self-direction, assuming that the person's rational choice reflects her
or his authentic desires, values, and so on. It is less clear, however, that these pro-
cesses of respect for rational choices will be sufficient to recognize programmatic
self-direction, although many significant health-care decisions affect programmatic
self-direction. Furthermore, as was indicated by some of Sherwin's comments, many
of the policies and practices that affect programmatic self-direction in the area of
health may be outside the scope of individual patients' decision making (e.g., legal
prohibitions on voluntary euthanasia, access to home-based care, availability of
community-based support for people with disabilities, provision of universal access
to health care, etc.). These features affect autonomy in a way that is not captured by
the approach to autonomy that only concerns itself with specific medical choices.
According to Meyers, to exercise autonomous programmatic or episodic self-
direction, autonomous people must develop an array of competency skills, those re-
quired for self-discovery (identifying what they want, value, care about, etc.), self-
definition (acting on their desires, values, etc.), and self-direction (correcting their
actions or understanding of their values when they misidentiry them).45 Exercising
these skills constitutes autonomy competency. Rational choice, understood as choos-
ing to act on that course of conduct that, of the available alternatives, best suits
a person's desires, values, or life plan may or may not reflect a person's autonomy
competency. Rational choices, choices made on the basis of principles of rational-
228 Relational Autonomy in Context

ity, display some of the skills of autonomy competency but do not constitute au-
tonomy.46 A person who is fully autonomous will make rational choices that reflect
their authentic desires or values. However, being able to make a rational choice does
not necessarily reflect full autonomy; rational choosers can make rational choices if
they have developed their capacity for rational deliberation in light of rational prin-
ciples, even if they lack skills in self-direction. Even less should making rational
choices be understood as constituting autonomy in Meyers's account.
Applying Meyers's conception of autonomy to health care requires broadening
the focus from a narrow examination of the conditions for choice to an assessment
of the conditions for the development of the skills that encompass autonomy com-
petency. Bioethicists concerned with protection, promotion, and respect for personal
autonomy in health care must direct their attention to the wide array of health-care
contexts that may impede the development of autonomy competency. The following
discussion sketches some implications of Meyers's approach to health-care ethics.

Autonomy Competency and Universal

Access to Health Care
Meyers recognizes the interests that people are likely to have in those things nec-
essary for the development and exercise of autonomy in a range of areas. She
writes that basic interests are to be understood as "generalized means to various
self-regarding ends." While Meyers does not wish the rational presumption that all
people share some basic interests, such as "health, strength, intelligence, power,
wealth, friendship, opportunities, and reputation,"' 17 to override individual prefer-
ences and choices, it seems clear that in the provision of basic welfare services
(health, education, housing, etc.) a society can promote personal autonomy by en-
abling access to those things needed to develop autonomy competency and to pur-
sue an autonomous life. Health-care systems are clearly part of this provisioning, in-
sofar as good health and decent access to health-care facilities will be a precondition
for the pursuit of a very great range of autonomous projects. Thus, one initial im-
plication of Meyers's account for bioethics is that the promotion and protection of
autonomy competency requires the provision of some level of universal access to
health care. Individual needs and wants with regard to services will vary, but it is rea-
sonable to assume that all autonomous people will have an interest in health care
provision sufficient to provide for the basic health-care interests of the members of
the community (consistent with individual choice and preferences). Health-care sys-
tems that do not provide equitable access for all members of the community will not
promote and protect autonomy to the same degree as those that do.

Socialization and the Development of Autonomy

Competency in Health-Care Settings
According to Meyers
Someone is minimally autonomous when this person possesses at least some dispo-
sition to consult his or her self and at least some ability to act on his or her own
beliefs, desires, and so forth, but when this person lacks some other skills trom the
 Choice and Control in Feminist Bioethics 229

repertory of autonomy skills, when the autonomy skills the person possesses are
poorly developed and poorly coordinated, and when the person possesses few in-
dependent competencies that could promote the exercise of available autonomy
skills. I shall say that someone is fully autonomous when this person possesses a
complete repertory of well developed and well coordinated autonomy skills cou-
pled with many and varied independent competencies. Medially autonomous peo-
ple range along a spectrum between these two poles.48

Meyers is particularly concerned to recognize the role of socialization in the devel-

opment of autonomy competency. In her view the capacity for developing and ex-
ercising full autonomy is affected by socialization, and given dominant influences in
current child-rearing and educational practices, most socialization fails to develop
the skills necessary for full autonomy. As a result, most women and men end up
lacking full autonomy; they are minimally or medially autonomous.49 Meyers dis-
cusses the role of appropriate parenting and school education both in creating an en-
vironment within which children can develop autonomy competency and in learn-
ing some basic autonomy competency skills. What she says in this context is applicable
to a number of health-care contexts.
Health-care settings are social environments that can play a socializing role in
the development of autonomy competency. Some health-care settings can signifi-
cantly inhibit the development of the skills that constitute autonomy competency,
quite independently of the exercise of those skills in medical decision making. In
particular, those people who receive long-term or residential care may be signifi-
cantly socialized through the health-care environment. Three specific examples are
(1) children born with severe disabilities or chronic serious illness and those who
suffer trauma of various kinds while young, (2) adolescents and young adults with
psychiatric illnesses, and (3) patients in nursing homes.
Health-care providers and policies can serve to promote full, medial, or mini-
mal autonomy in children. For example, the future possibilities presented to chil-
dren with disabilities or significant illnesses can come to dominate those children's
self-understandings in ways that limit their ability to imagine other alternatives. In
recent years there has been a revolution in the care of children with developmental
disabilities that has given some children much greater opportunities to develop and
exercise autonomy skills, within environments that both foster autonomy and pro-
tect the child's welfare.
Provision of counseling and therapy to assist adolescents and young adults with
psychiatric illnesses to better understand themselves, their illnesses, and their desires
—rather than simply providing treatment to curb undesirable behavior—can en-
hance the development of autonomy competency. Meyers argues that competency
skills can be taught and recommends a range of educational practices to foster per-
sonal autonomy in children.50 These are also applicable to pediatric and adolescent
health-care. Encouraging children with long-term illnesses to participate in decision
making, to ask questions about their care, to voice their feelings and desires, and to
take responsibility for aspects of their own care help them in the development of au-
tonomy competency. On the other hand, uncritically treating such children as ill,
needing total care, and unable to understand or accept their condition robs those
children of the opportunity to develop the skills needed for autonomy competency.
230 Relational Autonomy in Context

Health-care workers can also assist parents and guardians to understand the child's
concerns and condition and can challenge the preconceptions of parents and
guardians that might stifle the child's autonomy development. Children who have
experienced health-care institutions that foster rather than frustrate the development
of autonomy competency will more likely become autonomous selves who can then
make medical decisions that are autonomous. Those whose skills are frustrated or
limited by health-care environments are less likely to develop the capacity for au-
tonomous medical decision making and will be more likely to be minimally or me-
dially autonomous, rather than fully autonomous.
Similarly, the autonomy competency of residents in nursing homes can be
shaped by the health-care environment. Given that Meyers's account of autonomy
recognizes stages or degrees of autonomy, the approach can be used to inform pro-
cesses for preserving those autonomy skills that a person still retains when the per-
son's autonomy has been compromised and he or she is dependent on others for
care. Many elderly residents in nursing homes are physically very frail and ill, yet by
no means do all such residents lack some degree of autonomy. Policies and prac-
tices in nursing homes designed to protect residents or to deal with limited re-
sources very often have the effect of eroding lifelong dispositions to use an array of
competency skills, leaving residents passive, atrophying their competence, and
opening the way for rapid mental and physical deterioration.51 Establishing nurs-
ing home practices that both recognize the limitations experienced by residents and
seek to protect residents' remaining autonomy competency would require a signif-
icant reassessment of the role of nursing homes and health-care provision. Such a
reassessment might well contribute to the breaking down of artificial barriers be-
tween health care and the rest of people's lives in a way that recognized the full sig-
nificance of health care practices in the development, fostering, exercise, and pro-
tection of personal autonomy.

Health-Care Crises and Altered Self-Conceptions

Health-care emergencies very often require patients to confront new information
that significantly alters their self-perception and future possibilities.52 It may be ap-
propriate, then, to treat these crises as inherently autonomy threatening. Although
a fully autonomous person will have an array of skills that can be employed in com-
ing to terms with major life changes, the suddenness and significance of some health-
care crises force a radical reassessment of one's self-understanding, which may effec-
tively derail a persons autonomy competency. The extent of the threat to a person's
autonomy posed by these emergencies will depend on such features as the degree of
autonomy competency the person has, the nature and circumstances of the crisis,
and the various resources available to support the person through it (family, coun-
selors, support groups, etc.).
To adapt the earlier example of the white, able-bodied, professional man, a man
who has been brought to the hospital with a heart attack discovers that he requires
immediate cardiac bypass surgery. This revelation could force him to completely re-
assess his self-understanding. His integrated life plan may have been built, in part,
on the assumption that he was healthy, with a long professional career before him,
 Choice and Control in Feminist Bioethics 231

and that he had the material and psychological resources to cope with whatever
life might throw at him. The diagnosis and subsequent surgery may lead to a self-
realization of the limits to his capacity to simply "will" himself to accommodate his
changed circumstances. Although he may consent to the surgery and recuperative
treatment, the impact of the events surrounding his hospitalization may well cause
an enduring setback to his autonomy competency.53 Any approach that emphasizes
the significance of respect for autonomy at the front end (provision of information
about treatment and respect for informed consent) will be unable to recognize the
impact of trauma, terminal diagnosis, or personal medical emergencies on the self-
understanding of patients and will, therefore, not be able to explain the importance
of providing support for people in coming to terms with their altered circumstances
as part of the protection and respect for autonomy.

Exercise of Autonomy: Information and Alternatives

The discussion of medical emergencies also raises the question about whether the
presentation of more information and more alternatives in health care should be as-
sumed to indicate greater respect for patients' autonomy. If we focus on the devel-
opment and exercise of people's autonomy competency, rather than on the obliga-
tions of health-care workers to avoid interference with the informed consent or
rational choice of patients, there are good reasons to doubt that more information
and more alternatives being offered to patients will always enhance their autonomy.
In both medical emergencies and cases in which a serious condition has been di-
agnosed, which will require a series of treatments with significant side effects, the
provision of information and alternatives may enhance autonomous choice, but
the provision of extensive information and alternatives, without some other form of
support in decision making, may impede a person's decision-making capacity.
In these circumstances, being given vast amounts of information and oppor-
tunities to take full control of one's health-care is no guarantee that autonomy will
be protected. Depending on the array of autonomy competencies that can be sum-
moned to the task, a person may be better or less able to use information critically
to determine how to choose authentically. In many cases, processes of counseling
that help the person determine what it is he or she really wants in the context (and
why) may better promote autonomy than greater information. That is, once it is
recognized that significant medical decision making may confront a person with
new information that challenges his or her self-understanding and capacity to ex-
ercise self-direction, processes that assist the person to reexamine, when possible,
one's preferences, goals, values, and so on may better assist autonomous choice than
clear information provision alone. There may not always be time or opportunity to
realize this kind of support, yet when it is available to those who need it or want it,
it would support patients' autonomy. For example, recognition of what a particular
condition or treatment involves may be best understood by meeting with nonexpert
sufferers, survivors, and carers, and it might better enable the patients to put the in-
formation they receive together with their own self-understanding.
Assisting patients to make choices through active understanding of their wants
and expressed preferences may well better protect autonomy than detailing risk
232 Relational Autonomy in Context

probabilities, especially when the alternatives are limited and one or other of the al-
ternatives is clearly preferable in the circumstances. A full listing of all possible side
effects, for example, may well do nothing but agitate a patient who is clear that she
desires to be treated and understands that there are risks associated with achieving
that goal.54 The physician, standing as objective, neutral information source, is not
necessarily an aide to autonomy. A health-care worker who has sufficient informa-
tion; who can listen actively to the patients' identification of their concerns, desires,
fears, and so on; and who can ask them how much they want to know and why will
often better promote autonomy both in decision making and in the patient's capac-
ity to learn to accommodate or respond to the changes in their health, so they can
learn to live with, resist, or accommodate their altered circumstances.
By examining some feminist writings on bioethics and autonomy, I have shown
that we need to break away from an understanding of autonomy that focuses nar-
rowly on informed consent. In particular, 1 have extended both Sherwin's critique of
the dominant conception of autonomy in bioethics and Meyers's procedural ap-
proach to autonomy to show the implications for health-care ethics of a relational
account of autonomy modeled on Meyers's understanding of the development and
exercise of autonomy competency. It is worth repeating that 1 am not arguing for
doing away with some sort of improved consent process, of the kind that might be
developed from Sherwin's insights. Rather 1 am arguing that in bioethics it is im-
portant not to restrict the scope of respect for personal autonomy to the consent
process. The range of areas within health-care where the development, promotion,
and exercise of autonomy competency can be influenced by health-care institutions,
policies, and professionals must also be considered.
More work is needed in this area, however. At the practical level, the specifica-
tion of the kinds of practices and policies in health care that need reexamination in
light of the demands of autonomy must be elaborated in the context of existing
health-care systems and an articulation of the likely autonomy-affecting features of
different conditions and treatments. Finally, given the broader conception of auton-
omy articulated here, there is a need to reexamine the basis for assigning to health-
care workers responsibility for patients' autonomy. The kinds of responsibilities and
duties that individual health-care workers might have if this approach to autonomy
is adopted need to be closely evaluated. For example, should all responsibility for re-
specting and promoting personal autonomy in health care fall into the hands of in-
dividual health professionals, or should there be state and institutional responsibility
for some aspects of personal autonomy? Furthermore, there is the familiar issue of
the importance of autonomy promotion, protection, and respect relative to other
ethical considerations and legitimate demands on resources. Nonetheless, bioethi-
cists who wish to respect autonomy should ensure, among other things, that they
recognize autonomy in all its complexity.

Notes

1. This article has benefitted from comments from participants at conferences and sem-
inars in 1996 and 1997, as well as from discussions with Rebecca Albury, Anne Donchin,
 Choice and Control in Feminist Bioethics 233

Catriona Mackenzie, Katherine Morgan, Barbara Seeker, Susan Sherwin, and Natalie Stoljar,
to whom I ov/e thanks.
2. See Susan Brison's article, "Relational Autonomy and Freedom of Expression," in this
volume.
3. At least two, Susan Sherwin and Anne Donchin, have articulated relational concep-
tions of autonomy that they apply to health care: Susan Sherwin. "A Relational Approach to
Autonomy in Health-care," in The Politics of Women's Health: Exploring Agency and Autonomy,
The Feminist Health Care Ethics Research Nework, coordinator Susan Sherwin (Philadel-
phia: Temple University Press, 1998); Anne Donchin, "Understanding Autonomy Relation-
ally," Journal of Medicine and Philosophy 23, no. 4 (1998).
4. See, for example, Diana Meyers's discussion of this distinction in Self, Society, and
Personal Choice (New York: Columbia University Press, 1989), pp. 76-79, 101-102; see
also Harry Frankfurt, "Freedom of the Will and the Concept of a Person," The Journal of
Philosophy 68 (1971): 5-20; Gary Watson, Free Will (London: Oxford University Press,
1982); John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971).
5. Thomas L. Beauchamp and James Childress, Principles of BiomedicalEthics, 4th ed.
(New York: Oxford University Press, 1994).
6. W.D.Ross, The Right and the Go od (London: Oxford University Press, 1930).
7. Beauchamp and Childress, Principles of Biomedical Ethics, p. 121.
8. Ibid., pp. 190-91.
9. Ibid., p. 190.
10. Ibid., p. 327.
11. Ibid., pp. 14-26.
12. Sherwin, "Relational Approach to Autonomy"; Donchin, "Understanding Auton-
omy Relationally."
13. John Stuart Mill, On Liberty (1859); Joel Feinberg, "Legal Paternalism," Canadian
Journal of Philosophy 1 (1971).
14. Indeed, in a recent article on the role of families in informed consent, Mark
Kuczewski notes that it would not be much of an exaggeration to say that "all of medical
ethics is but a footnote to informed consent." Kuczewski, "Reconceiving the Family: The
Process of Consent in Medical Decision Making," Hastings Center Report, March-April
1996, pp. 30-37, 30.
15. Rosemarie Tong, "Feminist Approaches to Bioethics," in Feminism and Bioethics:
Beyond Reproduction, ed. Susan M. Wolf (New York: Oxford University Press, 1996), p. 74.
16. Ibid.
17. Laura Purdy, "Surrogate Mothering: Exploitation of Empowerment," Bioethics 3
(1989): 18-39.
18. Susan Sherwin, "Feminism and Bioethics," in Wolf, Feminism and Bioethics.
19. See, for example, Catriona Mackenzie's critique of bioethical and feminist ap-
proaches to the ethics of abortion, in which she queries the adequacy of these approaches be-
cause of their narrow understanding of pregnancy as an event in the lives of women over
which women lack control, in "Abortion and Embodiment," Australasian Journal of Philoso-
/.Aj/70 (1992): 136-155.
20. Tong, "Feminist Approaches to Bioethics," pp. 75-76.
21. Robyn Rowland, Living Laboratories (Sydney: Sun Books, 1992); Janice Raymond,
Women as Wombs: Reproductive Technologies and the Battle over Women's Freedom (New York:
HarperCollins, 1993).
22. Rowland, Living Laboratories; Raymond, Women as Wombs; Christine Overall, Ethics
and Human Reproduction: A Feminist Analysis (Boston: Unwin Hyman, 1987).
23. Mary Jeanne Larrabee, An Ethic of Care (New York: Routledge, 1993).
234 Relational Autonomy in Context

24. Nel Modelings, Caring: A Feminine Approach to Ethics and Moral Education (Berkeley:
University of California Press, 1994), p. 33.
25. See Susan Sherwins critique of Modelings in Sherwin, No Longer Patient (Philadel-
phia: Temple University Press, 1992), p. 47. See also Virginia Held, who focuses on the
mother-child relationship as the central ethical relationship, in Feminist Morality: Transform-
ing Culture, Society and Politics (Chicago: University of Chicago Press, 1993).
26. Joan Tronto, "Beyond Gender Difference to a Theory of Care," Signs, 12 (1987):
644-663; Jennifer Nedelsky, "Reconceiving Autonomy: Sources, Thoughts and Possibili-
ties," Yale Journal of Law and Feminism 1 (Spring 1989): 7-36.
27. Sherwin, No Longer Patient.
28. Sherwin, "Relational Approach to Autonomy."
29. Sherwin, "Feminism and Bioethics," p. 48.
30 Sherwin, "Relational Approach to Autonomy."
31. Genevieve Lloyd, The Man of Reason: "Male" and "Female" in Western Philosophy
(London: Methuen, 1984).
32. Sherwin, "Relational Approach to Autonomy."
33. Ibid.
34. Ibid.
35. Sandra Bartky, Femininity and Domination: Studies in the Phenomenology of Op-
pression (New York: Routledge, 1990), chap. 2.
36. Sherwin, "Relational Approach to Autonomy."
37. Jan Bowen, "A Dose of Their Own Medicine," Good Weekend, 6 December 1997,
pp. 30-34.
38. On the distinction between rational choice and autonomous choice, see Meyers,
Self, Society, and Personal Choice, pp. 77-80; on the principles of rationality and deliberative
rationality, see Rawls, Theory of Justice, pp. 410 — 423.
39. Meyers, Self, Society, and Personal Choice, p. 76.
40. Gerald Dworkin, The Theory and Practice of Autonomy (Cambridge: Cambridge
University Press, 1988); Robert B. Young, Personal Autonomy: Beyond Negative and Positive
Liberty (London: Groom Helm, 1986); Meyers, Self, Society, and Personal Choice.
41. In what follows I focus on Meyers's account because it draws on these earlier works
and develops them in the light of feminist and further moral-psychological concerns. Mey-
ers's account is particularly illuminating for developing an account of autonomy appropriate
for feminist bioethics. She focuses on the role of socialization in the development of auton-
omy competency, on the importance for autonomy of one's understanding of one's emo-
tions, and on the need to develop the capacity for self-definition.
42. Ibid., p. 76.
43. Ibid., p. 98.
44. Ibid., p. 48.
45. Ibid., 76.
46. Ibid., pp. 77-78.
47. Ibid., pp. 98-99.
48. Ibid., p. 170.
49. Ibid.
50. Ibid., pp. 195ff.
51. Jill A. Blakeslee, "Speaking Out: Untie the Elderly," American Journal of Nursing88
(1988): 833-834; Susan Dodds, "Exercising Restraint: Autonomy, Welfare and Elderly Pa-
tients," Journal of Medical Ethics 22 (1996): 160-163.
52. Susan Brison reflects on the immediate and ongoing effects of sexual assault ion a
 Choice and Control in Feminist Bioethics 235

person's self-understanding in "Outliving Oneself: Trauma, Memory, and Personal Identity,"

in Feminists Rethink the Self, ed. Diana T. Meyers (Boulder, Colo.: Westview, 1997).
53. The effect of medical crises need not all be damaging, however, as people may dis-
cover resilience they had not known themselves capable of, or they may come to reassess their
self-understanding and their goals in positive rather than self-damaging ways.
54. Bowen, "Dose of Their Own Medicine," pp. 32-33.
 10

AUTONOMY AND INTERDEPENDENCE

Quandaries in Genetic Decision Making1

Anne Donchin

Introduction
In 1996 a foundation funded by a British pharmaceutical company launched a proj-
ect to allay public fears about the new genetic detection techniques under develop-
ment in its industry.2 The focal point: of the project was a traveling road show, fea-
turing performances of a play called The Gift, which focused on a family affected by
an autosomal recessive genetic disorder (Friedreichs ataxia). The audience debate
that followed was structured and circumscribed by a staff facilitator, who offered a
preselected menu of options for resolving decisional quandaries dramatized by the
play's characters. Included were such questions as these: when such techniques be-
come feasible should they be available to individual parents, or should the state pre-
clude individual exercise of options directed to genetic enhancement? Suppressed
were questions involving the more subtle ambiguities suggested by the play's title
and a perplexing array of further issues implicit in the drama, for example, should
human choice supplant chance in the selection of genetic progeny? What stake does
the society have collectively in the uses to which such newly emerging technologies
are put? A complicated tangle of moral, social, and metaphysical issues was reduced
to a few simple options: either the state reaches out to regulate genetic technologies
or individuals are left free to select for their offspring whatever characteristics they
prefer, limited only by the options available in the commercial marketplace and their

236
 Autonomy and Interdependence 237

individual purchasing power. Implicit in the debate's structure was a predisposition

to leave individual family members free to make genetic decisions unencumbered by
constraints that would limit the effects of their action on others.
The agenda was tailored around a conception of autonomy that idealizes the
individual decision maker and abstracts options from both context and conse-
quences. Its intended message, apparently, was that the new genetics would be
problem-free if individuals would only utilize medical techniques to identify af-
fected embryos. Guiding alternative resolutions of the dilemma presented in this
well-meaning farce was a popular version of the individualistic conception of au-
tonomy entrenched in bioethical theory. This theoretical structure stifled all other
possible resolutions of the characters' dilemma and created the illusory expectation
that social problems endemic to the new genetics would vanish if only affected
embryos could be eliminated.
The project's deficiencies as a vehicle for public education helped crystallize my
own uneasiness about the reigning conception of personal autonomy that motivates
much health-care policy and practice. This conception bypasses concern about the
particular situations of agents as long as formal informed consent requirements can
be satisfied, ignores background conditions that shape options offered to patients,
and regards decision makers in isolation from significant relations with surrounding
others including health-care providers. My aim here is to show the need to reorder
bioethical theory so that it incorporates a conception of autonomy that stimulates
health-care professionals to respond appropriately to the social situation of patients
and their families, including disparities in their power relationships.

Weak and Strong Relational Autonomy

The dominant framework of North American bioethics is structured around a set of
cardinal moral principles, including autonomy, nonmaleficence, beneficence, and
justice. Since publication of The Belmont Report^ and Tom Beauchamp and James
Childress's first edition of Principles of Biomedical Ethics,^ this framework has dom-
inated bioethics-oriented teaching, consultation, and scholarship, as well as the pro-
nouncements of prestigious governmental ethics commissions. The prominence
given to patient autonomy within this framework has been evident even in federal
statutes such as the Patient Self-Determination Act?
In subsequent editions of their text, Beauchamp and Childress have sought to
adapt their version of this theoretical structure, now called "principlism," to address
critical objections directed primarily to their ranking and representation of the prin-
ciple of autonomy.6 They now stress the respect due autonomous agents rather than
an ideal image of fully autonomous selves. Respect, Childress explains in a related
article, is both an attitude and "an act of refraining from interfering with, or at-
tempting to interfere with the autonomous choices and actions of others, through
subjecting them to controlling influence . . . or manipulation of information."7 Its
implementation, he contends, entails autonomy-related rights such as liberty, pri-
vacy, confidentiality, and claims to positive assistance. It is compatible, though, with
individuals yielding decision-making power if they choose to subordinate themselves
to authorities like physicians or religious institutions.
238 Relational Autonomy in Context

Feminist criticisms of the conception that he and Beauchamp advanced in ear-

lier editions of their Principles pointed to its tendency to underestimate the impor-
tance of intimate and personal relationships and to rely on a misleading image of an
independent self.8 But in their most recent edition, Beauchamp and Childress re-
spond that theirs is not the "stark, individualistic conception of autonomy" that
feminists fault. 9 They offer what I call weak relational autonomy, for they do not deny
that communal life and human relationships provide the matrix for the develop-
ment of self. Values and beliefs shared with others may still be fully chosen as one's
own. However, their cautious acknowledgement that personal relations play a role in
self-development joins social tics only contingently to agents' self-understandings.
Bonds that link individuals can seemingly be detached at will. Individuals' beliefs
and values are assumed to be transparently self-evident and, barring coercive influ-
ences, voluntarily chosen. Their account is weakly relational because it does not
acknowledge that social relations are often at least partially constitutive of self-
conceptions or that social relations are often not voluntarily chosen by agents. I
argue later that in genetic contexts, many affected relations are biologically linked
and hence not voluntarily chosen. Moreover, Beauchamp and Childress overlook
the impact of the self's social situation in two respects. First, in their model, respect
for autonomy is compatible with a contract model of the physician-patient rela-
tionship; second, their model ignores macro-level (societal) influences on the auton-
omy of individual patients. I consider each of these problems in turn.
Despite Beauchamp and Childrcss's insistence that their conception is neu-
tral with respect to other theoretical perspectives, it fits most comfortably within
theoretical accounts that envisage physician and patient: as comparably situated
independent contractors. First, social position is largely irrelevant to transactions
between them. A physician-contractor can fully discharge her responsibilities by sat-
isfying the requisite formal conditions for securing informed consent, so there is no
requirement that she enter into an interpersonal relationship with her patient. 10 This
independent contractor model fails to capture significant dimensions of relations
between physician and patient, for sick people are vulnerable and dependent on the
skill and sensitivity of their caregivers. 1 ' Therefore, respecting such patients' auton-
omy typically requires attention to details of their life experience and surroundings.
Often patients do not fully recognize their own beliefs and values, so reaching an au-
tonomous decision about their care may require extended exploration of their his-
tories and motivational structures. Then, too, patients' self-understandings may be
so confused with others' perceptions of them (particularly in hospitals) that no de-
cision can be disentangled from their influence. Respecting autonomy would require
recognizing patients' struggles to break free of oppressive authoritative influences
and assisting them to sustain relationships essential to their self-identity and well-
being. In attending members of marginalized groups, providers would need to tran-
scend group stereotypes that jeopardise effective and equitable care without ignoring
group identities integral to patients' self-understanding.
Second, in Beauchamp and Childress's account, respect for autonomy is con-
structed around micro-level considerations and isolated both from macro-level, so-
cietal issues (which principlist accounts categorize under justice) and middle-level
 Autonomy and Interdependence 239

concerns. Susan Wolf rightly faults principlist schemes for making no structural pro-
vision for ethical problems that arise in relation to middle-level groupings such as
family, race, and gender.12 She points out that these shortcomings are looked on
largely as aberrations rather than problems endemic to the bioethical enterprise. To
illustrate: in a recent issue of a major medical journal, the authors of "Ethnicity and
Attitudes toward Patient Autonomy" surveyed members of several ethnic groups to
study attitudes toward end-of-life decision making. They found that some groups
were more likely to believe that the patient's family should be the primary decision
makers regarding end-of-life care. Whereas the Beauchamp and Childress account of
personal autonomy allows only for explicit individual delegation of authority, the
survey calls attention to an implicit group value structure that subordinates individ-
ual decision making. Harmonious family functioning was more central to the val-
ues of the research subjects than the autonomy in decision making of individual
members.13 In light of these findings, procedures for securing informed consent
that bypass value differences and rely on an identical set of formal standards are
morally suspect. Adequate respect for autonomy requires attention to the specifics of
physician-patient interactions, as well as to class, race, and other middle- and macro-
level influences that structure these relationships.
My own understanding of autonomy incorporates features of the weak rela-
tional conception insofar as it recognizes that social relations provide the ground for
the development of autonomy capacities and stresses the respect providers owe their
patients. But I move beyond this to a stronger conception that encompasses Jennifer
Nedelsky's central idea: that there is a social component built into the very meaning
of autonomy.14That is, the subject-centered activities of reflecting, planning, choos-
ing, and deciding that enter into self-determination are social activities in both a
subjective and an objective sense. Subjectively, material for reflection is built on the
foundation of a shared past and future expectations that involve others' participa-
tion. Objectively, the alternatives actually available for decision making depend on
background norms, practices, social structures, and institutions that configure and
limit options. As Martha Minow has noticed, it is true in varying degrees of all of us
that how we function depends on circumstances within our social and physical en-
vironment.15 In the balance of this section, I accordingly develop the components of
what I call strong relational autonomy.
Autonomy in the strong relational sense is both reciprocal and collaborative. It
is reciprocal in that it is not solely an individual enterprise but involves a dynamic
balance among interdependent people tied to overlapping projects. Moreover, the
self-determining self is continually remaking itself in response to relationships that
are seldom static. The experience of trauma, so common among hospitalized patients,
suggests that the self is affected, and perhaps partially constituted, by the changing
circumstances to which it is exposed. For, as Susan Brison explains, it is the loss of
connection experienced by trauma survivors that most imperils their autonomous
selfhood.161 share her conviction that the self exists fundamentally in relation to oth-
ers. Interconnections continue to shape and define us throughout our lifetime, so
that patterns through which we construct (and reconstruct) our self-identity and in-
fuse it with meanings are bound up with meanings given in the social world exter-
240 Relational Autonomy in Context
17
nal to us. The threat of serious illness or trauma both endangers these intercon-
nections and thwarts autonomous pursuits. The body one has trusted to pursue
one's plans and projects is shown to be vulnerable, fragile, and unprotected.
The other feature of the conception of strong relational autonomy that I advo-
cate is the collaborative component. An appropriate respect for personal autonomy
will require helping professionals to respond sensitively to the meanings illness has
for those in their care; to deploy their power and influence to restore and strengthen
autonomy competencies 18 ; and to support patients' struggles to create new personal
meanings out of the experience of disease, disorder, or disability. Any tenable con-
ception of personal autonomy is subject-centered; but a social conception that is
relational in this stronger sense will take into account the network of relation-
ships that bears on individual efforts to be self-determining, responsible agents. In
decision-making contexts prevalent in health-care settings, this conception encom-
passes providers of care, affected family members, and other individuals and groups
to whom a subject's identity is bound. The context that configures care, the ramifi-
cations of decisions for the life projects of patients and their families, and structures
of family power and authority that impinge on decision making are all integral to
one's capacity for autonomous self-determination.
My account is also feminist in several senses beyond the features of feminism to
which Beauchamp and Childless respond. First, it applies the resources of feminist
scholarship to a set of issues that have previously disregarded gender hierarchies.
Second, it envisages autonomy from within structures of power and authority rather
than assuming that pursuit of personal goals can be isolated from social aims and in-
volvements. Third, it is sensitive to structural inequities in the position of differently
situated women and members of social groups whose opportunities to shape their
lives in self-determining directions are often meager and inadequate. Thus it ad-
dresses the micro- and macro-level concerns they ignore. In these respects, my ac-
count is sensitive to Susan Sherwin's ambivalence about feminist proposals that pri-
oritize personal autonomy.19 I aim to steer a path between two risks: that individual
self-determination will be overwhelmed by more encompassing social aims and that
self-determination will be afforded so much importance that power imbalances are
allowed to persist under the guise of autonomous self-determination.

Empirical Support for a Strong Relational

Conception of Autonomy

Genetics and Biological Ties

The need for a strong relational conception of autonomy can be most readily shown
by examining practical problems that arise within families that are bound together
by biological connections. Genetic-related decision making, particularly, brings
clearly into focus underlying tensions between prevailing accounts of personal au-
tonomy dominant in bioethics (both individualistic and weakly relational ones) and
the interpersonal dependencies so prevalent in families with inherited genetic disor-
ders. Concerns easily overlooked in other medical contexts become more prominent
here, for genetic information carries more personal, familial, and social risks and
 Autonomy and Interdependence 241

burdens than virtually any other kind of medical information. Practitioners who
rely on a standard conception of autonomy may inadvertently exacerbate patients'
tensions. Policymakers accustomed to weighing only individual and societal interests
when considering the claims of third parties to genetic information may overlook
the bearing this information has on the autonomy of patients' families and racial
and ethnic groups to which they are linked.
I realize, however, that in focusing on situations in which social families are tied
together by biological connection, I risk being misread by feminists who view bio-
logical families as reactionary and repressive vestiges of patriarchal social arrange-
ments that should be supplanted by deliberately chosen voluntary relationships.
Surely, traditional families organized around biological connections do often replicate
and reproduce from other social domains hierarchical gender relations that reinforce
the continuing subjugation of women, sanction violence and abuse, and inhibit the
development of an autonomous sense of self.20 But contrary to the presumptions of
Beauchamp and Childress and some feminists who presume that social ties can be
severed at will, biological grounding inevitably influences social connection, so that
our most significant relations will often not actually be voluntary. Short of embarking
on radical social reform that severs biological from social connections,21 family poli-
tics could be reconfigured to rectify power imbalances so that the autonomy of all
family members is fully respected. Advancing this aim, however, requires recharacter-
izing personal autonomy in a way that acknowledges the influence of nonvoluntary
biological connections on personal identity and social relationships.
The host of issues surrounding newly developing genetic knowledge vividly il-
lustrates the extent to which a conception of family as a biologically linked social
unit is imbedded in social institutions, penetrates individual self-perceptions, and
perpetuates (often exacerbating) social injustices. For example, new genetic knowl-
edge imposes social and economic burdens on families that cannot be alleviated
without recognizing, at a social level, the injustice of the burdens and taking appro-
priate steps to remedy them. The acknowledgement of biological ties undermines
the conception of the individual implicit in both the individualistic and weakly re-
lational models, namely, the conception in which the individual has voluntary con-
trol over the network of social relations in which she is imbedded.
In the following sections, I argue that relational considerations are already im-
plicitly taken into account by medical practitioners who are working in genetic con-
texts, and I examine one example to develop the components of the strong relational
model. I first consider intrafamily problems that arise (1) when a family member
needs an organ or tissue donor, (2) when genetic knowledge intrudes on pregnancy
and child rearing, and (3) when a genetic disorder is detected in the bloodline. In
the final section I return to the gender-specific vulnerabilities of women.

Wresting a Relational Conception from

Prevailing Medical Practice
Transplantation and Tissue Donation. Contrary to much popular rhetoric and to
the assumptions presupposed in The Gift, genetic quandaries did not suddenly de-
scend on clinical practitioners like a bolt from the blue. Their appearance was ante-
242 Relational Autonomy in Context

dated by problems arising out of tissue-matching techniques used when a family

member needed an organ or tissue donor. New information emerged that frequently
threatened social bonds among biological kin. In responding to such situations,
practitioners improvised, evolving de facto norms. These reflected a double aware-
ness of family dynamics: tensions between the autonomy-advancing projects of in-
dividual family members and their relational bonds and a nascent realization that in-
dividuals' autonomy is relational. I cite several instances of individual and family
interests that butt up against one another, one that involves balancing family inter-
ests when children are too young to play a significant role in deliberations them-
selves and others that involve the deployment of sensitive family information.
The first instance concerns a minor child who risked dying from leukemia that
could be treated only by a bone marrow transplant. Only her mentally handicapped
sister was a good match. Since the parents had conflicting interests in the well-being
of both children, a judicial opinion was sought. The court in this case concluded
that the physical risk involved in surgery to retrieve the marrow from the handi-
capped sister was outweighed by the psychological benefits conferred on her. If the
sister with leukemia died, not only would her sibling be denied her companionship,
but the grief-stricken mother would find it difficult to provide the retarded sister
with the maternal attention she had grown accustomed to.22 Similar cases have come
before the courts in several countries. The courts have commonly used variants of
this balancing strategy, weighing actual physical burdens against hypothetical psy-
chological benefits to the donor. Where the children have had a social, as well as bi-
ological, relationship, the court has almost always ruled, in favor of donation. So the
healthy child has been compelled to undergo a nontherapeutic surgical procedure
for the sick sibling's benefit.
Implicit in such rulings is a relational ethic that obliges family members to
subordinate their individual interests to others' well-being. But there are difficulties
in a relational approach that overemphasizes the relations of care and connection
between siblings. Thus such rulings satisfy neither individualistic nor relational re-
quirements for autonomy, although for different reasons. Those who lean toward
an individualistic discourse fault the courts for not adhering more singlemindedly
to a decision-making standard based on the donor's best interests. Instead, the cal-
culus of burdens and benefits seems skewed, sacrificing the retarded sister's interests
to her "normal" sibling's. Were she a competent, autonomous decision maker her-
self, she could deliberate with other family members, reflect on the value of her re-
lationship to them, and voluntarily form her own judgment. Those who favor a
more relational discourse wonder why courts are so reluctant to recognize that the
moral relevance of family connection is not reducible to the interests of discrete in-
dividuals. In actual families, children have responsibilities to one another long be-
fore they develop capacities to avow them voluntarily. Preparation for autonomous
adulthood includes experience in caring for others one cares about and coming to
recognize one's own needs by responding to theirs. Adults who lack such grounding
are ill-fitted to build enduring voluntary relationships with others on a foundation
of mutual reciprocity.23
However, the individualistic discourse prevalent in these cases affords protec-
tion to the child asked to make the sacrifice that may be lost in a relational account
 Autonomy and Interdependence 243

that is inattentive to power relations among family members. Children need protec-
tion lest they be treated, Cinderella-like, as mere conveniences to advance others'
ends. So even if courts were to set aside the individualistic standard of best interests
and adopt a relational model, justice would not be served unless they set constraints
on moral sacrifices family members can be expected to make for one another.
An implicit relational approach is even more obvious in two further kinds of
situations: in cases in which a family member's blood is tested to determine compat-
ibility for organ or tissue transplantation and in circumstances in which a family is
being evaluated to find the most likely transplant donor. In such instances, practi-
tioners may feel obligated to act deceptively, selectively withholding information
from a patient who needs bone marrow, a kidney, or a slice of liver. For example,
they might withhold from their patient knowledge that a tested family member is a
good match but refuses the surgery required to retrieve the tissue. Knowledge of the
refusal, these practitioners claim, would have a disruptive effect on family relations
at a time when they are already under stress. This practice is sometimes criticized as
a kind of quasi-paternalism that elevates the interests of family members who refuse
to contribute over the interests of the patient, which should, in the critics' judg-
ment, be paramount.
Practitioners may also act on relational considerations when linkage studies are
done to assess risk for a genetic condition. Blood studies may reveal that the social
father is not the biological father, as family members believed. When a relationship
between the parents is ongoing, the practitioner faces a burdensome quandary. Who
should be informed? Is there any general rule that might provide guidance? Should
unwelcome information be imposed on people? When this information comes to
light in the process of screening for a suitable organ or tissue donor, the medical unit
might establish a de facto policy to disclose only information pertinent to the search
for a compatible donor. If a family member has already told the transplant team that
he will contribute a blood sample to provide a more complete family profile but is
unwilling to donate, the sample would be used only to establish linkage and then
would be set aside in the continuing search for a suitable donor. But when family
linkage studies are undertaken to determine carrier status for a genetic condition,
this dodge will not work. So many idiosyncratic aspects of the family situation may
have a bearing on the consequences of disclosure that it is difficult to envision a sus-
tainable general rule—apart from adopting truth telling as an absolute requirement.
However, all but the most committed deontologist would contend that when inter-
vening in family relations, practitioners should aim to minimize foreseeable adverse
consequences.24

Pregnancy and Child Rearing. The need to frame autonomy relationally, if it is to

be respected at all, becomes starkly evident when a women who is genetically at risk
becomes pregnant. A spouse may never have considered the full impact of marriage
to the carrier of a dominant genetic condition until the couple considers children.
Dominant carriers are bound to know about affected people within their own fam-
ilies, but recessive carriers may have no advance knowledge of the condition. Often
a woman is already pregnant before genetic testing for carrier status is recommended
to her. But by then, her options are already diminished. Pregnancy imposes pressures
244 Relational Autonomy in Context

on decision making that would not exist under less constraining conditions. If the
couple is willing to consider termination, they may have to decide very quickly. Sel-
dom do they know much about the particular genetic condition, prognosis for the
child, social support services, or how their own lives will be affected by a disabled
child. Test results compound these uncertainties since their degree of accuracy
varies. So suddenly a happy, untroubled pregnancy is transformed into a stressful ex-
perience, upsetting everyone's prior expectations.
So determined are many parents to learn the genetic susceptibilities of their
children that when a genetic disorder is not identified prenatally or evident at birth,
they often request that their child be tested. Here providers have squarely con-
fronted their obligation to consider the impact of their intervention on family rela-
tions. Several medical associations have established policies that limit parents' power
to demand that their child be tested. To safeguard the children's future autonomy
and protect confidentiality, it is recommended that practitioners refuse to test chil-
dren for genetic disorders unless there is a therapy available to treat them or a med-
ical benefit to be derived from monitoring their condition. Some parents object,
charging "paternalistic intervention" that infringes on their own autonomy. Al-
though many parents are reluctant to discuss their own risk with their children, they
are often so eager to learn their child's risk that they pressure practitioners to make
exceptions to the policy against testing children.25 But when testing affords no ther-
apeutic benefit to the child, it is difficult to build a case that parents' wishes should
prevail.
I turn now to examine more fully quandaries that arise when a genetic anomaly
is detected in the bloodline. 1 show why no simple formula can solve the problems
the new genetics thrust before us or shortcut the task of rethinking and renegotiat-
ing our relation to this new knowledge. Only a strong relational understanding of
autonomy will do here.

Disorders Detected in the Bloodline. A third example of a relational approach im-

plicit in genetic practice is offered by cases in which a genetic disorder is detected in
the bloodline. Trained and experienced counselors will tell people who consider test-
ing for a severe dominant disorder that regardless of the results, their futures will be
changed. The news will not be all good or all bad for anyone.26 Learning that one
belongs to an affected family is bound to have a profound impact on all family
members, both individually and collectively. The circumstances of testing magnify
this effect. First, for many genetic conditions the judgment that one is at high risk
involves pooling information among those who share the same bloodline. So the
medical risks will be compounded by psychological, familial, and social complexi-
ties. An individual who resists testing may not be able to avoid knowing if she is at
risk when other family members discuss their test results. Second, when a direct
gene test is available to the individual who is seeking diagnosis, the results may still
have striking implications for the future of others, so much so that even those who
test negative for the disorder may need to reconstruct their autobiographies and re-
configure their family relationships. Some, for instance, may feel a heightened sense
of family responsibility, believing (falsely) that their negative result will increase the
likelihood that other family members will test positive. Others may have trouble rec-
 Autonomy and Interdependence 245

onciling ambivalent feelings for their family members with this new information.
Thus any decision an individual makes about testing will penetrate both the biolog-
ical and the social family.
Individuals within the same family may use very different strategies to cope
with their sense of powerlessness in the face of a wholly random event. Some deny
the condition; others undertake a compulsive search for knowledge.27 Following de-
finitive testing, the more fortunate members may experience "survivor guilt." The
unfortunate ones may blame parents and hold them accountable for their disability.
If the disorder is X-linked, a mother may bear the brunt of resentment since it has
almost certainly been transmitted through her. Parents who undertake prenatal di-
agnosis and already have a similarly affected child sometimes feel deeply ambivalent
about aborting an affected fetus, fearing that their living child may later interpret
the abortion as a personal rejection. Brothers, sisters, and parents will all be drawn
into a tangled web, laden with tension and conflict. The strain will extend to others
in intimate committed relationships with those at risk, thus threatening relation-
ships beyond the biological family as well.28
Test results may disrupt family expectations, impelling members to reshape
their self-identities and renegotiate family positioning. Then "family scripts," preex-
isting systems of beliefs about family relations, are shattered. The self-conception
one has comfortably developed within a given constellation of relationships is desta-
bilized. All family members are compelled to rethink their personal histories and re-
cast their stories about how they want to live their lives.29 Although family scripts in-
fluence everyone's self-development, in families with inherited genetic disease they
may be far more difficult to renegotiate. Such families sometimes cope with uncer-
tainties about which member might inherit a familial condition by preselecting one
they believe is affected. Then they interpret rhe behavior of that individual in ways
that fulfill their prophesy. The individual who subsequently tests negative for the
condition may be even more severely ostracized by the family than if the result had
been positive.30
The vagaries of insurance and employment practices will intensify psychologi-
cal stresses. Unless medical confidentiality can be assured, even a family member
who has not been tested herself might lose her job or insurance if the test results of
others in the family leak out. In one case, a woman in her fifties requested a direct
gene test for Huntington disease (HD). Her son and daughter-in-law protested,
claiming they should have a voice in the woman's decision, too. According to stan-
dard accounts of autonomy, their interference would surely be unwelcome. But the
son had good reason to fear the effects of the test. His employer had been seeking
just this kind of information. If his mother tested positive, his own job would be
threatened.31
Among genetic conditions, HD intrudes particularly forcefully into the lives of
individuals and families and imposes complex demands on those who counsel af-
fected families. When families are reluctant to inform other kin who may be at risk,
practitioners may be tempted to intervene by passing on this information. A curi-
ous spin on this theme is illustrated by the case of a twenty-nine-year-old woman
who came to the genetics clinic for counseling because her mother had told her that
she (the mother) had tested positive for HD. The mother, the counselor learned,
246 Relational Autonomy in Context

had a long psychiatric history, involving drug and alcohol abuse. The daughter was
an only child whose father left home early in her life. The mother and stepfather
both abused her. The counselor gained access to the mother's records and found that
she had actually tested negative for HD. But the mother, when asked, refused to let
her tell the daughter that she did not have the disease.
Within a principlist biocthics framework, a practitioner's moral responsibilities
would be articulated in terms of two guiding principles: respect for the daughter's
right to make autonomous decisions and for the mother's right to confidentiality.
Geneticists and trained counselors who regard confidentiality as sacrosanct would
disregard the daughter's interests rather than breach confidential information. The
tested individual, they often argue, has a stronger claim to the information than
other family members, so she ought not to be compelled to disclose the informa-
tion. 32 In this view the counselor who overrides the mother's refusal is guilty of
"clinical imperialism." 33 Philosophers and legal theorists are more likely to resolve
the dilemma by stressing the importance of accurate information for the daughter's
future plans. They would argue that the geneticist is obligated to share her knowl-
edge of the deception so that the daughter can free herself from the influence of her
overcontroling mother. For those who follow this line of reasoning, the daughter's
claim to information so vital to autonomous pursuit of her own goals would trump
the mother's claim to confidentiality.
But as long as the daughter is deprived of this information, she has no com-
pelling reason either to strengthen or sever her relational ties. Unlike the clear-cut
personal histories Marilyn Friedman appeals to in her account of relational auton-
omy,34 the relevance of a genetically burdened relationship to personal autonomy re-
mains ambiguous in the absence of pertinent information. This case of misinfor-
mation illustrates that reflection alone may not be an adequate guide for an agent
who is seeking to expand autonomy that relates to her personal relationships. Even
in the absence of confounding misinformation, an adequate account of relational
autonomy would still need to provide criteria for distinguishing between autonomy-
fostering and autonomy-inhibiting aspects of social relationships, fn this regard, any
account that preserves even minimally Kantian features will recognize that whar we
want for ourselves may not enhance our autonomy if it can be attained only by
dodging responsibilities toward others who depend on us. Respect for their interests
and their autonomy may require us to relationali/e our own autonomy in the course
of advancing our plans and goals.35 Indeed, as a general rule, following one's own
sense of the good life usually binds one to schemes of social cooperation over which
one may have little direct influence or control. Some of these may be only instru-
mentally relevant to one's own pursuits, but others will inevitably be substantively
bound to one's projects, so that what one desires for oneself is inextricably inter-
twined with others' desires. Thus, outright disregard of other morally significant val-
ues is possible only in certain very limited contexts.
In all these instances, practitioners seem to be operating from not an individu-
alistic conception of autonomy but an implicitly relational model that appraises
family interests collectively. In so doing, they expose two fissures in the dominant
structure. First, they point to specific inadequacies in a framework that operates
from a noncontextual set of primary principles that may have significantly different
 Autonomy and Interdependence 247

implications as contexts vary. Second, they address the possibility (not always the
case) that disclosure will exacerbate tensions rather than relieve them. Here, too, like
the cases of incompetent donors, further protections are needed to ensure equitable
treatment for all family members affected by a medical intervention. Respect for au-
tonomy would seem to require respect for both communal matrices and individual
decision-making capacities.
Thus there is good reason not to abandon a rule-based approach altogether be-
cause the individual who is most vulnerable and has the most riding on the decision
deserves the protection that an adequately formulated principle of autonomy can af-
ford. Competing claims of other family members ought not to be allowed to trump
hers. Nor should her interests be reduced to the kind of balancing act we saw at
work in the judge's deliberations about the incompetent transplant donor. The pro-
tection of young children's future interests calls for an approach that gives less
weight to the immediate concerns of parents and more to the children's futures, un-
derstood broadly to encompass not only physical well-being but also capacities for
autonomous self-determination.

An Example of Reciprocal and Collaborative

Components of Relational Autonomy
In this section, I return to the example of the daughter who is seeking genetic infor-
mation to articulate in detail my conception of strong relational autonomy and to
show how collaborative and reciprocal components are often intertwined. Since au-
tonomy is not solely an individual enterprise and the self-determining self is con-
tinually remaking itself in response to relationships that are seldom static, personal
autonomy is not incompatible with adapting one's personal projects to the needs or
preferences of others as long as no one's autonomy is trampled on. Respect for every-
one's autonomy, however, requires long-term reciprocity and the equitable balancing
of power relations. These conditions limit an individual's freedom to opt out of a re-
lationship, but they might possibly require opting out if they cannot be met even
minimally. Mindful here of the tension Friedman alludes to, I focus on several fea-
tures of the relationship between the daughter who is seeking information pertinent
to her future plans and the mother who is withholding that information. I also con-
sider the counselor's responsibilities insofar as they bear on her professional relation-
ship to mother and daughter.
Insofar as the counselor is obliged to respect the autonomy of both parties, she
must decide how to manage the information the mother has withheld that is perti-
nent to the daughter's decision making. If the counselor is not, herself, to contribute
to severing their ties, she will explore possible resolutions to the dilemma. To bring
this about, though, she needs a decision strategy that is more open to the complexity
of genetic knowledge than the language of competing claims, one that recognizes
the interlocking relevance of genetic knowledge for the entire biological family. In
this one significant respect, at least, mother and daughter are not free to sever their
relationship.
It is imporrant to note that a rigid respect for personal privacy would inevitably
be confounded in such situations. The awareness that genetic information is shared
248 Relational Autonomy in Context

among kin is implicit in both the mother's deception and the daughter's quest for
further information about her own risk. Both know that their fates are interlinked.
If the mother is affected, the daughter has a 50% risk of acquiring the disease too.
If the mother docs not have the gene, the daughter's risk is reduced to nil (assuming
the father is unaffected). Thus a determination of moral responsibility that slights
the daughter's request for knowledge about her own susceptibility exaggerates the
mother's control over the information. It thereby impedes the daughter's capacity to
shape her own future autonomously and disregards the harms a family member in a
position of power can exert over vulnerable others.
Nonrelational accounts of autonomy may also distort the actual dynamics of
family relationships in a further respect, for an analysis focused on competing claims
presupposes that people arc normally at odds with one another. On some occasions,
they surely are, but a discourse framed around individual rights and interests is
bound to neglect other significant features of family bonds. Shifting attention to the
relational situation that frames the conflict opens possible routes to further strategies
for resolving tensions, reducing injustices, and reconfiguring self-understandings.
Mother and daughter could possibly be reconciled on a new level informed by the
backward- and forward-looking meanings that: this new knowledge has for their
conception of self and other. And if reconciliation is not possible on terms accept-
able to both, the counselor's cmpathetic support can nonetheless contribute to inte-
gration of this new knowledge into their individual self-understandings.
Pertinent to these observations is Susan Brison's evocative account of her own
experience as a trauma survivor, for it reveals the deeply social nature of the self
and the limits of an individual's capacity to control her own self-definition. i 6 To re-
constitute a unified self out of the disintegration that followed violent assault, she
needed empathetic and responsive listeners. Disclosure of vulnerability to genetic
disease can precipitate a similar crisis. In the face of a threat that one can neither
fight nor flee, which damages one's trust in the integrity of mind and body, connec-
tions between self and others arc needed to combat defensclessness in the face of
life-threatening forces and anxiety about a foreshortened future. 37
Collaboration among family members is necessary to preserve autonomy in
families plagued by disabling genetic disorders. Connections to family and other
intimates, Brison points out, provide not only nurturing support but also a reposi-
tory for collective memories and rituals that sustain the sense of a unitary self able
to persist through crises. Family narratives about those who coped courageously
with the threat of disease and withstood its withering assaults not only serve many
of the functions of support groups but also intensify other family members' sense
of theit reciprocal interdependence. Professional counselors encourage affected
families to seek out groups that can offer sutvival strategies, but recognition of the
family's centrality in sustaining the sense of bodily integrity of the affected member
is less common.
For example, John Hardwig's concerns about how prevalent norms of medical
decision making tend to slight family interests convey a sense of family as a group
of disparate individuals with a common pool of material resources and perhaps re-
ciprocal obligations to provide care in times of need. When a family will be affected
by a decision, fairness requires, he insists, that the family have a role in reaching it.
 Autonomy and Interdependence 249

In some situations, he believes, the interests of other family members might even
outweigh the patient's own.38 His is an upsetting and radical claim, say Hilde and
James Nelson, both because it runs counter to the priority given to individual au-
tonomy and because it ignores the fact that the patient has a greater stake in the de-
cision than anyone else. Certainly, family should have input to deliberation, but
family motives and concerns, they stress, are better aired in the open, where the in-
tricate network of their relationships can be protected from breakdown. Practition-
ers who cast people in a conflictual relationship increase the likelihood that they
will, in fact, see one another as adversaries.39
The disagreement between the Nelsons and Hardwig runs deeper than the
decision-making scenario on which it is focused. Hardwig's formulation of the issue
takes family conflict as static, a function of each individual's separate material inter-
ests, which are unaffected by the physician's intervention. Insofar as the Nelsons' ap-
proach holds providers responsible for fostering family communication and seeking
reorientation of the perspectives of interested parties, it is founded on an under-
standing of the family as a dynamic, interlocking structure that is not reducible to
its separate parts. Of the two, only the Nelsons' perspective is compatible with both
the collaborative and reciprocal senses of relational autonomy— collaborative inso-
far as family members recognize their dependence on one another for the achieve-
ment of personal plans and projects, and reciprocal in that recollections of their
shared history, overlapping memories, and rituals are directed to maintaining and
restoring the psychic integrity of their individual members.

Genetics and Gender

Any contemporary account of personal autonomy ought to be responsive to femi-
nist ambivalence about autonomy as a personal ideal that many still associate with
a distinctively masculine ideal of self-development. Relational accounts typically
aim for a transvaluation that frees the conception from such associations. But unless
that account is responsive to structural inequalities in the position of differently sit-
uated women, it risks reinstituting some of the most reprehensible features of the
masculine ideal, overvaluing individual self-determination so grossly that power im-
balances among the more and less advantageously situated are disregarded. Thus an
adequate relational account needs to connect personal self-development with more
inclusive social aims, emphasizing women's particular vulnerabilities to unjust treat-
ment, namely, their reproductive capacities and social positioning within families. In
health-care settings, providers and counselors need to be attentive to such vulnera-
bilities as they intervene in the deliberations of a family member under treatment
and her kin, who share genetic characteristics and participate in a common social
network. Providers, I argue, need to consider not only particular circumstances but
also rule-based principles if they are to guard against unjust treatment of family
members who, by virtue of gender or social position, bear a disproportionate share
of family responsibility. Family interests are doubly implicated when there is genetic
disability, both because of the care-giving and support functions likely to fall pre-
dominantly on women and because of the vulnerability of other family members
who are potential victims of the same biological incapacity.40 First, I note the major
250 Relational Autonomy in Context

respects in which women's lives are likely to be more disturbed by new genetic
knowledge than men's. Second, I identify the responsibilities this vulnerability
imposes on providers who disseminate this knowledge and advise women and
their families of their options. Once again, I emphasize how genetic knowledge can
thwart or advance autonomy, depending on how it is imbedded in a clinical rela-
tionship.
Genetic information enters into family relations and affects family life along
gender-specific paths. It is primarily women who are the central focus of the family-
based problems that genetic diseases create/'1 By virtue of both reproductive capac-
ity and social positioning, the role of family caregivers has traditionally fallen to
women. Clinical arrangements tend to reinforce this delegation of family responsi-
bilities. How burdensome such duties will be for any particular woman, of course,
will depend on her material circumstances arid support networks.42 Women bear
more of the responsibility for monitoring health issues in families than men. They
are the "kin-keepers" who do the family's genetic housekeeping. They tend to know
more about the obstetrical history of the man's family than he—how many preg-
nancies his mother had, how many ended in miscarriage, and so on. Since only
about one in four women is accompanied to a counseling session by a spouse or
partner, prudent women may have no choice but to arm themselves with all the in-
formation they can muster.
By virtue of reproductive capacity, too, genetic knowledge is likely to influence
women's lives far more pervasively than men's. First, only women undergo prenatal
testing, which is increasingly becoming a part of the customary prenatal care pack-
age. Not only women over thirty-five, but also younger women are increasingly
being urged to submit to preliminary genetic screening procedures and follow-up di-
agnosis, even when there is no prior knowledge of genetic disorders in their family
medical histories. Prenatal diagnosis is already available for hundreds of condi-
tions. 43 As DNA technology advances, laboratory procedures are automated, and a
skin prick replaces more intrusive procedures, far more screening tests are likely to
be included in routine prenatal care, further complicating prenatal decision making.
Second, women endure more of the psychological stress generated by testing
than male family members. Since therapies to treat genetic impairments following
prenatal diagnosis are virtually nonexistent, a woman whose fetus tests positive for
a genetic abnormality must assume primary responsibility for a difficult decision
about terminating a wanted pregnancy, often on the basis of incomplete informa-
tion. For their partners, prenatal diagnosis is often a less immediate worry, more eas-
ily postponed or denied.44 Genetic "therapies" now under development might even-
tually provide an alternative to abortion following diagnosis of fetal anomaly, but
these interventions will also be accomplished through the bodies of women. They
carry their own risks, which the pregnant woman will need to weigh in the light of
her own values arid priorities.
Preimplantation diagnostic techniques (PID) further complicate decision-
making options. This procedure (utilized in The Gift) minimizes recourse to abor-
tion as a way to avoid the birth of a genetically impaired child. Here too, however,
the mother's body must be invaded to facilitate laboratory inspection and selection
of embryos.45 No longer a futuristic technique, PID is already being marketed by di-
 Autonomy and Interdependence 251
46
agnostic laboratories for a fast-growing array of single gene disorders. Only ade-
quately informed women can evaluate whether the risk, disruption, and anxiety en-
tailed by PID are worth it to them.
Many genetic anomalies are not discovered until after birth. Since women are
the primary caregivers for infants (as well as for disabled people of all ages), they are
often the first to notice that their infant is not developing properly, and so they seek
out medical advice, battling their way through a health-care system that too often
depreciates a mother's observations. Most general practitioners have little or no
training in medical genetics. So a concerned mother may need exceptional tenacity
and determination to surmount disparaging misinformation ("your baby is just lazy
or placid" or "you're a neurotic mother") and reach a consultant qualified to offer an
informed diagnosis and the information and support needed to deal effectively with
her child's problems.47
Compounding the disproportionate burdens imposed on women by virtue of
their reproductive capacities are many gender-specific structural arrangements only
remotely related to biological differences. Poor women are particularly subject to
choice-minimizing health-care policies and practices. Women of color and other
underserved populations are doubly disadvantaged and less likely to find their way
to trained genetic counselors.48 For those who do, cultural differences may indirectly
hinder the communication and education process.
Class-based differences add a further layer of complexity to the inevitable dis-
parity between lay knowledge about inheritance and scientific accounts.49 Lay ac-
counts are often tied to cultural and psychological factors that resist scientific ex-
planation. Words like "uncertainty" and "risk" tend to have different meanings for
geneticists and lay people. For the professional, risk may mean the probability of a
negative outcome; for the patient, it is more likely to connote the severity of the
outcome.50 If such differences are too profound, barriers to a decision responsive to
the client's life circumstances and personal goals may be insurmountable.
A woman's ability to make an autonomous decision is determined not only by
the amount of information offered but also by the manner in which it is communi-
cated to her. Some providers are more directive in their approaches than others.51
Obstetricians often offer advice about pregnancy termination rather than options
for the patient's determination.52 Geneticists and genetic counselors are schooled in
an ethos of nondirectiveness that pulls in the opposite direction. The counselor is
expected to provide the facts a counselee needs to make an autonomous, informed,
and rational decision. These facts are to be balanced and given in a nondirective,
value-free neutral manner that respects the recipient's "beliefs, cultural tradition, in-
clination, circumstances, and feelings."53 Devised to protect patients from eugenicist
bias or other paternalistic impositions of the counselor's sense of rhe other's good,
the ideology of nondirectiveness fails to recognize, however, that the context that
frames counseling already presupposes the importance of detecting genetic anom-
alies and interceding.54 An overly naive commitment to nondirectiveness overlooks
at least four problematic moral dimensions of the relationship.
First, a nonselective barrage of information may overwhelm and paralyze a pa-
tient. Second, because nondirectiveness promises a degree of neutrality that no pro-
vider can possibly fulfill, patients may be caught off guard as they are guided toward
252 Relational Autonomy in Context

a course of action that defeats the primary intention of counseling: to enhance au-
tonomous decision making. Third, judgments about which conditions merit dis-
cussion with clients are often value-laden. 5 5 Unless counselors assume responsibil-
ity for the selectivity they exercise, they will inadvertently press their own values,
thereby jeopardizing the patient's effective reflection and deliberation. Fourth, inso-
far as counseling is committed to a traditional scientific model that bifurcates facts
and values, emotional dimensions of the client-counselor relationship will be sup-
pressed, isolating the client from the kinds of empathetic support needed to restore
a fractured identity or preserve her sense of wholeness in the face of trauma and
family conflict. Thus, the illusion that counseling is value-free conceals important
moral dimensions of the interaction that if recognized and addressed directly, would
clarify everyone's respective goals and empower the client to reorder her own moral
priorities and assign them sufficient weight to counterbalance the competing inter-
ests of other family members.
In the interest of promoting women's autonomy capacities, then, the nondirec-
tive paradigm should give way to a nonhierarchically structured model of genetic
counseling that fosters a collaborative relationship and enhances a client's capacities
to exercise agency within her own relational context. 56 The client might then make
her own determination about the centrality of family relations to her future goals.
Ideally, the counselor would not see her role terminating with the decision to abort
or continue a pregnancy but would strive to create a basis for a sustained relationship
with the client and her family that extends beyond the immediate crisis. Such revi-
sioning would advance genetic counseling beyond, the prevailing nondircctive model
toward a more holistic understanding of family dynamics imbedded in a relational
understanding of autonomy that incorporates both its collaborative and reciprocal
components.

Conclusion
The individualistic model of autonomy so dominant in medical practice hampers
effective medical decision making, gives insufficient attention to the impact of de-
cisions on family members, undermines the agency of those who are excluded from
the decision-making process, and often imposes unjust burdens on those affected by
others' decisions. Uncritical advocacy of this model by bioethicists fosters what
Susan Wolf calls a "bioethics for the privileged," serving the interests of those
who have ready access to health-care and an understanding physician who respects
them. 57 Medical practitioners who recognize limitations of the dominant model
sometimes modify their procedures to draw into discussion other family members
who are centrally affected by a decision. But these efforts tend to be fragmentary and
ad hoc, often inadvertently polarizing family conflict and jeopardizing the well-
being and autonomy of those most centrally affected. They are rarely supported by
institutional structures, which are increasingly being centralized. Without major re-
configuration of structural patterns, few will benefit from rearrangement of indi-
vidual provider-patient relationships. In the context of clinical medicine and prac-
tice, then, personal autonomy needs to be understood as involving self-determining
individuals who arc positioned relationally, both to their families and intimates and
 Autonomy and Interdependence 253

to the practitioners who care for them. Two major reorientations are critical in im-
plementing such a model.
First, all parties at risk must be adequately and appropriately identified as indi-
viduals within the institutional contexts that circumscribe their options. Such a shift
in perspective should bring into focus the disparate distribution of benefits and bur-
dens among family members. Decision-making strategies that attempt only to bal-
ance their individual interests risk silencing the voices of some who are immediately
affected by the action at issue. Such persons need the principled protection of a care-
fully nuanced recognition of relational autonomy and an unswerving commitment
to honor it.
Second, this acknowledgement needs to be implemented in ways that recog-
nize and respond to injustices in personal relationships, whether they are central or
tangential to institutional structures. Strategies that provide occasion for all family
members to articulate their own perspectives and air their disagreements before any
decision is imposed can help maximize their autonomy, both separately and collec-
tively. But to avoid reinscribing injustice in a new guise, a relational model must also
take full account of relevant ways in which individuals and families are differently
situated in relation to one another and to health-care resources. Its effectiveness
would depend on the sensitivity of providers to family dynamics, recognition of sit-
uations in which the voices of the less powerful members are silenced, and inter-
vention to facilitate their active participation in decision-making processes.
Implementation of such a relational model would impose increased demands
on health-care systems, which, belatedly casting off their paternalistic legacy, still
recognize individual autonomy only minimally. To aim for less, though, would be to
acquiesce in social arrangements that perpetuate disparate respect for the privileged
and the vulnerable.

Notes

1. I wish to express my gratitude to all those who contributed to my education in clin-

ical genetics: Indiana University Study in a Second Discipline Program, which enabled me to
devote an entire semester to the study of clinical genetics, and particularly to Dr. David
Weaver's patient mentoring. Thanks are due, too, to Dr. Ellen Solomon and her colleagues
in the Division of Medical and Molecular Genetics at Guy's Hospital, University of London,
who so generously gave their time to me so that I could compare the delivery of genetic serv-
ices in the United States and in Britain. My gratitude also to Catriona Mackenzie and Natalie
Stoljar for their careful and insightful comments on successive drafts of this article; to Helen
Bequaert Holmes, whose deep understanding of genetic issues enriched my own; and to Ed-
mund R Byrne, whose comments on previous drafts helped me to say what 1 meant to say.
Remaining inadequacies are, of course, my own doing.
2. The project was sponsored by the Wellcome Centre for Medical Science, an initiative
of Britain's largest medical research charity, the Wellcome Trust. The play was written by
Nicola Baldwin. Further info filiation is available from the Wellcome Centre, 210 Euston
Road, London NW1 2BE.
3. National Commission for the Protection of Human Subjects of Biomedical and Be-
havioral Research, The Belmont Report: Ethical Principles and Guidelines for the Protection of
 254 Relational Autonomy in Context

Human Subjects in Research, DREW Publication no. (OS) 78-0012 (Washington, D.C.: U.S.
Government Printing Office, 1978).
4. T. Beauchamp and J. Childress, Principles of'Biomedical Ethics (New York: Oxford
University Press, 1979). Subsequent references refer to the fourth edition, published in 1994.
5. U.S. Congress, Patient Self-Determination Act. Omnibus Budget Reconciliation Act
(1990), Pub I, no.101-508: 4206, 4751.
6. For a classic critique of principlism, sec K. D. Clouser and B. Gert, "A Critique of
Principlism," Journal of Medicine and Philosophy 15 (1990): 219-237. For more extended
discussion, see the series of articles in the Kennedy Institute of Ethics journal^,, no. 3 (1995).
7. James F. Childress, "The Place of Autonomy in Bioethics," Hastings Center Report 20
(1990): 13.
8. Beauchamp and Childress themselves cite Susan Sherwin's No Longer Patient: Femi-
nist Ethics and Health Care (Philadelphia: Temple University Press, 1992), p. 138. For Sher-
win's more recent reflections on principlist conceptions of autonomy, see "Feminism and
Bioethics," in Feminism and Bioethics: Beyond Reproduction, ed. Susan M. Wolf (New York:
Oxford University Press, 1996), pp. 47-66. Also relevant is Wolf's introduction, especially
pp. 15-16.
9. Beauchamp and Childress, Principles of Biomedical Ethics, "pp. 124-125.
10. For a more comprehensive criticism of the inadequacies of contractual accounts of
autonomy in health-care contexts, see Lucy M. Candib, Medicine and the Family (New York:
Basic Books, 1995). Candib draws on both Sherwin's earlier work, No Longer Patient, and
Virginia Held's critique in "Non-Contractual Society," in Science, Morality and Feminist The-
ory, ed. Marsha Hanen and Kai Nielsen (Calgary: University of Calgary Press, 1987).
11. Laura Purdy (unpublished manuscript) graphically illustrates this phenomenon in
her autobiographical account of her own illness: "Even normally articulate and intelligent pa-
tients may have been turned to jelly by pain and anxiety and thus be unable to take the lead."
12. Wolf, Introduction, Feminism and Bioethics.
13. Leslie J. Blackball et al., "Ethnicity and Attitudes toward Patient Autonomy," jour-
nal of the American Medical Association 274 (1995): 820-825.
14. Jennifer Nedelsky, "Reconceiving Autonomy: Sources, Thoughts and Possibilities,"
Yale Journal of Law and Feminism \ (1989): 7-36.
15. Martha Minow and Mary L. Shanley, "Relational Rights and Responsibilities: Re-
visioning the Family in Liberal Political Theory and Law," Hypatia 11 (1996): 4-29.
16. Susan J. Brison's remarkably insightful article, "Outliving Oneself: Trauma, Mem-
ory, and Personal Identity," is in Feminists Rethink the Self, ed. Diana T. Meyers (Boulder, CO:
Westview, 1997), pp. 12-39- The passage 1 refer to here appears on page 28.
17. Ibid., p. 14.
18. I borrow this usage from Diana T. Meyers, Self, Society, and Personal Choice (New
York: Columbia University Press, 1989).
19. In Sherwin's "Feminist Ethics and Ambivalence about Autonomy" (unpublished
1994 manuscript) and her "A Relational Approach to Autonomy in Health Care," in The Pol-
itics of Women's Health: Exploring Agency and Autonomy, ed. S. Shcrwin (Philadelphia: Temple
University Press, 1998), pp. 19-47.
20. For elaboration on this point, see Adrienne Asch and Gail Gellcr, "Feminism,
Bioethics, and Genetics," in Wolf, Feminism and Bioethics, pp. 318-350. Note especially pp.
341-342.
21. As some 1970s feminists advocated, such as Shulamith Firestone in The Dialectic
of Sex (London: Jonathan Cape, 1971).
22. This particular clase was heard in Britain [(Re Y (Mental Incapacity: Bone Marrow
Transplant) 1 996 2FLR 787J, but the U.S. courts have employed similarly contorted reason-
 Autonomy and Interdependence 255

ing in cases that involve incompetents. The most celebrated is Strunk v. Strunk (445 S.W. 2d
145, Ky. 1969). James Dwyer and Elizabeth Vig discuss several others in "RethinkingTrans-
plantation between Siblings," Hastings Center Report 25 (1995): 7-12. They build a case for
a less individualistic decision-making standard and draw analogies between organ donation
between siblings and more mundane situations in which children are commonly expected to
take responsibility for one another.
23. I have in mind here not only Dwyer and Vig's discussion but also the work of a host
of feminists who have been influenced by an ethic of care. An excellent introduction to this
literature is Justice and Care: Essential Readings in Feminist Ethics, ed. Virginia Held (Boulder,
Colo.: Westview Press, 1995). Communitarians also tend to underscore the importance of
rearing children to be responsive to others' needs. However, feminists are likely to fault their
tendency to reinscribe traditional expectations that women sacrifice their own needs to (more
powerful) others. I contrast aspects of these perspectives that pertain to accounts of personal
autonomy in "Reworking Autonomy: Toward a Feminist Perspective," Cambridge Quarterly
of'HealthcareEthics4 (1995): 44-55.
24. When testing indicates that the father who worried about his carrier status is not a
carrier after all, he might be so relieved to learn the truth that the accompanying information
about the child's biological paternity would seem insignificant by comparison. Even when the
practitioner has reason to believe that the social father would be distressed, the importance
of disclosure may clearly override any adverse consequences. In situations in which the study
provides little new information about the social father except that the child is not biologically
his, the more common practice is to disclose the results to the mother and let her decide
whether to inform her husband. Many practitioners believe that at least she should be in-
formed since this information might have considerable impact on future childbearing deci-
sions. When the mother is the primary patient, this strategy maximizes her capacity to make
future decisions autonomously. But if the couple were later to split up and the father contin-
ued to harbor guilt about an affected child, the same information might have an important
bearing on his future decisions, too.
25. Dorothy C. Wertz reports in "Parents' Responses to Predictive Genetic Testing in
Their Children," Journal of Medical Genetics 33 (1996): 313-318, that 23% of HD clinics
tested children under the age of twelve. See also Angus Clarke and Frances Flinter, "The Ge-
netic Testing of Children: A Clinical Perspective," in The Troubled Helix: Social and Psycho-
logical Implications of the New Human Genetics, ed. T. M. Marteau and M. Richards (Cam-
bridge: Cambridge University Press, 1996), pp. 164-176. Also relevant is David A. Ball,
A.Tyler, and P. Harper, "Predictive Testing of Adults and Children," in Genetic Counselling:
Practice and Principles, ed. Angus Clarke (London: Routledge, 1994), pp. 63-94.
26. This point is made in the Stanford Working Group, "Report on Genetic Testing for
Breast Cancer Susceptibility" (Palo Alto, Cal.: Stanford University Center for Biomedical
Ethics, 1996), p. 14.
27. Nancy Wexler develops this point in "Clairvoyance and Caution: Repercussions
from the Human Genome Project," in The Code of Codes: Scientific and Social Issues in the
Human Genome Project, ed. Daniel J. Kevles and Leroy Hood (Cambridge, Mass.: Harvard
University Press, 1992), pp. 211-243. Among themselves, genetic counselors refer to these
people as monitors and blunters. Monitors can't live with uncertainty and want to know
everything. Blunters don't welcome any information and cannot be engaged in any substan-
tive discussion at all.
28. For a fuller account of such situations see the fascinating chronicles of personal ex-
perience with genetic disease included in Marteau and Richards, Troubled Helix.
29. In Self, Society, and Personal Choice, Meyers calls this kind of project "programmatic
autonomy," and she distinguishes it from "episodic autonomy," which focuses on particular sit-
256 Relational Autonomy in Context

nations. "Personal autonomy" in her account is a function of these two. My own account of au-
tonomy resembles hers at some junctures but puts less emphasis on the inner psychological
states of the individual subject and more on outer social relations that figure in inner represen-
tations. Shifting the locus of consideration in this way brings out interconnections between in-
dividual strivings to shape an autonomous identity, social structure, and personal history. This
shift in emphasis brings patterns of power and authority into clearer focus. For further discus-
sion of their distinction, see Kathryn Addelson, Impure Thoughts (Philadelphia: Temple Uni-
versity Press, 1991), especially chap. 11. Later 1 refer to such rethinking of individual self-
conceptions insofar as they depend on shared family history as "reciprocal relational autonomy."
30. This point is stressed by Martin Richards in "Families, Kinship and Genetics," in
Marteau and Richards, Troubled Helix, pp. 249-273.
31. Personal communication. Unless otherwise indicated, all cases are direct reports of
actual cases generously told to me by staffs of genetics clinics.
32. See Angus Clarke's introduction to Genetic Counselling, p. 14.
33. According to Martin Richards in "Families, Kinship and Genetics," p. 266.
34. Marilyn Friedman, "Autonomy and Social Relationships: Rethinking the Feminist
Critique," in Meyers, Feminists Rethink the Self. Note her remarks on pp. 55-56.
35. In "Abortion and Embodiment," Australasian journal of Philosophy "/'0 (1992):
136-155, Catriona Mackenzie provides a carefully nuanced analysis of the responsibilities of
the pregnant woman toward her fetus, which aims to show how abortion is most fundamen-
tally about women's self-determination. The fetus's moral status, she argues, depends not only
on its intrinsic properties but also on relational properties it has with others, particularly the
woman carrying it. These relational properties arc continually in a process of change as the
fetus's intrinsic properties develop and the pregnant woman's bodily subjectivity alters. Her
analysis of relationships among self-determination, embodiment, and responsibility has far-
reaching implications for other contexts, too, some of which I allude to (too briefly) here.
Marilyn Friedman offers a somewhat different reading of interconnections between auton-
omy and responsibility in her appraisal of Gauguin's "moral luck" in this volume.
36. Brison, "Outliving Oneself," p. 30. In her many autobiographical notebooks, May
Sarton owns up to comparable limitations in her own capacity to define herself.
37. Ibid., pp. 13-14.
38. John Hardwig, "What about the Family?" Hastings Center Report-20 (1990): 5-10.
39. Hilde Lindemann Nelson and James Lindcmann Nelson, The Patient in the Fam-
ily: An Ethics of Medicine and Families (New York: Routlcdge, 1995). Note especially pp.
114-116.
40. See Ann Williams, "Genetic Counselling: A Nurse's Perspective," in Clarke, Genetic
Counselling, 44-62. Note remarks on p. 48.
41. Note Peter S. Harper's remarks on this point in "Personal Experiences of Genetic
Diseases: A Clinical Geneticist's Reaction," in Marteau and Richards, Troubled Helix, pp.
54-59.
42. My thinking on this issue has been sparked by many sources, but I owe a particu-
lar debt to insights from Mary B. Mahowald et al., "The New Genetics and Women," Mil-
bank Quarterly 74 (1996): 239-283.
43. These range from profound mental retardation and early death, in the instance of
some trisomies and Tay-Sachs, to disorders that affect daily living and shorten the life span
but do not cause serious mental incapacity—cystic fibrosis and hemophilia, for instance. For
more details, see Lori B. Andrews, Jane K. Fullarton, Neil A. Holtzman, and Arno G. Mo-
tulsky, eels., Assessing Genetic Risks: Implications for Health and Social Policy (Washington,
D.C.: National Academy Press, 1994), particularly p. 75.
44. Gwen Anderson's 1990 (unpublished) study of a group of couples she followed
 Autonomy and Interdependence 257

through pregnancy illustrates this phenomenon. The women in her study worried consider-
ably more than their husbands about what they would do following prenatal diagnosis if the
results were positive. In one case the worry was borne out and the worrier was far better pre-
pared to cope with the news than her nonworrying husband.
45. To make a single cell from the newly developing embryo available for testing, the
woman must undergo in vitro fertilization (IVF), a costly procedure that is seldom covered
by insurance and that is uncomfortable, time consuming, and subject to failure at several
junctures. She must endure the risks inherent in the surgical procedures to retrieve her ova
and transfer the embryo to her uterus, and she needs to weigh the probabilities of failure, for
often the embryo fails to implant. Success rates from IVF continue to be disappointing, hov-
ering around 15%—higher at the most experienced clinics and lower at those with little ex-
perience. Although success is more likely when there is no evidence of infertility and may
even approach pregnancy rates achieved through unprotected sexual intercourse, whether the
success rate justifies the bodily risk and expense is doubtful. Undertaking IVF for genetic in-
dications intrudes on the woman's life even more extensively than IVF alone because she is
pursuing that route out of a desire not just for any child but also for a child of a certain de-
scription. For recent figures on IVF success rates in the United States, see "Assisted Repro-
ductive Technology in the United States," Fertility and Sterility 71 (1999): 798-807. For
British data (which are likely to be more accurate since all clinics are requiredto report), see
Sixth Annual Report, Human Fertilisation and Embryology Authority, 1997.
46. Currently these include Tay-Sachs disease, muscular dystrophy, hemophilia, fragile
X, Down syndrome, HD, and achondroplasia.
47. Misinformation is very common with such conditions as spinal muscular atrophy.
For a fuller account, see Harper, "Personal Experiences of Genetic Diseases," in Marteau and
Richards, Troubled Helix, pp. 54-59.
48. For a sensitive account of their situation, see Lavanya Marfatia, Diana Punales-
Morejon, and Rayna Rapp, "Counseling the Underserved: When an Old Reproductive Tech-
nology Becomes a New Reproductive Technology," in Strategies in Genetic Counseling: Repro-
ductive Genetics and New Technologies, ed. B. A. Fine et al. (March of Dimes Birth Defects
Foundation: Birth Defect: Original Article Series, vol. 26, 1990), pp. 109-126.
49. See Richards, "Families, Kinship and Genetics," p. 270.
50. On this point see Shoshana Shiloh, "Decision-Making in the Context of Genetic
Risk," in Marteau and Richards, Troubled Helix, pp. 82-103. Note especially pp. 90-91.
51. See Clarke's introduction in Genetic Counselling, p. 18. He attributes this informa-
tion to M. Holmes-Seidle et al., "Parental Decisions Regarding Termination of Pregnancy
Following Prenatal Determination of Sex Chromosome Abnormality," Prenatal Diagnosis 7
(1987): 239-244, and A. Robinson et al., "Decisions Following the Intrauterine Diagnosis
of Sex Chromosome Neuploidy," American Journal of Medical Genetics 34 (1989): 552-554.
T. M. Marteau, H. Drake, and M. Bobrow confirm this observation in "Counseling after
Fetal Abnormality: The Differing Approaches of Obstetricians, Clinical Geneticists, and Ge-
netic Nurses," Journal of Medical Genetics 31 (1994): 864-867; and Josephine Green and
Helen Stratham cite additional sources supporting it in "Psychosocial Aspects of Prenatal
Screening and Diagnosis," in Marteau and Richards, Troubled Helix, pp. 140-163.
52. A survey that sampled obstetricians in England and Wales showed that over a third
would not refer a woman for genetic testing unless she agreed in advance to rerminate an af-
fected pregnancy. For elaboration of this point, see Green and Stratham, "Psychosocial As-
pects," pp. 150-151.
53. Quoted from the National Society of Genetic Counselors Code, 2.2, 1991. Cited
in Fern Brunger and Abby Lippman, "Resistance and Adherence to the Norms of Genetic
Counseling," Journal of Genetic Counseling4 (1995): 151-167.
258 Relational Autonomy in Context

54. For a persuasive account, sec Abby Lippman, "Prenatal Genetic Testing and Screen-
ing: Constructing Needs arid Reinforcing Inequities," American Journal of Law and Medicine
17 (1991): 15-50. (Reprinted in Clarke, Genetic Counselling, pp. 142-186).
55. See Brunger and Lipmann, "Resistance and Adherence," and Arthur L. Caplan,
"Neutrality Is Nor Morality: The Ethics of Genetic Counseling," in Prescribing Our Future:
Ethical Challenges in Genetic Counseling, ed. Diane M. Barrels, B. LcRoy, and A. L. Caplan
(New York: Aldine De Gruyter, 1993), pp. 149-165.
56. In "Interviewing Women: A Contradiction in Terms," in Essays on Women, Medi-
cine, and Health (Edinburgh: Edinburgh University Press, 1993), Ann Oaldcy formulates
such a model specific to the relationship between a research interviewer and her interviewees.
She develops this model more fully in her subsequent book, Social Support and Motherhood
(Oxford: Blackwell, 1995). The model she proposes, with appropriate modifications, could
be transported to the counselor-client relationship.
57. Wolf, Feminism andBioethics, p. 18.
 11
RELATIONAL AUTONOMY,
S E L F - T R U S T , A N D H E A L T H CARE
FOR PATIENTS WHO ARE O P P R E S S E D 1

Carolyn McLeod
Susan Shenvin

Relational Autonomy and Self-Trust

Traditional autonomy theory, especially as it is deployed in the context of health-
care ethics discussions, has focused on specific factors that interfere with autonomy:
coercion, internal compulsion, and ignorance.2 Depending on the circumstances
and the severity, the presence of any of these factors may make it impossible, or at
least very difficult, for an agent to act autonomously.3 What traditional accounts
tend to overlook, however, is that there is a fourth way in which an agent's auton-
omy may also be compromised, namely, by the forces of oppression.4 Oppression
may itself involve dimensions of coercion, compulsion, and ignorance, but it func-
tions in complex and often largely invisible ways, affecting whole social groups
rather than simply disrupting isolated individuals; as a result, its effects tend to be ig-
nored within the traditional autonomy framework that focuses solely on individuals.
Moreover, addressing these features requires changing the broad social conditions
that constitute oppression and not merely changing some of the specific circum-
stances of an individual agents situation. It is necessary, then, to explore the distinct
ways in which oppression may interfere with a persons ability to act autonomously.
To make explicit the role of oppression in autonomy, we appeal to a concept of
relational (or contextual) autonomy. We understand relational autonomy to involve
explicit recognition of the fact that autonomy is both defined and pursued in a so-
cial context and that social context significantly influences the opportunities an

259
260 Autonomy and the Social

agent has to develop or express autonomy skills. In relational autonomy, it is neces-

sary to explore an agent's social location if we hope to evaluate properly and respond
appropriately to her ability to exercise autonomy. Whereas traditional accounts con-
cern themselves only with judging the ability of the individual to act autonomously
in the situation at hand, relational autonomy asks us to take into account the impact
of social and political structures, especially sexism and other forms of oppression, on
the lives and opportunities of individuals. By making visible the ways in which au-
tonomy is affected by social forces, especially oppression, relational autonomy chal-
lenges assumptions common to much bioethical literature that autonomy be viewed
as an achievement of individuals.
We must, therefore, evaluate society and not just the individual when deter-
mining the degree to which an individual is able to act autonomously. Insofar as op-
pression has reduced or undermined an agent's ability to act autonomously in vari-
ous contexts, relational autonomy seeks politically aware solutions that endeavour to
change social conditions and not just expand the options offered to agents. In par-
ticulat, a relational view of autonomy encourages us to understand that the best way
of responding to oppression's restrictive influence on an individual's ability to act au-
tonomously is to change the oppressive conditions of her life, not to try to make her
better adapt to (or simply to manage to "overcome") those conditions privately.
\Ve live in a society complicated by many intersecting forces of oppression. Al-
though we may challenge them and work to reduce their influence, we cannot hope
to see their disappearance in the near future. Many members of our society have
been damaged to varying degrees by these forces of oppression, and many have
diminished ability to exercise autonomy as a consequence of their experiences as
members of oppressed groups. Unfortunately, the damage done by oppression may
be so deep that the effects can never be completely erased, 5 but reducing oppressive
circumstances can increase opportunities to repair some of the damage and prevent
its occurrence in future generations. Because oppression is always an impediment to
autonomy, any society truly committed to promoting autonomy must work to elim-
inate the forces of oppression that reduce citizens' opportunities to become maxi-
mally autonomous. 6
Of course, this is not to say that every person who belongs to one or more
groups that is subject to oppression is incapable of exercising autonomy or that
everyone who belongs to a dominating group is fully autonomous. Rather, it is to
make clear that oppression tends to interfere with an agent's ability to develop or ex-
ercise autonomy effectively in specific ways. Individual members of oppressed
groups are affected to varying degrees by the forms of oppression that are endemic
to their society; some manage to overcome the oppressive circumstances of their
lives largely unscathed. Furthermore, since there arc many intersecting forms of op-
pression, each person belongs to multiple groups with multiple locations on any map
of oppression and may well be privileged in some respects even as she or he is op-
pressed in others, and this complex positioning may have an impact on the agent's
opportunity to develop the abilities needed to act autonomously. Also, the features
that support the development of some of the skills necessary to exercise autonomy
are products not only of large social forces but also of particular personal relation-
ships; the latter vary significantly from one individual to the next even within the
 Relational Autonomy, Self-Trust, and Health Care 261

same social groups. Membership in one or more oppressed groups is an obstacle,

then, but not an insurmountable barrier to the development of relational autonomy.
Here we focus specifically on one particular dimension of oppression that in-
terferes with autonomy, namely, oppression's effect on self-trust. We argue that an
agent requires a certain degree of self-trust to be able to act autonomously; because
self-trust is undermined by oppression, oppression reduces an agent's ability to act
autonomously. To make our argument more concrete, we review some areas of health-
care where it is evident that the effects of oppression on self-trust interfere with pa-
tients' ability to act autonomously, and we consider some of the options available to
health-care providers in such circumstances. We discuss a variety of cases, concen-
trating particularly on problems associated with caring for women with serious ad-
dictions, as this is an area where the links among oppression, self-(dis)trust, and au-
tonomy create especially difficult challenges for health-care providers. We explore
how health-care providers ought best to respond to these patients. In our view, there
is no single solution to this difficulty but a range of responses that must vary with
the health needs and the situation of each patient.
We begin by looking at some of the distinct ways in which oppression can
limit an agent's ability to act autonomously in order to appreciate how self-trust fits
within this complex matrix. First, oppression may involve explicit or implicit limi-
tations on the options available to members of oppressed groups. An important ex-
ample is the way in which oppression may limit the options that are reasonable for
an agent to choose under various circumstances by placing her in a double bind;7 in
such cases, whatever she does, she will suffer. For example, it is not unusual for
women who are living with violent partners to feel trapped in those dangerous re-
lationships. They may feel that they cannot leave the relationship because the dan-
ger of such action would be too great: generally, the most serious attacks of do-
mestic violence occur when the battered partner attempts to separate. Moreover,
separation may bring not only the heightened risk of attack but also harassment,
financial difficulties, the risk of losing custody of her children, and for immigrant
women the threat of deportation. On the other hand, remaining with the abusive
partner clearly represents a serious danger to the woman and her children. Unless
she has access to excellent social services for battered women, there may be no safe
option available to her. Less dramatic examples of oppression-related double binds
involve the ways in which women are often considered suspect as job candidates if
they are single and if they are married, if they have children and if they do not, and
if they quit their previous job and if they were fired.
Second, and more deeply, oppression may shape agents' values and desires in
ways that undermine their capacity for autonomous choice in certain matters. In
fact, women's oppression typically involves circumstances in which the agent's im-
mediate interests appear to support her active participation in practices that actually
promote her oppression. For example, part of women's oppression is that they are
primarily valued (by themselves, as well as by others) in terms of their instrumental
role as mothers. Many women who have difficulty in conceiving or maintaining a
pregnancy seek medical interventions to circumvent infertility problems, using pow-
erful fertility-enhancing drugs and often surgical treatments in the pursuit of a baby.
In so doing, they reinforce the social norms that proclaim the profound importance
262 Autonomy and the Social

of women being able Co reproduce. The relentless efforts of so many infertile

women to spare no risk or expense to produce a biologically related child help to
deepen social expectations about the normalcy of childbearing for all women and to
raise questions about the nature of other women who choose childlessness. What-
ever their intentions, their participation in practices that take for granted the desir-
ability of biological reproduction helps to further entrench social attitudes about
women's fundamental role as mothers. 8
Our main focus in this article is on two other ways in which oppression can in-
terfere with autonomy. One is that oppression tends to deprive a person of the op-
portunity to develop some of the very skills that are necessary to exercise autonomy
by restricting her opportunity to make meaningful choices and to have the experi-
ence of having her choices respected. As Diana Meyers argues, exercising personal
autonomy involves certain types of competency that depend on the development of
corresponding skills. 9 If an agent is never exposed to an environment that fosters the
development of those skills, she will lack the ability to exercise autonomy. Moreover,
she will have been deprived of the opportunities to develop the level of self-trust
that is necessary for her to gain and use these skills effectively. She may then not be
in a position to exercise autonomy even when invited to do so. Hence, it is not suf-
ficient simply to offer a person an uncoerced choice; it is also necessary to ensure
that she has had the opportunity to learn to exercise choice responsibly.
Lastly, oppression can lead to the internalization of a sense of social worthless-
ness and incompetence that is translated into a lack of self-worth and self-trust.
When a group is oppressed, the society at large operates as if that group is less wor-
thy and less competent than others and devalues its members. Members of op-
pressed groups may then internalize these attitudes; many are inclined to accept
society's devaluing of their personal worth on at least an unconscious level and to
doubt their own worth and ability to make appropriate choices.10 This lack of self-
worth arid self-trust may be devastating to agents' autonomy competency, interfer-
ing with their ability to act according to their own interests at all.
Many feminists have explored the effect: of these various aspects of oppression
on autonomy, and several have addressed the necessity of having an adequate sense
of self-worth. To date, none have focused explicitly on the effect of oppression-
related reductions in self-trust on the agent's autonomy. 11 Our aim here is to bring
out the significance of self-trust in discussions of oppression and autonomy. We are
interested in exploring the ways in which oppression interferes with the develop-
ment of self-trust and with the fact that reduced self-trust tends to reduce an agent's
ability to act autonomously. And we seek to understand the special challenges
health-care providers face in caring for patients with the diminished autonomy that
is associated with low levels of self-trust.
Self-trust (and its correlate, self-distrust) must be considered when evaluating or
promoting autonomy skills because autonomy is dependent on self-trust. 12 Exercis-
ing autonomy involves, in part, reflecting on one's beliefs, values, and desires; mak-
ing reasonable decisions in light of them; and acting on those decisions. It is essen-
tial in developing the capacity to be autonomous that the agent trusts her capacity to
make appropriate choices, given her beliefs, desires, and values; that she trusts her
ability to act on her decisions; and also that she trusts the judgments she makes that
 Relational Autonomy, Self-Trust, and Health Care 263

underlie those decisions. These include judgments about the values and motivations
that lie behind her decisions and about the efficacy of her own decision-making
skills. Without trust in these judgments and trust overall in her ability to exercise
choice effectively, any agent would have little motivation to deliberate on alternative
courses of action.
In recognizing some degree of self-trust to be a precondition of autonomy, we
are offering a substantive view of autonomy that builds on existing theories formu-
lated by Paul Benson and Robin Dillon.13 Autonomy theories can be either sub-
stantive or procedural. Procedural accounts require that the agent subject her beliefs,
values, and desires to some procedure or method of evaluation and that she act on
whatever beliefs or desires satisfy that procedure. Meyers, for example, offers a pro-
cedural account in which the goal of the procedure is to determine what the agent
really believes or desires, and being successful in this process of self-discovery re-
quires that the agent possess certain competency skills.14 We accept the procedural
dimensions of autonomy and believe that self-trust has a role to play within them,
but we also believe that an autonomy theory must be supplemented with a limited
substantive demand. What distinguishes substantive accounts from procedural the-
ories is that the latter, unlike the former, are "content-neutral"; the procedure of
evaluation does not dictate that the agent must really believe anything specific about
herself or the world to be autonomous. Substantive theories, on the other hand, put
restrictions on the kinds of beliefs, values, and desires an agent must have in order to
be autonomous. The restriction Benson and Dillon argue for is to have a positive
conception of one's own worth and respect for one's capacities. The requirement
that the agent value or respect herself puts a limit (albeit a minimal one) on the sorts
of beliefs or feelings that she can have about herself if she is to be autonomous.15
We propose a different, though similar limit in requiring that the agent trust her
own judgment. As we explain below, we believe that self-trust is a criterion for au-
tonomy distinct from that of self-respect or self-worth.
The role of self-trust within the procedural dimensions of autonomy includes
the following aspects. First, to be motivated to exercise her own choices, the agent
must trust her capacity to choose effectively, a type of self-trust that we refer to as
Type I. Having this capacity involves having good decision-making skills and also
being situated to choose well, meaning that the agent is adequately informed of al-
ternative courses of action and of whatever facts are relevant to her decisions. For
her to trust herself to make good decisions, she must trust her competency skills and
the accuracy and adequacy of the information available to her. Distrust in these
areas inhibits her autonomy because it makes it very difficult for her to formulate
decisions about what she should believe or desire and how she should act. As noted,
oppression can cause self-distrust in the agent's decision-making skills by depriving
her of sufficient opportunities to develop and exercise those skills. People who be-
lieve that she is less competent than others by virtue of being a member of a partic-
ular social group will deny her these opportunities. Oppression can also limit her
knowledge base for making decisions by ensuring that most of the information cir-
culating throughout her society is about the lives of members of the dominant
group and the risks or benefits they would incur by making particular decisions,
such as those about their own health care.
264 Autonomy and the Social

Second, an agent must trust her ability to act on the decisions she makes, a type
of self-trust that we call Type II. She may lack this form of self-trust because she
lacks the courage to act on her judgments and consequently distrusts her ability to
do so. She may also lack it because of conflicting desires that stand in the way of her
acting on her judgments. Such desires are commonplace among people who are op-
pressed and who are attempting to fight their oppression. Their trust in their ability
to act on decisions that oppose their oppression can be compromised by the desire
simply to get along with others or the desire to experience the benefits of conform-
ing to dominant stereotypes or interests. For example, some women find it very dif-
ficult to stick with a decision not to appear unwise or innocent around men in po-
sitions of power, who would not pay much attention to them otherwise.
Third, the substantive demand on autonomy concerned with self-trust is that
the agent must trust the judgments she makes that underlie her own choices. We call
her trust in her own judgment Type III. Judgments that are relevant to her auton-
omy are, for example, her judgments about the trustworthiness of her own decision-
making skills. Whether she trusts or distrusts her evaluations of her own capaci-
ties as an agent will depend on her level of self-knowledge, or the degree to which
she feels that she understands her own strengths or weaknesses as an agent. Other
autonomy-related judgments concern the values that inform her decisions. She must
have some confidence in the appropriateness of her values for her to be motivated to
make decisions that reflect them. If her values have been shaped by oppression, it
will not be easy for her to trust them because they will encourage her participation
in practices and behavior that undermine her moral worth and that may cause her
severe suffering.
Each of these three different types of self-trust or distrust can be situadonal or
more general, meaning that they can apply only to specific situations or to most
situations in which the agent finds herself. 16 For example, most of us distrust our
decision-making capacities in at least some situations, in particular, those in which
we know that we lack the knowledge required to make an appropriate choice. People
trained only in philosophy probably would (and should) distrust their ability to
make good decisions in situations in which training in engineering is called for. Type
II self-trust is also often situational; for example, whether we have the courage to act
on our decisions usually depends on the context in which we have made them. If we
lack that courage in most situations, our autonomy will suffer, and that happens
whenever any of the three types of self-distrust occur generally rather than only sit-
uationally. On the other hand, situational self-distrust of any type may be minor
and, hence, trivial if it occurs in an area of little importance to the agent. However,
if it occurs in an area that relates to how she defines herself or to what she values
most, ir will be nontrivial in the sense that it will interfere with her autonomy. For
example, if she defines herself as a proud lesbian but lacks trust in her ability to act
on judgments that reflect that pride in relevant situations, that situational Type II
self-distrust will negatively affect her autonomy.
Just as it is appropriate sometimes to distrust others, it is also appropriate some
times to distrust oneself. There are situations in which it is justified for an agent to
distrust her ability to act on her judgments. The justification lies in a history of fail-
ing to act on the relevant judgments in the relevant situations. Often situational self-
 Relational Autonomy, Self-Trust, andHealth Care 265

trust and distrust develops, at least in part, through inductive reasoning about past
successes or failures at using or attempting to use our various abilities. Depending
on the soundness of this reasoning, the agent's self-trust or distrust will be well
formed or ill formed. For example, an alcoholic may reason that she can trust her
ability to act on her decision to refrain from drinking, but her reasoning may be un-
sound, based on her past experience, in which case her trust in her ability to refrain
is ill formed.
Clarifying how the different types of self-trust develop and the exact nature of
self-trust (that is, what sort of mental attitude it is) is not something we can accom-
plish here, but we make some preliminary remarks about both. First, regarding the
development of self-trust, the fact that oppression and self-distrust are interrelated
means that self-trust does not always or merely develop through inductive reasoning
by the agent. The level of support that the agent receives within her social environ-
ment will have a profound influence on her self-trust. That support can exist on two
different levels: (1) the agent can be given opportunities to develop and use her var-
ious capacities and, through these opportunities, learn to trust her capacities; (2) the
agent can receive encouragement from others to trust her own capacities. On the first
level, the self-trust is relational in a causal sense; supportive social conditions provide
the materials for its development. On the second level, self-trust is relational in a
constitutive sense; the agent's trust in herself exists in part because others reinforce
that trust in their relationships with her. Our self-appreciation is influenced by the
opinions that others have of us, particularly when we are young. It is doubtful that
anyone could ever avoid the constitutively relational aspect of self-trust and distrust.
The way in which a particular instance of self-trust or distrust develops will de-
termine whether that trust is informed primarily by beliefs or feelings. The primary
influence on self-distrust that arises through inductive reasoning about past failures
will probably be a belief, whereas the primary influence on self-distrust that exists
because of subtle attempts by others to undermine self-trust will probably be a feel-
ing about our own incompetency. Such subtle attempts, which may not be fully
conscious on the part of those who make them, are common in relationships that
involve oppression.17 Oppressed people often receive the subtle (and occasionally
not so subtle) message from others that their opinions are not as credible or as im-
portant as the opinions of the dominant group. Often the message is vague, and so
it produces a vague sense or a feeling in the agent, rather than a belief, that her judg-
ment is untrustworthy.
Thus, self-trust, like interpersonal trust, is an attitude that is shaped by beliefs
and/or feelings about the trustworthiness of the trusted. This general description of
trust is consistent with the view of many trust theorists, such as Annette Baler,
Richard Holton, Karen Jones, and Trudy Govier, even though they differ in their
elaborations on the view that trust is an attitude.18 For example, Jones says that trust
is an attitude of optimism about the goodwill and competence of the one trusted,
and it involves an expectation that that person will be "directly and favorably moved
by the thought that we are counting on her."19 Holton agrees that it is an attitude
that entails assumptions about the future motivations of the trusted person, but he
disagrees that it is necessarily informed by a belief about the goodwill of that person.
We can infer from where the agreement lies about the nature of interpersonal trust
266 Autonomy and the Social

that the attitude of self-trust concerns the competence of the self and expectations
about how one will be motivated to act in the future.
Thus, we can see how self-trust differs from self-worth and self-respect; it is not-
interchangeable with either of them, although that is how some authors, such as
Benson and Dillon, treat those forms of self-appreciation.20 Self-trust differs from
self-worth in being about our competence rather than our worth. It makes sense, for
example, to talk about trusting our competence to judge our worth. Self-respect can
also be about our competence, but there is a sense in which the positive evaluation
it gives of our competence is grounded in our past behaviour.21 In trusting ourselves,
we are optimistic that we will be able to carry that evaluation forward into the fu-
ture. Thus, one can trust oneself to act in the future in a way that maintains one's
self-respect. However, it is doubtful that one could have that trust in the absence of
self-respect, and moreover, that one could acquire self-respect regarding one's com-
petence without some self-trust. Those self-regarding attitudes are distinct, yet they
are mutually reinforcing.
Some theorists, particularly Holton, argue that we can decide to develop trust-
ing attitudes, whereas we cannot simply decide to believe. These attitudes are con-
strained somewhat by evidence that contradicts them and by our current beliefs and
feelings; for example, we cannot decide to trust a person whom we believe to be un-
ttustworthy. However, outside of some minimal constraints, we can will ourselves to
trust. As Holton explains, willed attitudes of trust that are unreasonable, given the
evidence and our relationship with the trusted, are ill founded.22 Likewise, trusting
attitudes toward the self can be ill founded; for example, if I develop a trusting at-
titude toward my ability to take a bobsled down a mountain without killing myself
but I have no evidence of my ability to do so, that attitude is unreasonable.
In conclusion, we have reviewed how different types of self-trust can damage
our autonomy and explained how they can be related to oppression. Being fully au-
tonomous in our view requires having all nontrivial forms of the different types of
self-trust. We recognize that autonomy exists in degrees and that lacking some in-
stances of nontrivial self-trust docs not make an agent completely nonautonomous.
However, our interest lies primarily in cases in which the agent's autonomy has been
damaged severely by oppression-related reductions in self-trust, and we intend to
discuss how health-care providers should respond to such patients. In the next sec-
tion, we explore the influence on health-care provision of the impact of oppression
on self-trust and autonomy. At the end of it, we consider cases in which that impact
is severe, specifically, among women with addictions.

Responding to the Effects of Oppression

on Self-Trust in Health Care
It is at present widely agreed throughout the biomedical health-care community that
patients should make autonomous decisions about their health care to the greatest de-
gree possible. There is no clear consensus, however, on how health-care providers are
to proceed when caring for patients who are not fully autonomous. We are concerned,
specifically, with the question of determining how health-care providers should re-
spond to patients whose autonomy is diminished as a result of their oppressed status.
 Relational Autonomy, Self-Trust, and Health Care 267
Typically, the health-care system has responded to patients with reduced levels
of autonomy (as identified by the traditional analysis) by exercising paternalism
(making decisions on behalf of patients without their full consent). Indeed, exer-
cising paternalism has been customary throughout the history of medicine, even
when dealing with autonomous patients. It is only in the past few decades, pater-
nalism has become widely recognized as a direct threat to autonomy; as a result,
health-care providers have been formally discouraged from exercising paternalism
when dealing with patients who are capable of making autonomous decisions.
Nonetheless, a certain degree of paternalism pervades modern practice. Health-
care decisions can be exceedingly complex, requiring understanding of a great deal
of technical information and careful weighing of options. Even the most inde-
pendent and self-reliant patient often feels overwhelmed and is inclined to defer to
medical authority when facing serious health matters. In such a context, when
health-care professionals are faced with patients whose autonomy is suspect for any
reason, it may seem appropriate to exercise paternalism to ensure the best possible
care for these patients.
Paternalism on the part of health-care professionals is always questionable, how-
ever, for it may involve significant distortion of the patient's real interests. The exer-
cise of paternalism is especially problematic when applied to patients whose auton-
omy is reduced by virtue of their history of oppression. Oppression involves unjust
distributions of power, and health-care settings are sites of very uneven power dif-
ferentials. If health-care professionals, especially physicians, further consolidate their
already disproportionate power in relation to patients, especially those from op-
pressed groups, they exacerbate a problematic power differential and further reduce
the already limited autonomy of their patients. Moreover, they are unlikely to be in
a position to know what is ultimately in the best interests of patients whose life ex-
periences are very different from their own; hence, they are unlikely to be in a posi-
tion to exercise paternalism wisely.23 Other solutions to the problem of reduced au-
tonomy must be found.24
The impact of oppression on patients' ability to exercise autonomy can be felt
in many different ways. For example, the ability to exercise autonomous choice re-
quires, among other things, access to appropriate information. Competent health-
care providers strive to simplify the relevant information and to ensure that patients
appreciate the meaningful choices before them, but most shape the decisions their
patients will make by tailoring the information to ensure the selection of what the
health-care expert considers the best choice for each patient—an indirect form of
paternalism. Patients' autonomy is generally reduced to the exercise of "informed
choice" in which the information provided is restricted to that deemed relevant by
the health-care provider (and by the health-care system, which has determined what
information is even available by pursuing certain sorts of research programs and ig-
noring others). Even in "ideal" cases in which patients have strong autonomy skills
and full access to all the available information, it is important to recognize the influ-
ence that oppression may have on the information base and, thereby, on the mean-
ingful options available to patients.
Specifically, oppression tends to restrict the relevant knowledge base that can be
called on for making health-care decisions. It may limit the health-care providers'
268 Autonomy and the Social

ability to appreciate the type of information the patient might need to know in
order to choose wisely in her circumstances. For example, the need to arrange for
flexible work hours to adapt to a treatment regime may not seem problematic to a
physician but may pose an insurmountable barrier to women in certain types of
jobs. Furthermore, oppression may limit the knowledge base itself; for example, al-
though heart disease is the leading cause of death in women, many important stud-
ies of heart disease have used only male subjects, so there are insufficient data about
the effectiveness of prevention or treatment programs for women (e.g., the value of
taking a daily dose of aspirin as a preventive measure). Similarly, research into treat-
ment for AIDS has concentrated on the progress of the disease on men; as a result,
women must often choose drug regimens whose effects on women's bodies have not
been sufficiently explored. 25 In these sorts of cases, relational autonomy helps us to
recognize that circumstances may limit the autonomy of all patients who belong to
oppressed groups by creating an inappropriately limited knowledge base on which
such patients must rely.
These sorts oi gaps in available knowledge are problematic in that they reduce
the agent's ability to choose well. They also raise issues of trust and of Type I self-
trust. 1 hese gaps mean that often patients who are members of oppressed groups
have particular reason to distrust the information their health-care providers supply.
In addition, they mean that patients are limited in their ability to obtain independ-
ent and appropriate information on their own. In such circumstances, patients must
lack trust in their own abilities to acquire the necessary knowledge to make well-
informed decisions. Even a patient who has managed to develop strong autonomy
skills will lack Type 1 self-trust when she must choose in the face of inadequate or
skewed data. She may have complete confidence in her own ability to make such de-
cisions, but if she knows that she is making them without adequate data, she will
still not be able to fully trust the outcome. Many menopausal women feel precisely
this unease as they grapple with the question of whether or not to begin hormone
replacement therapy. Although the self-distrust at issue is quite limited in scope and
located around a particular range of choices—and, hence, is a relatively benign
form of self-distrust—it is still relevant to the autonomy of the agent. In this case,
it is a form of well-founded distrust of the results of her own deliberations because
of the limitations inherent in the circumstances in which these decisions are made.
Oppression is relevant here because the data needed to make health-care decisions
are often particularly limited for members of oppressed groups; as a result, patients
from these groups find that they must often distrust the knowledge base they need
to rely on and that distrust: will affect their confidence in many sorts of health-care
decisions they are required to make.
In such circumstances, autonomy is compromised. Paternalism is not any more
reliable, however, since health-care providers will face the same limits in knowing
what will best meet these patients' needs. Thus health-care providers have a respon-
sibility not to take over decision making from patients but to ensure that patients
understand the limits of the knowledge the former can provide. Moreover, health
workers should appreciate their collective responsibility to work toward filling in
these important knowledge gaps.
Yet more complex questions arise when we consider cases in which patients
 Relational Autonomy, Self-Trust, and Health Care 269

choose medical procedures that seem inseparable from their oppression. In these
cases, patients who are aware of the ways in which socialization may have shaped
their values and desires to conform to oppressive stereotypes will again find it diffi-
cult to trust their own deliberations. They will lack Type HI self-trust insofar as it
concerns the values and desires that inform their decisions. For example, the deci-
sion about whether or not to use hormone replacement therapy at menopause is fur-
ther complicated by factors other than inadequate and inconsistent evidence. There
is also the fact that women are encouraged to use such treatments to continue to
look young; in doing so, they are participating in the norms of a culture that prefers
its women young and beautiful. Of course, hormone replacement therapy is seldom
a straightforward, single-dimension decision; women are also encouraged to use it to
reduce their risk of heart disease and osteoporosis and now, it is suggested, of Alz-
heimer's disease. The multiple dimensions of such decisions mean that many factors
must be taken into account, making such decisions especially difficult and complex
for many women. Those who find that maintaining a youthful appearance is an im-
portant consideration are in a particularly awkward position. They may well have
excellent reasons for wanting to continue to look young for as long as possible: their
careers and romantic possibilities may well depend on appearing youthful, and their
own aesthetic sensibilities may also be an issue. The problem is that cultural atti-
tudes that consistently value young (looking) women over older women are oppres-
sive to every woman. To devalue women who clearly have reached a certain matu-
rity is to devalue important aspects of all women's lives. It reflects a value system that
cares more about women's appearance than their wisdom or experience; in such a
system, women are valued more for their ornamental role than for their personhood.
Women who are aware of this cultural prejudice against aging women, yet who feel
a strong desire to look young, will find themselves uncertain about their own moti-
vations to use hormone replacement therapy. They may also lack a form of Type II
self-trust, one relating to their ability to act on their judgment that in making
choices that affect themselves, they should be respectful of their own personhood.
They may distrust this ability because they are confused about whether choosing
hormone replacement therapy is consistent with that judgment.
Similar problems are associated with the use of cosmetic surgery to better meet
society's beauty standards. As more women make the effort to fit these norms, the
pressure grows on other women to overcome their natural "handicaps" and adapt to
the expectations that apply to women. In such ways, the evaluation of women by ex-
ternal standards of appearance becomes ever more normalized, further contributing
to the oppression of all women by overshadowing efforts to recognize them in other
terms.26 For women to participate in this value system is to reinforce it rather than
challenge it; their compliance helps to perpetuate its oppressive power.
This issue is made especially problematic because it is an area where women
have reason to distrust their own value schemes. Some of the operative values are
part of a cultural worldview that is oppressive to women in general and tends to be
especially oppressive to women who belong to marginalized groups, including dis-
abled women, lesbians, women of color, and poor women (since the norms pro-
moted are those of young, affluent, slim, fit-looking, white women). Women whose
positive evaluations of their own bodies are not attached to these exploitative social
270 Autonomy and the Social

values may find it very difficult to identify or maintain their own value schemes in
light of the availability of "cures" for certain body shapes, such as small breasts and
large tummies, which are treated as subjects for invasive medical responses. Their
difficulty in maintaining their own value systems suggests a lack of Type 11 trust in
their ability to act on decisions that reflect those systems or a lack of Type III trust in
their own evaluations of their values. Cosmetic surgeons effectively reduce many
women's body parts to material for surgical manipulation, invoking technologies
otherwise reserved for heaiing serious illness and conveying the sense that such de-
ficiencies are important enough to warrant dramatic solutions. At the same time,
they join their medical colleagues in acting as authorities for women on the health of
their bodies. The multiple messages involved make it hard for many women to trust
their own evaluations of their values regarding body shape and size when those eval-
uations are at odds with sexist beauty norms.
Health-care providers have a responsibility, then, to refrain from encouraging
such use of medical resources and to refrain from promoting the youth and beauty
their procedures may engender. They need to reflect on their role in a culture that
cares more about the superficial aspects of women's appearance than their characters
or talents. If they wish to promote the autonomy of patients who seek these proce-
dures, they should not simply respond to informed requests for surgical "correc-
tions" but, at least, also encourage their patients to consider the forces that lead to
these choices, as well as alternative responses.
In other cases, health problems are even more directly related to the patient's
oppression. For example, both poverty and stress, two conditions highly correlated
with oppression, are associated with countless illnesses and are inevitably aggravating
factors in any illness. Moreover, many health problems arc far more common in
members of oppressed groups than in the rest of the population; for example, AIDS
is more common among the poor and lupus is more common among people of
African descent in North America than in the population at large. Aboriginal
women die of cervical cancer in Canada at many times the rate of white women,
and suicide has reached epidemic proportions in some native communities in North
America. Although women tend to outlive men in much of the world, their lives are
plagued by more chronic and debilitating illnesses. Furthermore, one of the charac-
terizing features of oppression is that members of groups subject to oppression are
highly vulnerable to violent attacks; women, for example, suffer disproportionately
from the effects of domestic and sexual violence, and disabled women in particular
experience an exceptionally high rate of attack. These correlations are inseparable
from the oppressive conditions that affect disadvantaged populations. As we have
noted, such experiences undermine the autonomy of members of oppressed groups
in multiple ways; in particular, systematic abuse often interferes with an agent's sense
of self-trust, as we discuss below.
When patients appear in the health-care system with conditions in which their
oppression seems ro be a contributing factor, then, it is not sufficient simply to try to
correct the immediate damage, for that leaves the underlying contributing factors
intact. It is important for patients—and others—to understand the social and po-
litical dimensions of their condition. To restore their sense of self-trust, those who
have been assaulted need to appreciate that this violence is part of an endemic pat-
 Relational Autonomy, Self-Trust, and Health Care 271

tern and not a consequence of their own behavior; those suffering from nutrition-
related disorders because of their low incomes need to appreciate the role that
poverty (and not necessarily incompetence) plays in limiting their access to a nutri-
tious diet; and those with poorly understood illnesses should understand that there
has been inadequate research into such conditions as lupus that primarily affect dis-
advantaged populations. This knowledge should help patients feel validated in their
legitimate claims for care and should help them to avoid blaming themselves for the
conditions in which they find themselves. It may help strengthen the patients' trust
in their ability to recognize their need for help and give them guidance on how to
pursue both personal and political strategies to improve their situation. Such guid-
ance may require directing patients toward self-help activist groups that will pro-
mote a sense of empowerment and build skills and forms of self-appreciation that
are necessary for autonomy, including self-trust.
Identifying ways in which health-care providers can promote the autonomy of
patients with certain oppression-induced problems, as well as seriously low levels of
autonomy, can be exceedingly difficult, however. We now consider some of these
difficulties when dealing with a group whose members have faced especially serious
problems in developing the conditions necessary for autonomy— women with se-
rious addictions (i.e., addictions to dangerous quantities of such harmful substances
as heroin, crack cocaine, alcohol, and solvents). Most of these women tend to have
several problems in exercising autonomy. The first and most obvious is one recog-
nized by all autonomy theories: addictions are a form of compulsion, so by their
very nature, they interfere with autonomy. This feature in itself makes treatment
of serious addictions a significant problem for those committed to respecting au-
tonomy.
The compulsive nature of addiction may make it very difficult, for example, for
an addict to make a voluntary (let alone an autonomous) choice to enter a treatment
program. Hence, the question arises, when faced with an addict with a low level of
autonomy who does not wish to seek treatment or who denies that she has a drug
problem, should providers be permitted to get her into treatment by deceiving her
somehow or by simply forcing her? Despite the fact that addicts have low levels of
autonomy, methods that involve force or deception to admit them for treatment are
problematic for several reasons, most notably because the treatment that follows is
likely to be ineffective. The general consensus in the field of addiction treatment is
that many of the addict's beliefs and attitudes must change if she is to modify her
behavior, and this change will not occur in treatment if she is there unwillingly.27
Hence, whatever the ethical arguments are about coercion in such cases, there is a
strong pragmatic case against coercing any addict into treatment.
Whatever account of autonomy we might choose, then, there are good reasons
to doubt that women with serious addictions are fully autonomous and also good
reason to refrain from forcing them into treatment. We believe, though, that there
is an even deeper concern that emerges when we adopt the perspective of relational
autonomy—the fact that the autonomy of women with serious addictions is un-
dermined not only by the compulsive nature of their addiction but also by the ways
in which their personal history has inhibited their ability to acquire the conditions
necessary for autonomy. Many women with serious addictions have experienced se-
272 Autonomy and the Social

vere violence or abuse. A strong correlation between women's addictions and their
abuse has been proven in various studies (most but not all of which focus on alco-
holic women). Brenda Miller and her colleagues discovered that compared to non-
alcoholic women, it is more common for female alcoholics to have suffered child-
hood sexual abuse, emotional abuse by their fathers and spouses, and spousal
physical abuse.28 One study found that 74% of alcoholic women have experienced
sexual abuse, compared to 50% of nonalcoholic women, and a similar difference in
percentages was found for physical and emotional abuse: 52%, compared to 34%,
for physical abuse; 72%, compared to 44%, for emotional abuse. 29 Moreover, the
experiences of female alcoholics with different forms of abuse tend to be more fre-
quent and more severe than their nonalcoholic counterparts. 30 For illicit drug use,
one study found that teenaged girls who reported a history of sexual or physical
abuse used drugs more frequently than other girls. 31
What makes most forms of violence or abuse examples of oppression is that
they are so systemic that they could be defined as social practices.32 This definition is
appropriate given that these forms of oppression occur in a social environment that
makes them permissible, either explicitly or implicitly. 33 The emotional, physical,
and sexual abuse of women are not isolated problems that concern only individual
women; they are political issues because they are encouraged by sexist stereotypes of
women as inherently more passive and vulnerable than men, as primarily sexual ob-
jects, and as caregivers as opposed to care receivers.34
Most forms of abuse tend to have a negative impact: on the abused person's level
of self-appreciation. Different kinds of abuse can prevent the development of or can
destroy existing self-trust of all the three main types, resulting in the diminished au-
tonomy of the agent. Emotional abuse during childhood, as well as during adult-
hood, that involves continual criticism and labeling the victim as worthless and in-
competent can damage all types of self-trust. Experiencing frequent criticism for her
opinions and choices can cause the agent to seriously distrust her decision-making
capacities, to lack the courage to act on her own decisions, and to distrust her judg-
ment overall. Predominantly physical forms of abuse can also damage self-trust,
mostly because the victims often blame themselves for the anger or sexual desires of
their abuser. Victims of incest often think that they somehow provoked their abuse,
either because that is what their abusers tell them or because blaming themselves
seems to offer a way of reasserting control by allowing them to believe that the abuse
will stop if they behave differently in the future. 35 When the abuse continues, they
are likely to begin to profoundly distrust their judgment and decision-making ca-
pacities because every decision they made about what they needed to do to avoid the
abuse turned out to be wrong. Hence, their self-trust of Types f and ffl will be dam-
aged as a result of the abuse. Some empirical support for these claims can be found
in a small study by Doris Brothers, which shows that the greatest problems relating
to trust caused by incest and rape lie in the victim's trust in themselves.36 These
problems may not only relate to Types f and f II self-trust; in some cases, they may
also concern the agent's trust in her ability to act on her judgments, trust that she
may lack because of having been threatened with harm or violence if she were to
alter her behavior or report her abuse to anyone.
Rather than blaming themselves for the abuse, some victims alternatively block
 Relational Autonomy, Self-Trust, and Health Care 273

it out of their minds, sometimes by retreating to a fantasy world. One survivor of

incest explains that "I hid myself; I became invisible. I did whatever I could do with
my mind to leave the situation."37 It is not uncommon for incest survivors who used
this method to forget about what happened to them until years later. Their trust in
their ability to accurately recall memories of their abuse years later may diminish be-
cause of disbelief on the part of others who are close to them and because of a so-
cietal acceptance of "false memory syndrome."38 As Sue Campbell argues, those who
have encouraged the acceptance of this so-called syndrome, in particular the
founders of the False Memory Syndrome Foundation, have succeeded in making it
more unlikely than it was in the past that people will believe the testimonies of in-
cest survivors.39 The so-called syndrome implies that incest survivors, most of
whom are women, are "bad rememberers,"40 and it encourages them to distrust their
memory to a degree that could seriously limit their autonomy. Someone who be-
lieves that she is a bad rememberer will have difficulty in trusting her capacity to re-
flect rationally on her beliefs. This capacity is a necessary decision-making skill, and
it relies on the memory of the agent. Reflecting on and evaluating our beliefs often
involves determining whether our memory of the situation in which we formed
them is consistent with the beliefs themselves. Those who are taught to distrust their
memory will lack a form of Type I self-trust that is necessary for autonomy.
Not all survivors of incest and other forms of abuse suffer from diminished self-
trust and autonomy. Many of them have resisted the label "passive victim," prefer-
ring to define themselves as survivors who have succeeded in developing their own
coping strategies for abuse.41 Not all coping strategies lead to survival, however;
abused women who cope with their abuse by becoming severely addicted to harmful
substances are not yet survivors.42 Although using these substances may have been
the only strategy available to them in the short term, prolonged and severe substance
use puts them at risk of serious health problems and may worsen whatever psycho-
logical problems they developed as a result of the abuse.43 Furthermore, prolonged
use inhibits their ability to adopt less harmful coping strategies.44 Before the term
"survivor" can be accurately applied to addicted women who suffered abuse, they
need to receive treatment for their addiction that addresses not only their addictive
behavior but also the abuse and its psychological effects.
Effective women-centered treatment must attend to the fact that many female
addicts probably have diminished self-trust and, hence, reduced relational auton-
omy because of experiences with abuse that typically undermine self-trust. Those
who care for these women face the dual tasks of helping them to break free of their
addictions, as well as improving their low levels of relational autonomy by helping
them to build higher levels of self-trust. These tasks in most cases will be insepara-
ble, which makes the latter essential in addiction treatment for women. Studies have
shown that for addicts who lack confidence in their own abilities, increasing that
confidence increases their chances of successful recovery from drug or alcohol ad-
diction.45 One study performed at the Amethyst Women's Addiction Centre in Ot-
tawa, Canada, revealed that women whose self-esteem and assertiveness skills in-
crease during treatment are more likely to curtail their drinking.46 Being assertive or
confident to express your own opinions and feelings has a lot to do with trusting
your own judgment about their accuracy and relevance in discussions with others.
274 Autonomy and the Social

Thus, improving assertiveness usually requires improving self-crust; if the former is

essential in addiction treatment for women, the latter must be as well. It stands to
reason that if an addict continues to have a psychological problem such as low self-
trust that diminishes her autonomy, her ability to succeed at the extremely onerous
task of quitting a serious drug habit will be inhibited.
Our analysis of the importance of discussing relational autonomy in the context
of treatment for women with addictions and histories of abuse provides an addi-
tional, and powerful, argument against coercing these women into treatment. Forc-
ing them into treatment that will probably be ineffective (since coerced treatment
typically is ineffective) will have the likely consequence of further undermining
their already limited autonomy. Imposing treatment will increase the powerlcssness
of these addicts because all that it achieves is a further reduction in their decision-
making power. Taking this power away from them in the context of treatment for
their addictions could only be justified if the intent and the most probable conse-
quence of doing so would be to improve their level of autonomy in the long run. In-
effective treatment would not have this consequence, and thus, for addicts with low
self-trust, it is especially urgent that means other than coercion be sought to encour-
age them to escape from the compulsive nature of their addictions.47
For health-care providers to have any effect on whether addicts seek treatment,
it: is important that providers develop trusting relationships with them.48 In the ab-
sence of such relationships, addicts are unlikely to listen to encouragement from
providers or will interpret their encouragement as an attempt to harm or humiliate
them. Gaining the trust of patients requires that providers honestly display moral
concern for their well-being and competence in addressing their health-care needs.49
Addicts who are members of oppressed groups may have good reasons to doubt that
providers will have either of these qualities, reasons that may relate, for example, to
the fact that various health-care professions have tended to ignore the health-care
needs of members of oppressed groups. It is important for providers to strive to
overcome barriers to trust in their relationships with these addicts by showing them
that they are morally concerned for their welfare and are committed to providing
whatever help they can. Yet, a delicate balance is called for. Providers should not ex-
pect or encourage addicts, in particular those with abusive histories, to trust them
before the addicts have any evidence of their trustworthiness. 50 The root of many
female addicts' problems lie in having trusted someone whom they should have been
able to trust but who instead betrayed them severely.
The ability of providers to address the health-care needs of addicts with dimin-
ished self-trust and relational autonomy is largely determined by the availability of
forms of treatment that are compatible with their needs. Many feminists have ar-
gued that traditional forms of addiction treatment, like Alcoholics or Narcotics
Anonymous are male-biased,51 and the American Medical Association (AMA) has
agreed with them.52 Some feminists have argued in favor of approaches to addiction
treatment that are feminist or, in other words, that see oppression as a relevant factor
in their addictions.53 We agree with this position and would add to it that feminist
approaches should be informed by a theory of self-trust that is feminist or, in other
words, that explains how oppression could be relevant to self-distrust. 54 These pro-
grams must pay attention to the ways in which different forms of oppression, in par-
 Relational Autonomy, Self-Trust, and Health Care 275

ticular violence and abuse, can undermine self-trust; more positively, they also must
develop ways to promote self-trust.
The environment and some of the group sessions in existing feminist programs
seem to promote self-trust to some degree already. Many women who have been
through the Amethyst program say that they feel more confident about their own
judgment (and therefore have built greater Type III self-trust) because of the "egali-
tarian" environment at the center. There women are comfortable in expressing their
own opinions because they know that others will listen to them and take them seri-
ously, an experience that some Amethyst clients have never had before.55 Amethyst
also conducts a special session on sexual abuse; one woman reports that from this
session she learned that having been abused was not her fault and that knowing this
has given her "the strength and courage [she needs] ro be a survivor."56 Gaining that
strength and courage probably translated into greater Type II and Type III self-trust;
learning about the dynamics of abusive relationships can help abused persons realize
that what they were led to believe or feel about their own judgment was unfounded
and that they should be more confident about acting on their judgments.57 As well
as giving them the opportunity to explore these dynamics, it may also be helpful for
women with low self-trust to explore how oppression may have shaped their values
and to consider whether adopting nonsexist, nonracist (and other antioppression)
values would make them more comfortable with their decisions. Another helpful
method for improving self-trust might involve giving women educational and em-
ployment opportunities that allow them to develop autonomy skills and prove their
competency to themselves. Most female addicts have lacked these opportunities in
the past to a greater extent than male addicts.58
Because oppression is morally objectionable and its continuing existence threat-
ens to undermine the ability of its victims to develop self-trust, feminist treatment
centers should also work along with other groups to try to eliminate oppression in
the lives of female addicts. For some addicts, however, oppression may have under-
mined their self-trust in such a profound way already that removing oppressive
forces from their lives would have little effect. As Susan Babbitt argues in Impossible
Dreams, feelings of incompetence and worthlessness can be so internalized that im-
provements to one's social environment would do little to change them. One form
of self-distrust that may be extremely difficult to dislodge is Type III distrust in the
agent's judgments about her own decision-making capacities. Even if she has been
very successful in using these capacities in the past, if she distrusts her judgments
about them, she will be inclined to interpret each success as a fluke. No matter how
hard she reasons about their origin, she could interpret every good decision she
makes in this way. When psychological damage caused by oppression is that severe,
it may be unreasonable to expect addiction counselors to heal it. Still, it is possible,
though certainly not guaranteed, that a supportive and loving environment in group
sessions where people trust one another will heal such damage. What addiction
counselors can do is figure out what sort of group dynamics can have that effect and
try to reproduce them in future group sessions. What they need to do minimally is
to understand the depth of the problem, lest they blame the clients for their own
sense of frustration from ineffective treatment.
Treating any patient whose autonomy and self-trust are reduced because of her
276 Autonomy and the Social

oppression is a complex matter. It must begin with understanding the political na-
ture of oppression and recognizing the importance of finding ways to empower pa-
tients by helping to restore their autonomy, in addition to dealing with their physi-
cal symptoms. Much of this work is beyond the scope of health-care providers;
it requires broadscale social and political change. Health care by itself cannot, of
course, correct all of the evils of oppression. It cannot even cure all of the health--
related effects of oppression. If health-care providers are to respond effectively to
these problems, however, they must understand the impact of oppression on rela-
tional autonomy and make what efforts they can to increase the autonomy of their
patients and clients. We have argued that this work must include efforts to help pa-
tients develop or strengthen their trust in themselves.

Notes

1. We wish to thank the editors, Catriona Mackenzie and Natalie Stoljar, for their very
insightful and detailed comments on an earlier draft of this article. Their guidance has been
invaluable. Much of our thinking on relational autonomy emerged through Susan's partici-
pation in the Feminist Health Care fit hies Research Network, supported by a Strategic Re-
search Network Grant from the Social Sciences and Humanities Research Council of
Canada. Carolyn completed her work on this article while on doctoral scholarships from the
Social Sciences and Humanities Research Council of Canada and the Izaak Walton Killam
Memorial Foundation.
2. We focus specifically on the ways in which autonomy is used in discussions of health-
care ethics. We concentrate, particularly, on the approach of Tom I,. Beauchamp and James
F. Childrcss, Principles of Biomedical Ethics, 4th ed. (New York: Oxford University Press,
1994), because it is widely seen as the most influential account in the discipline.
3. For example, Beauchamp and Childrcss say: "We analyze autonomous action in
terms of normal choosers who act (1) intentionally, (2) with understanding, and (3) with-
out controlling influences that determine their action' (ibid., p. 123). They do not dwell on
the nature of such "controlling influences" but cite examples that deal with compulsion and
coercion.
4. For analysis of the concept of oppression, see Iris Marion Young's Justice and the Pol-
itics of Difference (Princeton, N.J.: Princeton University Press, 1991).
5. This is a central issue in Susan Babbitt's Impossible Dreams: Rationality, Integrity, and
Moral Imagination (Boulder, Colo.: Westview, 1996).
6. For further development of this argument, see Susan Shcrwin, "A Relational Ap-
proach to Autonomy in Health Care," in The Politics of Women's Health: Exploring Agency
and Autonomy, The Feminist Health Care Ethics Research Network, coordinator Susan Sher-
wiu (Philadelphia: Temple University Press, 1998).
7. See Marilyn Frye's The Politics of Reality: Essays in Feminist Theory (Freedom, Cal.:
Crossing Press, 1983).
8. See Susan Sherwin, No Longer Patient: Feminist Ethics and Fietilth Care (Philadelphia:
Temple University Press, 1992).
9. Diana Tietjens Meyers, Self, Society, and Personal (Choice (New York: Columbia Uni-
versity Press, 1989).
10. See Paul Benson, "Free Agency and Self-Worth," The journal of Philosophy 91
(1994): 650-668; Robin Dillon, "Toward a Feminist Conception of Self-Respect," Hypatia
7 (Winter 1992): 52-69; and Babbitt, Impossible Dreams.
 Relational Autonomy, Self-Trust, and Health Can 277

11. In "Self-Trust, Autonomy, and Self-Esteem," Hypatia 8 (Winter 1993): 99-120,

Trudy Govier draws a connection between self-trust and autonomy, but she does not explore
the impact of oppression on self-trust, as we do later.
12. The relation works the other way as well: developing self-trust requires some degree
of autonomy. If we are not autonomous beings, we cannot know that the reasons we have for
believing or feeling that we are competent and that we have good judgment relate to our own
successes rather than successes for which we are not responsible. In Self-Trust: A Study of Rea-
son, Knowledge, and Autonomy (New York: Oxford University Press, 1997), Keith Lehrer
agrees that autonomy is necessary for self-trust, but he fails to see that the opposite is also the
case. He argues that trusting ourselves involves knowing that we prefer and accept things that
are worth accepting and preferring. He points out that if every evaluation we make is "im-
posed or fortuitous," we not only do not know that we are the ones preferring and accepting
what is valuable or correct but also we cannot even trust our own judgment about what is
valuable or correct in the first place (see especially p. 95).
13. Benson, "Free Agency," and Dillon, "Toward a Feminist Conception."
14. See Meyers, Self, Society, and Personal Choice, especially part 2.
15. Benson, "Free Agency," p. 664.
16. Govier explains in "Self-Trust, Autonomy" that self-trust can be merely situational.
She contrasts situational self-trust with what she calls "core" self-trust (pp. 112-114).
17. For discussions of the subtle workings of oppression, see Sandra Lee Bartky, Femi-
ninity and Domination: Studies in the Phenomenology of Oppression (New York: Routledge,
1990), especially "Shame and Gender," pp. 83-98, and Robin Dillon's "Self-Respect: Moral,
Emotional, Political," Ethics 107 (January 1997): 226-249.
18. See Annette Baier, Moral Prejudices: Essays on Ethics (Cambridge, Mass.: Harvard
University Press, 1995), p. 10; Richard Holton, "Deciding to Trust, Corning to Believe," Aus-
tralasian Journal of Philosophy 72 (March 1994): 63-76; Karen Jones, "Trust as an Affective
Attitude," Ethics 107 (October 1996): 4-25; Trudy Govier, "Is It a Jungle Out There? Trust,
Distrust and the Construction of Social Reality," Dialogue33 (1994): 237-252.
19. Jones, "Trust as an Affective Attitude," p. 4.
20. For example, in "Self-Respect," Dillon argues that an important part of self-respect
is "appreciating" (or trusting) our capacity to live up to expectations we have set for ourselves.
Benson discusses the feeling free agents have of being "worthy to act," and at least part of
what he means here is that they have trust in their "competen[cy] to answer for [their] con-
duct" ("Free Agency," p. 660).
21. Determining whether in fact self-trust and self-respect are distinct is more com-
plicated than these comments suggest. The issue is complicated because the philosophical
literature refers to different forms of self-respect (and they are reviewed by Dillon, "Self-
Respect," pp. 228-232). Self-trust would have to be distinguished from each of these
forms.
22. Holton, "Deciding to Trust," p. 71.
23. See Sherwin, No Longer Patient.
24. Of course, we recognize that some exercise of paternalism will be appropriate on
occasion, but only in very limited circumstances. We do not want to deny the need for pa-
ternalism in classic cases of emergency care, for example. We do want to insist, though, that
oppression should never be taken as license to presume consent that would not be presumed
of other patients in comparable emergency situations.
25. See Anna C. Mastroianna, Ruth Faden, and Daniel Federman, eds., Women and
Health Research: Ethical and Legal Issues of Including Women in Clinical Studies, 1 vols. (Wash-
ington, D.C.: National Academy Press, 1994).
26. The use of cosmetic surgery to promote oppressive norms is not restricted to gen-
278 Autonomy and the Social

der norms. Cosmetic surgeons also help transform people of Asian or African origin so that
they appear more Caucasian, a more highly valued look in many parts of the world.
27. See James Foulks, "Should the Treatment of Narcotic Addiction be Compulsory?" An-
nals of the Royal College of Physicians and Surgeons 13 (July 1980): 232-239. Theorists about ad-
diction treatment tend to focus now on ways to enhance the motivation of addicts to enter into
treatment on their own. See Mark Schuckit, Drug and Alcohol Abuse: A Clinical Guide to Diag-
nosis and Treatment, 3rd cd. (New York: Plenum Medical Book Co., 1989), p. 263, and W. R.
Miller, "Increasing Motivation for Change," in Handbook of Alcoholism Treatment Approaches: Ef-
fective Alternatives, ed. R. K. Hester and W R. Miller (New York: Pcrgamon, 1990), pp. 67-80.
28. See B. A. Miller, W. R. Downs, D. M. Gondoli, and A. Keil, "The Role of Childhood
Sexual Abuse in the Development of Alcoholism in Women," Violence and Victims 2 (1 987):
157-172; W. R. Downs, B. A. Miller, and D. M. Gondoli, "Childhood Experiences of Parental
Physical Violence for Alcoholic Women as Compared with a Randomly Selected Household
Sample of Women," Violence and Victims 2 (1987): 225-240; and B. A. Miller, W. R. Downs,
and D. M. Gondoli, "Spousal Violence among Alcoholic Women as Compared to a Random
Household Sample of "Women" Journal of Studies on Alcohol 50 (1989): 533-540.
29. S. S. Covinton and J. Kohen, "Women, Alcohol, and Sexuality," Advances in Alco-
hol and Substance Abused (1984): 41-56.
30. Ibid.; Miller et al., "Spousal Violence."
31. M. Bayatpour, R. D. Wells, and S. Holford, "Physical Violence and Sexual Abuse as
Predictors of Substance Use and Suicide among Pregnant Teenagers," The Journal of Adoles-
cent Health 13 (March 1992): 128-132.
32. See Young, Justice and the Politics of Difference, p. 62.
3.3. Ibid., p. 61.
34. Abuse or violence experienced by men who are not members of marginalized
groups is not inherently political. Although their abuse may cause similar psychological dam-
age as that motivated by oppressive social stereotypes, the underlying causes of the damage in
each case differ.
35. See Govier, "Self-Trust, Autonomy"; William May, "The Molested," Hastings Cen-
ter Report (May—June 1991): 9-17; Diane Lepine, "Ending the Cycle of Violence: Over-
coming Guilt in Incest Survivors," in Healing Voices: Feminist Approaches to Therapy with
Women, ed. Toni Ann Laidlaw, Cheryl Malmo, and associates (San Francisco: Jossey-Bass,
1990), pp. 272-287; Lena Dominelli, "Betrayal of Trust: A Feminist Analysis of Power Re-
lationships in Incest Abuse and its Relevance for Social Work Practice," British Journal of 'So-
cial Work 19 (1989): 291-307.
36. See Doris Brothers, "Trust Disturbances in Rape and Incest Victims," Ph.D. disser-
tation, Yeshiva University, New York, 1982; cited in Govier, "Self-Trust, Autonomy," pp.
99-101.
37. Lepine, "Ending the Cycle of Violence," p. 284.
38. For a definition of this syndrome, see Sue Campbell's "Women, 'False' Memory,
and Personal Identity," Hypatia 12 (Spring 1997): 51-82, especially pp. 69, 70.
39. Ibid., p. 74.
40. Ibid.
41. See Dominelli, "Betrayal of Trust," p. 303.
42. It is important to point out that it has not been shown that all women who have ad-
dictions and have been abused develop their addiction after their abuse. Their addiction may
have come first, and it also may have been a factor in their abuse. Female alcoholics, for ex-
ample, may be more susceptible than other women to sexual abuse because they tend to be
seen as "sexually loose" (Miller et al., "Spousal Violence," p. 538).
43. This point is made by Colleen Hood, Colin Mangham, and Don McGuire, "De-
 Relational Autonomy, Self-Trust, and Health Care 279

tailed Analysis of Literature Pertaining to Substance Use and Mental Health," draft prepared
for Health Canada (March 1995), pp. 21, 24.
44. Ibid., p. 59.
45. Stanton Peele, "What Works in Addiction Treatment and What Doesn't: Is the Best
Therapy No Therapy?" InternationalJournal of'the Addictions 2^ (1990-1991): 1409-1419;
Amethyst Women's Addiction Centre, "Here's to You Sister. Creating a Women's Addiction
Service: Amethyst's Story" (Ottawa: Amethyst Centre, 1997).
46. Ibid., p. 26.
47. The connection between autonomy and self-trust may not be relevant to all addicts
because not all of them suffer serious problems with autonomy or self-trust. Among those
who do not are most coffee drinkers, smokers, and people who have fairly mild addictions.
Lack of autonomy is not a serious issue for people whose addictive behavior does not impair
their ability to perform daily tasks.
48. See Sissela Bok's reply to a letter to the editor by Dr. Quentin Regestein on Bok's ar-
ticle, "Informed Consent, Deception, and Discovering Drug Abuse" Journal of the Ameri-
can Medical Association 268 (12 August 1992): 790, 791.
49. See Caroline Whitbeck, "Trust," The Encyclopedia ofBioethics, 2nd ed. (New York:
Macmillan, 1995), pp. 2499-2504.
50. See Lepine, "Ending the Cycle of Violence," especially p. 275.
51. This criticism is described in Iris Marion Young, "Punishment, Treatment, Em-
powerment: Three Approaches to Policy for Pregnant Addicts," in Expecting Trouble: Surro-
gacy, Fetal Abuse, and New Reproductive Technologies, ed. Patricia Ann Poling (Bloomington:
Indiana University Press, 1992), pp. 109-134; J. Yaffe, J. M. Jenson, and M. O. Howard,
"Women and Substance Abuse: Implications for Treatment," Alcoholism Treatment Quarterly
13 (1995): 1-15; and Ann Abbott, "A Feminist Approach to Substance Abuse Treatment and
Service Delivery," Social Work in Health-care 19 (1994): 67-83.
52. AMA, "Legal Interventions during Pregnancy: Court-Ordered Medical Treatments
and Legal Penalties for Potentially Harmful Behavior by Pregnant Women," Journal of the
American Medical Association 264 (28 November 1990): 2663-2670.
53. Among these feminists are Abbott, "Feminist Approach to Substance Abuse Treat-
ment"; Young, "Punishment, Treatment, Empowerment"; and the authors of Amethyst Cen-
tre's "Here's to You Sister."
54. There are already at least two feminist programs in Canada—the Amethyst pro-
gram in Ottawa and the Matrix program in Halifax. Both offer services for promoting self-
esteem and dealing with sexual and other forms of abuse, and as we discuss below, these serv-
ices may help their participants develop greater self-trust.
55. Amethyst Centre, "Here's to You Sister," pp. 6, 24.
56. Ibid., p. 35.
57. However, some female addicts should not be encouraged to feel more confident
about at least certain aspects of their judgment during treatment. Whereas many female ad-
dicts were abused themselves, some also participated in abuse, as the following statistics re-
veal: over 95% of child sexual abusers are men (Dominelli, "Betrayal of Trust," p. 299), but
almost 50% of child physical abusers are women, and a relevant factor in a large percentage
of child physical abuse cases is alcoholism; see Linda Gordon, Heroes of Their Own Lives: The
Politics and History of Family Violence, Boston 1880-1960 (New York: Viking, 1988), pp.
173-175. Addiction counselors face the challenge of encouraging some women to increase
their trust in their own judgment without letting them off the hook for having exercised very
bad judgment in the past.
58. Judith Grant, "The Women and Substance Abuse Project: The Invisible Problem,"
Final Report submitted to Health Canada, Moncton, New Brunswick, February 1997, p. 8.
 12

RELATIONAL AUTONOMY AND

FREEDOM OF EXPRESSION1

Susan J. Brison

Introduction
A currently prevalent defense of the right to freedom of expression maintains
that it rests on the right to autonomy. 2 On this account, governmental restric-
tions on an individual's speech and on a person's access to another's speech violate
the person's right to autonomy. 3 In recent years, the autonomy defense of free
speech has frequently been invoked in arguments against restrictions on hate
speech,4 although it is rarely made clear in these arguments what is meant by "au-
tonomy." In an earlier article, "The Autonomy Defense of Free Speech,'"' I exam-
ined six accounts of autonomy in the philosophical literature and argued that
none yields a satisfactory argument against restrictions on hate speech and that
five fail as accounts of autonomy. I suggested that the sixth—a relational account
of autonomy — succeeds as an account of autonomy but does not yield the con-
clusion that an autonomy-based defense of free speech precludes restrictions on
hate speech.
In this article I develop and defend such a relational account of autonomy, one
based on Amartya Sen's account of capability, 6 and I discuss the ways in which it
can, and cannot, be used to defend the right to free speech. I focus particularly on
the application of this relational account of autonomy to recent controversies over
hate speech, and I argue that although this relational account helps to explain why
the right to speak and to receive others' speech is important, it does not yield a cie-

280
 Relational Autonomy And Freedom of Expression 281

fense of the view that speech is special, requiring greater justification for its regula-
tion than is needed for the regulation of other conduct.
I define "hate speech" for purposes of this article as speech that vilifies individ-
uals or groups on the basis of such characteristics as race, sex, ethnicity, religion, and
sexual orientation and that (1) directly assaults its target(s), (2) creates a hostile en-
vironment, or (3) is a kind of group libel.7 Some forms of pornography are put into
the category of hate speech by those who consider it to be, as Susan Brownmiller put
it, "the undiluted essence of antifemale propaganda."8 Certainly not all that is la-
beled pornographic counts as hate speech against women. But much pornography
presently on the market—especially the extremely violent and degrading variety
that has been on the rise over the past decade—does count as hate speech when it
is used to vilify individuals or groups by means of assaultive speech, the creation of
a hostile environment, or group libel.9
The disjunctive definition of hate speech used here is based on definitions
found in the ordinances of municipalities and on university campuses. The first dis-
junct, defining hate speech as directly targeted assaultive vilification, uses the so-
called "fighting words" doctrine and was employed in the following code adopted by
Stanford University:
Speech or other expression constitutes harassment by vilification if it:
a) is intended to insult or stigmatize an individual or a small number of individu-
als on the basis of their sex, race, color, handicap, religion, sexual orientation, or
national and ethnic origin; and
b) is addressed directly to the individual or individuals whom it insults or stigma-
tizes; and
c) makes use of "fighting words" or non-verbal symbols [which] are commonly
understood to convey direct and visceral hatred or contempt. . . .10

The second disjunct, defining hate speech as hostile environment harassment is

based on a University of Michigan code regulating "[a]ny behavior, verbal or physi-
cal, that stigmatizes or victimizes an individual on the basis of race, ethnicity, reli-
gion, sex, sexual orientation, [etc.], and that. . . creates an intimidating, hostile, or
demeaning environment. . . ."" The preceding is a partial definition taken from the
University of Michigan Policy on Discrimination and Discriminatory Harassment,
a policy that was ruled unconstitutional in Doe v. University of Michigan. This defi-
nition was based on existing sexual harassment law, prohibiting "hostile environ-
ment" harassment.12
The third disjunct, which characterizes hate speech as a kind of group libel, is
modeled on criminal libel law, as well as on the tort of defamation. It holds that
speech counts as hate speech if it (a) vilifies individuals or groups on the basis of sex,
race, color, handicap, religion, sexual orientation, or national and ethnic origin and
(b) harms the reputation of individuals or groups because of their sex, race, color,
handicap, religion, sexual orientation, or national and ethnic origin. In Beauharnais
v. People of the State of Illinois, the U.S. Supreme Court upheld an Illinois group
libel law making it "unlawful for any person, firm or corporation to manufacture,
sell . . . or exhibit . . . any publication [which] portrays depravity, criminality, un-
chastity, or lack of virtue of a class of citizens, of any race, color, creed or religion
282 Autonomy and the Social

[and thereby] exposes [them] to contempt, derision, or obloquy or which is produc-

tive of breach of the peace or riots."13 Although critics of Beaubarnais consider its
ruling on group libel laws to have been implicitly overturned by New York Times Co.
v. Sullivan,H it has never been officially overruled by the Court and some legal the-
orists still consider it to be good law.15 Recent court opinions—while conceding
that hate speech, including some pornography, can cause considerable harm, for ex-
ample, by undermining equality of opportunity—have struck down campus and
municipal hate speech codes as well as an antipornography ordinance. 16
Some opinions have claimed that the right to free speech outweighs the right to
equality, without providing arguments for this claim. A number of influential liberal
political philosophers and legal theorists, however, have argued that hate speech is and
ought to be constitutionally protected, and they have defended their view by invoking
a fundamental right to autonomy, which restrictions on such speech would violate.
The autonomy defense of free speech is typically presented as a nonconsc-
quentialist defense. A number of consequentialist defenses of free speech have also
been given: the argument from truth, the argument from diversity, the argument
from democracy, the argument from distrust, the argument from tolerance, the
pressure release argument, and the slippery slope argument. Mixed views combin-
ing two or more of these defenses have also been defended. I discuss these defenses
elsewhere,17 arguing that none succeeds as a defense of free speech and that none
would preclude restrictions on hate speech. The autonomy defense is supposed to
have the advantage over consequentialist accounts of showing why the right to free
speech is immune to being balanced against other interests, but as I have argued,
not all of the versions of the autonomy defense are nonconsequentialist, and even
those that are do not show why restrictions on hate speech should be considered
impermissible.18
A major problem with the legal literature on autonomy and free speech is that it
frequently does not define the term "autonomy," with the result that the autonomy
defense of free speech, in some of its versions, can be difficult to explicate and to
evaluate. There are, however, hints of at least six different philosophical accounts of
autonomy in the free speech literature. I have argued that five of these are unsatis-
factory as accounts of autonomy and that they also fail to yield a defense of free
speech that explains why hate speech should be protected. 19 These accounts are (1)
autonomy as negative liberty,20 (2) autonomy as moral independence,21 (3) auton-
omy as a constraint (labeled the Millian principle) on the kinds of justification that
can be used for governmental interference, 22 (4) autonomy as a good that consists
of the actual ability to be rationally self-legislating, 23 and (5) autonomy as self-
fulfillment or self-realization.24 I mentioned a sixth theory of autonomy, a relational
account,25 claiming that, although this is more plausible, it does not explain why
speech should be given special protection. In the next two sections of this article, I
elaborate on and defend these claims.

Relational Autonomy—A Capability-Based Account

The theory of autonomy f consider the most defensible is one kind of relational ac-
count of autonomy that holds that an agent's autonomy is dependent on her having
 Relational Autonomy and Freedom of Expression 283

an adequate capability to function, where "capability" is denned, following Sen, as

"the various combinations of functionings (beings and doings) that the person can
achieve."26 In Sen's account, "Living may be seen as consisting of a set of interrelated
'functionings,' consisting of beings and doings."27 The well-being of a person is de-
pendent on functionings that range from "such elementary things as being ade-
quately nourished, being in good health, avoiding escapable morbidity and prema-
ture mortality, etc., to more complex achievements such as being happy, having
self-respect, taking part in the life of the community, and so on."28
I, like Sen, intentionally leave incomplete the specification of a capability set
adequate for autonomy since any such specification, at this time, is bound to be too
culturally specific. Although Martha Nussbaum has attempted to list the function-
ings necessary for (1) having a human life and (2) having a good, or flourishing,
human life,29 her lists, with their strong emphasis on the "separateness" of human
individuals, are certainly controversial and arguably overly "Western" in orientation.
However, I am not yet convinced that the specification of a capability set adequate
for autonomy is impossible. Perhaps it could eventually be drawn up, after extensive
cross-cultural dialogue, by arriving at an "overlapping consensus" among all cul-
tures.30 Or, alternatively, there may be no such consensus, and the best we can do is
to come up with a cluster concept of an adequate capability set such that possession
of a certain number of functionings in the set is sufficient to establish a family re-
semblance among various culturally contingent capability sets.31 For the time being,
I assume that a noncontroversial, minimal list of functionings specified in a capabil-
ity set adequate for autonomy includes not only "such elementary things as being
adequately nourished, being in good health, avoiding escapable morbidity and pre-
mature mortality, etc.," as Sen suggests, but also avoiding such autonomy-hindering,
brain-disordering psychological and physiological states as extreme fear or anger, de-
bilitating depression, and shattering pain.32
My view is that one has to be capable of achieving some—yet to be specified—
set of functionings in order to be autonomous (and that one can be more or less au-
tonomous depending upon one's capability set). It is the capability, rather than the
achieved functionings, that is necessary for autonomy since one could autono-
mously choose not to achieve a functioning one is capable of achieving, for exam-
ple, by choosing to fast when it is an option to eat. From the standpoint of auton-
omy, this is vastly different from starving. A person's capability set, on this view, is
determined in large part by her relations to others and the functionings they enable
or prevent her from achieving. On this theory, whether and to what extent we are
autonomous depends on our relations to other people. This is so in at least three dif-
ferent senses.
First, we are, in Annette Baier's term, "second persons," dependent on others to
enable us to develop and sustain the competencies required for autonomy.33 Auton-
omy is thus causally relational—that is, it comes about (or fails to be brought about)
as a result of relations with others in society. It can also be seen to be constitutively
relational, in that it requires the right sorts of ongoing relations with others for it to
be sustained. Second, to make autonomous choices, as I argue below, we need to
have a range of significant options to choose from, and so we need to live in a soci-
ety that makes these possible. Our personal, familial, social, political, and economic
284 Autonomy and the Social

relations with others are what enable or inhibit our access to a range of significant
options. Third, to perceive the availability of these options and to recognize their
achievability by us, we need to live in a culture in which the norms and expectations
do not preclude such recognition.
I his account of autonomy is, thus, more substantive and normative than most
in the literature. However, I share with Harry Frankfurt and Gerald Dworkin the
view that an autonomous person is (to the extent that she is autonomous) able to
govern herself—to choose which desires to act on—and this requires the ability to
reflect on and/or endorse her desires. On the Frankfurt-G. Dworkin view, auton-
omy is an internal relation (between a higher and a lower self or between second-
order desires and first-order desires). The relation is one of rational self-legislation.
Frankfurt's account stresses the necessity of the autonomous agent's endorse-
ment of or identification with her action. 5 '' Since one way of achieving endorse-
ment, however, is simply to modify one's second-order desires in order to bring them
into line with one's first-order desires (which may still seem to us, intuitively, to be
nonautonomously chosen), a preferable account is that of G. .Dworkin, who counts
as crucial to autonomy, not the agent's endorsement of the first-order desires acted
on, but rather the agent's ability to critically reflect on and revise or reject first-order
desires.35
However, a problem G. Dworkin's view of autonomy shares with Frankfurt's
(and the variant on it defended by Baker and Redish) is that it is ahistorical, requir-
ing us to look only at an internal relation within an agent at the time of choosing or
acting to determine whether the choice or act is autonomous. A number of critics
have raised objections to this aspect of the Frankfurt-G. Dworkin account of au-
tonomy.36 The "critical reflection" condition, to have any motivation, requires a his-
torical account of the formation of higher-order desires. Some conditions under
which higher-order desires are formed arc autonomy undermining and others are
autonomy enhancing. So the capacity for higher-order evaluation must be devel-
oped in the right way for it to yield autonomous decisions. John Christman at-
tempts to specify this in a content-neutral way (by requiring that the agent be able
to later endorse the means by which the capacity was developed), as does Diana
T. Meyers (by requiring that the agent have and successfully employ "autonomy
competency—the repertory of coordinated skills that makes self-discovery, Self-
definition, and self-direction possible").37
My account of autonomy is, like Christman's and Meyers's, a historical one, in
that it takes into acccount socialization and other forms of conditioning, specifying
some conditions on the formation of higher-order desires.38 If, for example, one has
been socialized, in large part as a result of others' speech, to expect very little of her-
self or to defer to others, she is hardly in a position to make autonomous choices.
To determine whether or not someone is autonomous or is choosing autonomously
now, we need to know how that person came to have her preferences, including
those leading her to the present choice. We need to know whether she developed the
competencies necessary for autonomous preference formation and ranking.39
The historical account of the development of the competencies necessary for
the exercising of autonomy provides a helpful elaboration of the first sense in which
the capability account of autonomy is relational, that is, the sense in which we are
 Relational Autonomy and Freedom of Expression 285

"second persons." We develop the capacity for autonomy only after considerable in-
teraction with others (parents, teachers, etc.) and socialization into the language,
norms, and other aspects of a culture. As second persons, we also require for au-
tonomous personhood the right sorts of ongoing relations with others. Since some
forms of conditioning, socialization, and ongoing interactions with others are
autonomy undermining, whereas others are autonomy enhancing, the degree to
which one is able to be autonomous depends on one's past and present relations to
others.
The second sense in which the capability account of autonomy is a relational
one is that it takes into account not only the ways in which secondary desires get
formed but also the range of important options available to people. As Joseph Raz
argues, "If having an autonomous life is an ultimate value, then having a sufficient
range of acceptable options is of intrinsic value, for it is constitutive of an au-
tonomous life that it is lived in circumstances where acceptable alternatives are pres-
ent."40 The ruling idea behind what Raz calls "the ideal of personal autonomy" is
"that people should make their own lives."41 On this view, autonomy admits of de-
grees, and one is more or less autonomous according to, among other things, the
range of goals one is aware of having as real options. As Raz puts it: "The au-
tonomous person is a (part) author of his own life. The ideal of personal autonomy
is the vision of people controlling, to some degree, their own destiny, fashioning it
through successive decisions throughout their lives."42 The sheer number of options
is not what is relevant here. What matters is that there be a wide enough range of
available significant options to yield an adequate capability set. To see if there is such
a range, one must determine which options count as significant and what range of
these options is necessary for autonomy. In this way, the capability account of au-
tonomy is a substantive—and an explicitly normative—one, and thus differs from
the purely procedural accounts of Christman and Meyers. If one has an inadequate
range of significant options to choose from, one's autonomy is diminished, and the
extent to which significant options are available to someone depends on the kind of
society she lives in, as well as on her more personal relations with others.
The third sense in which a capability account of autonomy is relational is that it
is a person's ability to recognize significant options as options for her that is crucial,
and this also depends on others. For example, the capability theory of autonomy
takes into account the phenomenon of adaptive preference formation, discussed by
Jon Elster and by Sen, among others.43 The sour grapes phenomenon (in which
choosers eliminate or fail to form preferences that can't be satisfied because their ob-
jects are unattainable) is very common, especially among those historically deprived
of equality of opportunity.
A capability account of autonomy requires not only a historical approach to
studying the process of preference formation but also a normative specification of
the capability set essential to autonomy. The two can be seen to be linked if we look
at cases in which people's expectations are diminished, relative to what we think they
ought to be, because of entrenched inequalities (legitimized by social norms) and
long-term deprivations. As Sen observes: "a thoroughly deprived person, leading a
very reduced life, might not appear to be badly off in terms of the mental metric of
desire and its fulfilment, if the hardship is accepted with non-grumbling resigna-
286 Autonomy and the Social
44
tion.' Such resignation can be seen to be a rational response to a situation one
cannot change. In such cases, "'prudential reasoning would suggest that the victims
should concentrate their desires on those limited things that they can possibly achieve,
rather than fruitlessly pining for what is unattainable." 45 As Sen notes, the problem
of entrenched deprivation "applies particularly to the differentiation of class, com-
munity, caste, and gender."'1'1 1 would add that the processes by which depriva-
tions become entrenched and expectations are lowered are largely ones that involve
speech, that is, the social and cultural dissemination and reinforcement of, for ex--
ample, racial and gender norms.

Relational Autonomy and Hate Speech

What: does a capability account of autonomy have to do with freedom of expression
in general and hate speech in particular? 47 One needs access to others' speech to
know about one's available options. One also needs to be exposed to alternatives to
propaganda and indoctrination. In addition, one needs to have some say about cul-
tural norms to achieve the kind of self-definition necessary for autonomy. These
considerations might lead one to suppose that unregulated speech is most conducive
to autonomy.
However, unregulated speech can also undermine autonomy by diminishing
one's capability set in critical ways. As defined at the beginning of this article, hate
speech is speech that vilifies individuals or groups on the basis of such characteristics
as race, sex, ethnicity, religion, and sexual orientation and that (1) directly assaults
its target(s), (2) creates a hostile environment, or (3) is a kind of group libel. Hate
speech of each of these three kinds can diminish one's capability set in distinctive
ways. I examine each in turn.
1. Hate speech, defined as assaultive speech or "fighting words," can diminish
one's capability set by incapacitating certain competencies required for autonomy.
Assaidtive speech—directed at an individual and made face to face—causes imme-
diate injury in the form of an uncontrollable emotional response akin to the reac-
tion to a slap in the face. It can also short-circuit reason, leading to a fearful "freeze,
flight, or fight" response, as well as to responses such as rage, humiliation, and ciis-
orientation, all of which can interfere with the ability to reason and, thus, with the
ability to think autonomously. I have discussed such effects of hate speech at length
elsewhere.48
2. Hate speech that creates a hostile environment can diminish someone's capa-
bility set by affecting her self-esteem, her views about her options, her beliefs about
her abilities, and the formation of her preferences. As Cass Sunstein argues, libertar-
ians and many liberals have failed to give adequate attention to the ways in which
social norms (as well as social roles and social meanings) influence preferences and
choices. "We should agree," he argues, "that social norms play a part in determining
choices; that people's choices are a function of their particular social role; and that
the social or expressive meaning of acts is an ingredient in choice."49 Of particular
relevance to debates about hate speech is the observation that "individual agents-
have little control over social norms, social meanings, and social roles, even when
they wish these to be very different from what they are."50 This should not be taken
 Relational Autonomy and Freedom of Expression 287

to be an argument for the incompatibility of autonomy with the existence of social

norms, meanings, and roles since "human beings can live, and human liberty can
exist, only within a system of norms, meanings, and roles; but in any particular form,
these things can impose severe restrictions on well-being and autonomy."51 But if
the autonomy of private individuals can be threatened or undermined by norms
over which they can, individually, have no control, this may provide justification for
state intervention to modify the norms.
One form of such intervention—motivated by respect for autonomy—might
be regulation of speech that interferes unduly with autonomous preference forma-
tion. If one grants the empirical claim conceded by the courts in the hate speech
cases mentioned earlier (that failure to restrict such speech can impair the ability of
individuals in targeted groups to act on their choices)—if the absence of restrictions
can limit individuals' employment options, their political potential, and even un-
dermine their ability to take advantage of those options that are available to them—
then autonomy, in this sense, cannot be invoked to defend such a policy unless one
adds the implausible claim that the threat to the would-be speakers' autonomy is
even greater.
Preferences are affected not only by one's options but also by one's beliefs about
the actual attainability of these options, and these beliefs are typically arrived at by
receiving others' speech. First- and second-order desires are heavily influenced, if
not instilled, by others' speech. Given that some kinds of influence are more more
likely than others to interfere, illegitimately, with the formation of desires and voli-
tions, an adequate theory of autonomy should count as autonomous those "deci-
sions reached with a full and vivid awareness of available opportunities, with refer-
ence to all relevant information, and without illegitimate or excessive constraints on
the process of preference formation."52 If we are "second persons"—not just in the
sense of having been formed by others in childhood, but also in that we continue to
be shaped and sustained by others—then others' speech to and about us and ours to
and about them are crucially important in the development and endurance of our
autonomous selves. To be arbitrarily (or methodically) deprived of the ability to
speak or of access to others' speech, when such speech is essential to autonomous
selfhood, is to have one's autonomy impeded. But it is not only the government that
can put up such impediments. As Owen Fiss, Frank Michelman, Cass Sunstein, and
others have argued, private speakers (alone or as contributors to market forces) can
interfere with others' speech as much as (or in some cases more than) the state can.53
And, as Virginia Held and Judith Lichtenberg have argued, the absence of govern-
mental regulation of speech does not always enhance access to informative speech.54
In any case, it is useful information, not misinformation or excessive (i.e., too costly
to process) information, that is needed for autonomy,55 and it is an empirical ques-
tion whether speech as unregulated by the government as it is now (of course, it is
already regulated) is as "free" and as conducive to autonomy as is would be with
more state control. It is not a question that can be decided a priori.
Judith Butler, in Excitable Speech?6 acknowledges that speech can cause harm to
individuals and groups, but she argues that this harm can be countered by linguis-
tic intervention alone (and that governmental regulation would be counterproduc-
tive). Harmful racist or other epithets can, on her account, be reappropriated by
288 Autonomy and the Social

their targets, defused by subversive usage, and turned into terms that empower
rather than degrade those so labeled. This reappropriation has occurred, for exam-
ple, with the term "queer." But this required a massive political and cultural move-
ment, extended over a period of time, during which gays and lesbians continued to
suffer because of others' use of the term. Even now, its use during a physical assault
that is motivated by homophobia can add to the harm endured by the victim.
Although Butler's discussion of the ways in which hate speech can harm its
targets is a sophisticated and welcome addition to the literature, her proposed
remedy—reappropriating the terms used in hate speech, in effect, redressing the
harms of hate speech with more speech—overestimates the extent to which indi-
vidual victims can counter verbal assaults by reappropriating the very terms that
historically have been used to degrade them. This remedy may be available to
those who (1) do not feel isolated in their victimization, (2) are not silenced by the
verbal assault, (3) are—and perceive themselves to be—members of powerful po-
litical groups, and (4) arc — and perceive themselves to be—able to work in con-
cert with others in their group to change the linguistic and cultural norms that
gave and continue to give the hate speech the power to harm. There is consider-
able debate, for example, among African Americans about whether the "n-word"
can be recuperated and given a positive spin. Although it is already used among
some African Americans as a term of endearment, there are those who claim that
it always carries a message of inferiority, no matter who uses it with whom. 57 A
further worry about the increasing acceptance of the term is that those who are not
African American may begin to consider it acceptable for them to use to label blacks,
even though in this context the word can still wound in autonomy-undermining
ways.
3. Hate speech that constitutes group libel can reduce a person's available op-
tions by affecting others' beliefs about her, as well as their behavior toward her. Libel
is written (or printed or depicted) defamation, as opposed to slander, which is spo-
ken.'58 Defamation is defined (in Prosser's classic text on torts) as communication
that "tends so to harm the reputation of another as to lower him in the estimation of
the community or to deter third persons from associating or dealing with him."59
Prosser goes on to note that "communications are often defamatory because they
tend to expose another to hatred, ridicule or contempt."60 In tort law, however, only
those who defame individuals are subject to liability: "One who publishes defama-
tory matter concerning a group or class of persons is subject to liability to an indi-
vidual member of it if, but only if, (a) the group or class is so small that the matter
can reasonably be understood to refer to the member, or (b) the circumstances of
publication reasonably give rise to the conclusion that there is particular reference
to the member."61
Richard Dclgado, in his now classic article "Words That Wound," has called
for an independent tort action for racial insults and epithets, Such a tort would,
Delgado argues, "protect the interests of personality and equal citizenship that are
part of our highest political traditions and moral values, thereby affirming the right
of all citizens to lead their lives free from attacks on their dignity and psychologi-
cal integrity."62
There is already, however, ;t cotisfinmonyl precedent for considering group libel
 Relational Autonomy and Freedom of Expression 289

to be sanctionable as criminal behavior. In 1952, the U.S. Supreme Court upheld

the constitutionality of a state law that criminalized group libel, in Beauharnais v.
People of the State of Illinois.6^ The law in question stated: "It shall be unlawful for
any person . . . to manufacture, sell. . . present or exhibit in any public place in this
state any [speech that] . . . exposes the citizens of any race, color, creed or religion
to contempt, derision, or obloquy or which is productive of breach of the peace or
riots."64 In Collin v. Smith,6') the court considered Beauharnais to be no longer au-
thoritative, in light of rulings in such cases as New York Times Co. v. Sullivan.66
However, Beauharnais has not been formally overturned, nor has the discussion of
libel of a public figure in New York Times v. Sullivan been explicitly applied by the
courts to the question of group libel. Kenneth Lasson has argued that Beauharnais is
still good law and has noted that group libel statutes are still in the criminal codes
of five states.67
Furthermore, there are grounds for rejecting the individualism that underlies
the dichotomy between private and group libel. As Isaiah Berlin wrote in his famous
essay "Two Concepts of Liberty":
Even Mill's strenuous effort to mark the distinction between the spheres of pri-
vate and social life breaks down under examination. Virtually all Mill's critics have
pointed out that everything that I do may have results which will harm other
human beings. Moreover, I am a social being in a deeper sense than, that of inter-
action with others. For am I not what I am, to some degree, in virtue of what
others think and feel me to be?68

The view that people carrbe libeled by hate speech has gotten short shrift in the
legal literature, not only because of liberal and libertarian views concerning privacy
and individualism, but also because of the dominant focus on audience-centered
views of free speech. In First Amendment jurisprudence, as Thomas Scanlon has
noted, different interests come to the fore in discussions of the various categories of ex-
pression. This has yielded the following set of approaches to free speech: the speaker-
centered (or what Scanlon calls "participant-centered") approach, the audience-
centered approach, and the bystander-centered approach (in which bystanders are
those affected by the speech, although not in the willing audience for it).69 Political
speech, it is argued, lends itself to the first approach, commercial speech more ap-
propriately falls into the second, and libel is in the third. The question remains of
whether a speaker-centered, audience-centered, or bystander-centered approach to
a theory of free speech is the most appropriate in the area of hate speech.
One might be able to use an audience-centered approach to argue for the regu-
lation of hate speech (or at least that variety of it that encourages false views about
the groups targeted by it). Such an argument would mirror rhe defense of restric-
tions on false or misleading advertising: Those individuals who try to apply these
false hate speech-induced beliefs in their relations with real people may be led to
make serious, culpable mistakes in judgment. So an audience-centered approach
could conceivably lead to arguments for restricting hate speech, as well as false or
misleading commercial speech.
A weightier argument for regulation of hate speech is one that takes seriously
the question of bystanders' interests, that is, the interests of rhose who are not in the
290 astomty and soci
willing audience for the speech bur who are nonetheless affected by it. In discussing
libel, as well as hate speed) of the group libel variety, I think it is important to divide
the category of bystanders into two categories: targets (those the hate speech is di-
rected to or about and who an-? affected by it) and bystanders (those affected by the
speech who arc not the targets of it), A speaker may have a very great interest in
writing (falsely) that Jones committed some heinous crime (if he wants to frame
Jones to get himself off the hook or if he is angiy Math Jones and feeling vindictive),
and a given audience may have a very great interest in reading this bit of libel (per-
haps they arc predisposed to believe u since they dislike Jones or Jones's race for
other reasons and this makes them feel vindicated in their prejudice, or perhaps they
are put at ease by the "fact'' that the criminal has been pinpointed).
This combination of speaker's and audience's irirerests is simply not enough to
warrant protection for die libclous speech. We have to take the target (Jones) into
account, and in this case her interests are overriding. (Joel f'einberg would say this
is so even in the case of very damaging truths about a private person that the public
simply doesn't need to know.70) In addition, the bystanders-••- that is, others not tar-
geted by the speech, not in the willing audience for it, but nonetheless affected by
it -may have interests in avoiding exposure (and reducing others' exposure) to it. In
this balancing of competing interests, a great deal of weight is given to the target's
interests, so much so that it makes sense to call the first Amendment jurispru-
dence of individual libel (of private figures) "target centered." If one accepts a target-
centered account of free speech in the area of (individual) libel, one ought: to accept
it in the area of group libel, including hate speech that takes that form, 7 1
lo see why, consider the following example, Suppose you are a black male and
someone publishes a newspaper article claiming (falsely, obviously) that black males
have a propensily to rape. Why should this not count as libelous material, damag-
ing to your interest in your character and reputation?' /7 - As the majority stated in
Bet.iuha.nMis v. Illinois,
ft would . , , be arrant dogmatism , , . to deny that . . . a man's job and his cduca-
rional opportunities and the dignity accorded him may depend as much on the
reputation of the racial and religious group to which he willy-nilly belongs, as on
his own merits. This being so, we are precluded from saying that speech conccd-
cdly punishable when immediately directed al individuals cannot be outlawed if
directed at groups with whose position and esteem in society the affiliated indi-
vidual may be inextricably involved,-'-'

However, although it lias to be shown that statements or depictions that defame

even large groups can harm individual members of those groups, this is easily done.
It is acknowledged, in discussions of affirmative action, for example, that individu-
als can suffer harms in virtue of being members of stigmatized groups. And few
would deny that defamatory statements and depictions contribute to and perpetuate
such harmful stigrnatization. It has been argued, though, that the harm to an indi-
vidual member of an attacked group cannot, at any rate, be very great if the targeted
group is large. As Feinbcrg argues: "Racist and porno films do not directly insult spe-
cific individuals, but: rather large groups, thus diluting the impact of the insult, or
at leas! its directed personal character, proportionately," l i e compares porno films
 Relational Autonomy and Freedom of Expression 291

with an invented genre of popular racist films that cater to secret haters of blacks,
the main features of which would be
stories of uppity blacks put in their place by righteous whites, taunted and
hounded, tarred and feathered, tortured and castrated, and in the climactic
scenes, hung up on gallows to the general rejoicing of their betters. The aim of
the films would be to provide a delicious catharsis of pent-up hatred. It would be
prudent, on business grounds, to keep advertisements discreet, and to use eu-
phemistic descriptions like "folk films" (analogous to "adult films").74

He goes on to argue that since the degree of insult to an individual varies inversely
with the size of the insulted group, "the 'folk films' might be more serious affronts
in this respect than the porno films since their target is a much smaller group than
half of the human race A black man might be more likely to feel a personal
grievance at the folk film he does not witness than a woman would to a porno film
she does not witness. . . ."75 How black women might appropriately react to both
genres, especially given the actual proliferation of racist pornography, is not con-
sidered.
The flaw in Feinberg's proportionality-of-injury argument is that the ease with
which the individual is identified with the group, the social standing of the group,
and the history of its treatment are all much more relevant to the degree of harm
suffered by individuals in the targeted group than is the relative size of the group. If
you are a black woman saxophone player, you are more likely to be harmed by
defamatory attacks on blacks and on women than by attacks on saxophone players.
And, in any case, as David Riesman pointed out, in an early article on group libel:
"It is hard to see why courts should be solicitous of a defendant who has hurt many
people rather than a few or one. In many political and social situations, it is likely
that the larger the defamed group—if the defamation can actually be made to stick
as 'truth'—the more likely that social dangers will ensue."76 In the case of already
stigmatized groups, defamatory attacks all too frequently "stick as 'truth.'" Members
of such groups are easily stereotyped, which enables a degrading description or de-
piction of the group to lead to diminished respect for all its members.77
A critic of my position might concede that a description of a group, say, women,
might well defame them but deny that there could be a depiction that defames all the
members of a group. For example, it might be denied that women as a class could
be defamed by a degrading, misleading, pornographic picture, photo, or film that
counts as hate speech and targets women.
To reply to this objection, I must first establish that such a defamatory depic-
tion can refer to women in general, given that the depiction itself is a particular and
is sometimes a depiction of a particular woman. Is it possible to depict a group by
means of a particular depiction or a depiction of one of its members? We know,
from Berkeley's critique of Locke's theory of abstraction,78 that representing a gen-
eral category, say, of triangles, is not a matter of constructing an abstract mental
image that somehow manages to capture all the features of particular triangles and
none of the features that aren't held in common. Rather, having the general idea of
triangularity involves being able to use an image of a particular triangle to stand for
all triangles. We frequently use pictures in this way, and not only the highly stylized
292 Autonomy and the Social

pictograms found on restroom doors and highway signs. In dictionaries and ency-
clopedias, a photograph of an individual plant or animal is frequently used to refer
to the entire species to which it belongs.
What many degrading symbols or depictions have in common with such pic-
tograms and photographs is their ability to be used as general terms. A photograprn
or a caricature of an individual woman can be seen as a depiction of women in gen-
eral.79 This is a function of how it is used, and the viewing context plays a large role
in determining the generality of the reference. In this way, depictions, as well as de-
scriptions, can libel groups of individuals.
Group libel, in addition to causing or entrenching harmful stereotypes, can also
undermine the autonomy of individuals by depriving them of a multicultural envi-
ronment that enhances opportunities for choice. As Raz argues, one needs the op-
tions afforded by a multicultural society to be fully autonomous and one also needs
to live in a society in which one's culture is not merely tolerated but respected and
encouraged. 80 Hate speech of the group libel variety can undermine multicultural-
ism, as well as lead to increased disrespect for those in the targeted cultural groups.
It is not only the autonomy of targets that is undermined by group libel. Par-
ticipant, audience, and bystander autonomy may also be diminished rather than en-
hanced by unregulated hate speech of this variety. It should be noted that those who
engage in hate speech are as much products of the cultural norms and practices
around them as are their targets. One cannot assume that they are acting au-
tonomously in uttering such speech.81 Willing audience members, such as the
young, impressionable, and socially vulnerable skinheads recruited by Tom Mer-
ger's Aryan Resistance League, may succumb to autonomy-bypassing indoctrination
techniques. Even those of us bystanders who do not use hate speech, who are not a
willing audience for it, and who strive to rid ourselves of all racist beliefs, for exam-
ple, can find unbidden and unwelcome racist epithets popping into our minds. Mari
Matsuda reports, for example, that upon seeing a Hindu woman, she found the
word "dot-buster" involuntarily entering her consciousness, instead of something
she would have chosen to think, say, "what a lovely sari."82
Thus the extent to which hate speech itself is the result of an autonomously
made choice is not clear. Although those who have given the autonomy defense of
free speech have assumed that the speaker is always speaking autonomously, this
is not always the case. As Kendall Thomas has pointed out, at least some of thos;e
engaged in hate speech, especially young persons, appear to be "channeling" the voices
of others, unthinkingly parroting their speech.83 On a capability account of au-
tonomy, if one's second-order volitions (to engage in hate speech, for example)
may be formed by indoctrination or brainwashing and not in the right (autonomy-
respecting) way, one's acted-on first-order desires are not necessarily autonomously
chosen.
Speech is nonetheless important to autonomy: it can enhance and undermine it.
Freedom of expression, when construed as a positive freedom actually to speak and to
be heard, can at times be seen to be instrumental in the development of autonomous
persons. A capability account of autonomy does not, however, yield a distinct princi-
ple of freedom of expression. That is, it docs not provide a defense of the principle
that the government requires greater reason to regulate speech than it needs to justify
 Relational Autonomy and Freedom of Expression 293

the regulation of other conduct. In particular, unregulated hate speech is not neces-
sarily autonomously chosen speech, and it can sometimes undermine its targets' ca-
pacity for autonomy by diminishing, in significant ways, their capability sets. We
need to look at the autonomy-enhancing and the autonomy-undermining effects of
such speech—on the speakers, audience, bystanders, and targets—on a case-by-case
basis, to determine whether regulating it would promote or undercut autonomy.

Notes

1. I am extremely grateful to Catriona Mackenzie and Natalie Stoljar for their encour-
agement, patience, and extraordinarily helpful comments on an earlier draft. I also thank
them, Sarah Buss, Diana T. Meyers, Yael Tamir, and Thomas Trezise for ongoing discussions
about the nature of autonomy. I am indebted to the School of Social Science at the Institute
for Advanced Study for supporting this work through a fellowship funded by the National
Endowment for the Humanities, during 1997-1998.
2. I use the term "free speech" interchangeably with "freedom of expression," and I use
rhem both to cover freedom of the press as well. I am drawing primarily on literature by free
speech theorists in the United States who assume the existence of a constitutional right to
free speech, as specified in the First Amendment. Autonomy defenses of free speech can be
found in David A. J. Richards: "Autonomy in Law," in The Inner Citadel: Essays on Individual
Autonomy, ed. John Christman (New York: Oxford University Press, 1989); "Free Speech and
Obscenity Law: Toward a Moral Theory of the First Amendment," University of Pennsylvania
Law Review 123 (1974): 45-91; "Pornography Commissions and the First Amendment: On
Constitutional Values and Constitutional Facts,"Maine Law Review 39 (1987): 275-320;
and "Toleration and Free Speech," Philosophy and Public Affairs 17 (1988): 323-336; Ronald
Dworkin: A Matter of Principle (Cambridge, Mass.: Harvard University Press, 1985), pp.
353-372; "Liberty and Pornography," New York Review of Books, 15 August 1991, pp.
12 — 15; "Women and Pornography," New York Review of Books, 21 October 1993, pp.
36-42; and "The Coming Battles over Free Speech," New York Review of Books, 11 June
1992, pp. 55-64; Thomas Scanlon, "A Theory of Freedom of Expression," Philosophy and
Public Affairs 1 (1972): 204—226, and "Freedom of Expression and Categories of Expres-
sion," University of Pittsburgh Law Review 40 (1979): 519-550; Charles Fried, "The New
First Amendment Jurisprudence: A Threat to Liberty," The Bill of Rights in the Modern State,
ed. Geoffrey R. Stone, Richard A. Epstein, and Cass Sunstein (Chicago: University of
Chicago Press, 1992); David Strauss, "Persuasion, Autonomy, and Freedom of Expression,"
Columbia Law Review 91 (1991): 334-371; Martin H. Redish, Freedom of Expression: A
Critical Analysis (Charlottesville, Va.: Michie Company, 1984); C. Edwin Baker, "Scope of
the First Amendment Freedom of Speech," UCLA Law Review 25 (1978): 964-990, and
Human Liberty and Freedom of Speech (New York: Oxford University Press, 1989); Diana T.
Meyers, "Rights in Collision: A Non-punitive Compensatory Remedy for Abusive Speech,"
Law and Philosophy 14 (1995): 203-243; Thomas Nagel, "Personal Rights and Public
Space," Philosophy and Public Affairs 24 (1995): 83-107. Kent Greenawah, who criticizes
some attempts to ground free speech in the singular value of autonomy, nonetheless appeals
to autonomy as one of "the subtle plurality of values that does govern the practice of freedom
of speech." See "Free Speech Justifications," Columbia Law Review %9 (1989): 119-155, 119.
3. Although not all of those giving the autonomy defense state this explicity, I assume
that they hold that some kinds of speech, for example, false advertising, perjury, and insider
trading, may be restricted without anyone's autonomy being violated. If a person shouted
294 autouny and pjrn
"drop dead," causing everyone within earshot to drop dead, presumably the banning of such
speech would not be prohibited by respect for autonomy.
-4, See the sources in note 1.
5. Susan J. Brison, "The Autonomy Defense of I ; ree Speech,™ Ethics 108 (January
1998): 312-339.
6. For presentations of Sen's account of equality of capability, see Amartya Sen, "Gen-
der Inequality and 'Theories of Justice," in Women, Culture, and Development, eel. Martha
Nussbaurn and Jonathan Glover (New York: Oxford University Press, 1995); Inequality Re-
examined (Cambridge, Mass.: Harvard University Press, 1992); "Well-being, Agency, and
Freedom: The Dcwcy Lectures 1984," The Journal of Philosophy 82 (April ! 985): 169- 220;
and "Rights and Agency," Philosophy and Public Affairs \ 1 (Winter 1982): 187- -223.
7. The definition that follows is similar to the one given in "Autonomy Defense of free
Speech," pp. 313-31 5, although some ol the terminology has been changed. For a selected
list of reported incidents of hate speech on college campuses from 1986 to 1 988, sec Howard
J. thrlidb, Campus Ethnoviolence and the Policy Options (Baltimore: National Institute against
Prejudice cV Violence, 1990), appendix. For discussions of numerous incidents of hate
speech through 1994, sec The Price We Pay: The (.'.use against Racist Speech, Hate Propaganda,
and Pornography, ed. Laura]. Lederer and Richard Delgado (New York: Hill & Wang, 1995).
8. Susan Brow.nmi.Uer, Against Our Will: Men, Women and Rape (New York: Bantam
Books, 1975), p. 443.
9. For numerous examples of pornography that could be construed as hate speech
under this definition, see the analyses ot pornography in Catharine M a c K i n n o n , Feminism
Unmodified: Discourses on Lift: and Laid (Cambridge, Mass.: Harvard University Press, 1987);
Only Words (Cambridge, Mass.: Harvard University Press, 1993); and Toward a Feminist The-
ory of the Stale (Cambridge, Mass.: Harvard University Press, 1993); Attorney General's
Commission on Pornography, Uintd Report, U.S. Department of Justice, ]tily 1986; Laura
Lederer, ed., Take Back the Night: Women on Pornography (New York: William Morrow,
1980); Catherine Itzen, ed,, Pornography: Women, Violence and Civil Liberties (New York: Ox-
ford University Press, 1992); and Lederer and Delgado, Price We Pay.
10. Quoted in Charles R. Lawrence 111, "If He I toilers Let Him Go: Regulating Racist
Speech on Campus," in Mari J. Matsuda et al., Words That Wound: ('.rtlical Race Theory, As-
saultive Speech, and the 1'irst Amendment (Boulder, Colo.: Westview Press, 1993), p. 67.
"Fighting words" were originally defined by the U. S. Supreme Court as words or symbols
"which by their very utterance inflict injury or tend to incite an immediate breach of the
peace." Chaplimky v. State of New Hampshire, 31 5 U.S. '568 (1942), 572. For a discussion of
the fighting words doctrine, see Susan j. Brison, "Speech, Harm, and the Mind-Body Prob-
lem in First Amendment Jurisprudence," Legal Theory 4 ( 1 998): 39-61. The Stanford code
was struck down by the court because of the Leonard Jaw, which requires even private edu-
cational institutions in California to abide by the U.S. Constitution. Carry v. Iceland Stanford
Junior University, California Superior Court: at Santa Clara, case no. 740309, 27 February
1995.
I f . Doev. University of'Michigan, 721 F.Supp. 852 (E.D.Mich. 1989), 856.
12. See Meritor Savings Bank v. Vmson, 477 U.S. 57 (1986).
1.3. Beauharnais v. People of the State of Illinois, 72 S. Ct. 725 (1952), 728.
14. New York Times Co. v. Sullivan, 376 U.S. 254 (1964).
15. Sec, for example, Kenneth Lasson, "Group Libel versus Free Speech: When Big
Brother Should Butt In," Duquesne Law Review 23 (1984): 77-130. I employ the disjunc-
tive definition of hate speech because it captures most of what has been called hate speech in
the legal literature. 1 do not claim that this is the only or the best definition. I am not argu-
ing that: it would pass constitutional muster at a public university or that it would be: good
 Relational Autonomy and Freedom of Expression 295

policy at a private one. If used in a hate speech code at a public university and then chal-
lenged in the courts, this definition would certainly raise questions about vagueness and over-
breadth, and the group libel disjunct would probably be ruled unconstitutional for lack of a
clear legal precedent. I am using this definition in an attempt to delimit the class of what
counts as hate speech to examine whether the autonomy defense of free speech explains why
this speech should be protected.
16. See Doe v. University of Michigan; UWM Post v. Board of Regents of the University
of Wisconsin, 774 F.Supp. 1163 (E.D. Wis. 1991); American Booksellers Association v. Hudnut,
111 F.2d 323 (7th Cir. 1985). Since the courts have made this concession, I am not defend-
ing this controversial empirical claim here. I discuss the harms of hate speech in "Speech,
Harm, and the Mind-Body Problem" in response to those who deny that speech can harm
and to those who attempt to make a sharp, morally and legally relevant distinction between
speech-caused harms and other harms. In this article, 1 discuss the ways in which some of the
harms of hate speech undermine autonomy.
17. Susan J. Brison, Speech, Harm, and Conflicts of Rights (Princeton, N.J.: Princeton
University Press, 2000).
18. Brison, "Autonomy Defense of Free Speech."
19. Ibid.
20. On this view, autonomy is simply freedom from governmental interference in some
specified domain. Thus a right to autonomy is a side constraint on governmental action. The
term "negative liberty" was defined by Isaiah Berlin and contrasted with "positive liberty" in
his "Two Concepts of Liberty," in Four Essays on Liberty (New York: Oxford University Press,
1969). Ronald Dworkin employs and defends this account of autonomy in "Liberty and
Pornography."
21. In this account, defended by Ronald Dworkin in Matter of Principle, to restrict peo-
ple's speech (or their access to others' speech) out of contempt for their way of life or their
view of the good violates their right to moral independence or autonomy. This amounts to
an unacceptable failure to treat them with equal respect and concern. Since rights, according
to Dworkin, trump considerations of social utility: "if someone has a right to moral inde-
pendence, this means that it is ... wrong for officials to act in violation of that right, even if
they (correctly) believe that the community as a whole would be better off if they did" (p.
359). Dworkin also uses this approach in two more recent articles: "The Coming Battles over
Free Speech" and "Women and Pornography."
22. Thomas Scanlon ("Theory of Freedom of Expression, p. 213) defines the Millian
principle as follows:
There are certain harms which, although they would not occur but for certain
acts of expression, nonetheless cannot be taken as part of a justification for legal re-
strictions on these acts. These harms are: (a) harms to certain individuals which
consist in their coming to have false beliefs as a result of those acts of expression;
(b) harmful consequences of acts performed as a result of those acts of expression,
where the connection between the acts of expression and the subsequent harmful
acts consists merely in the fact that the act of expression led the agents to believe
(or increased their tendency to believe) these acts to be worth performing.
(Although Scanlon calls this "The Millian Principle," it owes more to Kant than to
Mill.)
23. This view is defended by Scanlon in a more recent article, "Freedom of Expression,"
p. 533. This account of autonomy suggests a view of the self divided against itself—a true,
higher, sovereign self and an impulsive, lower self that occasionally needs to be brought into
line. An autonomous person is able to control herself—to choose which desires to act on —
and this requires the ability to reflect on her desires. Harry Frankfurt has argued that it is this
296 Autonomy and the Social

ability to have second-order desires concerning first-order desires that sets humans apart from
the beasts. See "Freedom of the Will and the Concept of a Person," in The Importance of
What We Care About (New York: Cambridge University Press, 1988), p. 12. Although Scan-
Ion himself does not invoke Frankfurt's account of autonomy, it is the position in the philo-
sophical literature on autonomy that appears to be closest to the view Scanlon invokes in this
later article on free speech.
24. Versions of this kind of account are advocated by Baker, "Scope of the First
Amendment," and by Redish, Freedom of Expression.
25. Influential relational accounts of autonomy include those advocated by Diana T.
Meyers, "Personal Autonomy and the Paradox of Feminine Socialization," Journal of Philos-
ophy 84 (1987): 619-628, and Self Society, and Personal Choice (New York: Columbia Uni-
versity Press, 1989); Jennifer Nedelsky, "Reconceiving Autonomy: Sources, Thoughts, and
Possibilities," Yale Journal of Laiu and Feminism \ (1989): 7 — 26; and Marilyn Friedman,
"Autonomy and Social Relationships: Rethinking the Feminist Critique," in Feminists Re-
think the Self, cd. Diana T. Meyers (Boulder, Colo.: Westvicw, 1997).
26. Sen, Inequality Reexamined, p. 40.
27. Ibid., p. 39.
28. Ibid.
29. Martha C. Nussbaum, "Human Capabilities, Female Human Beings," in Women,
Culture, and Development: A Study of Human Capabilities, ed. Martha C. Nussbaum and
Jonathan Clover (Oxford: Clarendon Press, 1995), pp. 61-104.
30. Such an account of a cross-cultural overlapping consensus is given by Charles Tay-
lor, "Conditions of an Unforced Consensus on Human Rights," in The East Asian Challenge
for Human Rights, ed. Joanne R. Bauer and Daniel A. Bell (New York: Cambridge University
Press, 1999), pp. 124-144. I thank Leslye Obiora for drawing my attention to this article.
31. This suggestion is inspired by Natalie Stoljar's work on feminism and essentialism.
For her account of "woman" as a cluster concept, see "Fissence , Identity and the Concept of
Woman," Philosophical Topics 23 (1995): 261-294.
32. Sarah Buss defends a similarly normative "human flourishing condition" on auton-
omy, arguing that it "both reflects and and illuminates the widely shared intuition that au-
tonomy can be undermined by pain, fear, depression, anger, and other 'negative' psychologi-
cal and physiological conditions." "How to Express Yourself in Your Actions," unpublished
manuscript, available from author, p. 17.
33. Annette Baicr, Postures of the Mind: Essays on Mind and Morals (Minneapolis: Uni-
versity of Minnesota Press, 1985); Lorraine Code, "Second Persons," in What Can She Know?
(Ithaca, N.Y.: Cornell University Press, 1991), pp. 71-109; and Virginia Held, Feminist
Morality (Chicago: University of Chicago Press, 1993), pp. 57-64.
34. Frankfurt, "Freedom of the Will." See also "Three Concepts of Free Action," "Iden-
tification and Externality," and "Identification and Wholcheartedness," all in Importance of
What We Care About.
35. Gerald Dworkin, The Theory and Practice of Autonomy (New York: Cambridge Uni-
versity Press, 1988); see especially pp. 15-17 for his reasons for rejecting the endorsement ac-
count.
36. For an excellent summary of discussions of the ab initio and the infinite regress
problems in such an account, see Christman, Inner Citadel, pp. 6-12.
37. John Christman, "Autonomy and Personal History," Canadian Journal of Philoso-
phy1\ (1991): 1-24; Meyers, Self, Society, and Personal Choice, p. 76.
38. For other discussions of the role of socialization in the development of the capac-
ity for autonomy, see Joseph Raz, The Morality of Freedom (New York: Oxford University
 Relational Autonomy and Freedom of Expression 297

Press, 1986), and Cass Sunstein, "Preferences and Politics," Philosophy and Public Affairs 20
(1991): 3-34.
39. For a discussion of these competencies, see Meyers, Self, Society, and Personal
Choice.
40. Raz, Morality of Freedom, p. 205; see also pp. 373-380.
41. Ibid., p. 369.
42. Ibid. Note that being an author does not require making up all of one's own lan-
guage, literary conventions, character types, plot lines, and so on, although it does require
making some choices within (or against) available options.
43. Jon Elster, Sour Grapes (New York: Cambridge University Press, 1983); Sen, In-
equality Reexamined; see especially p. 55, where Sen discusses the effects of entrenched in-
equalities and deprivations on preferences. See also Meir Dan-Cohen, "Conceptions of
Choice and Conceptions of Autonomy," Ethics 102 (1992): 221-243.
44. Sen, Inequality Reexamined, p. 55.
45. Ibid.
46. Ibid.
47. A capability account of autonomy has not, however, been explicity invoked in the
literature in defense of free speech, although Meyers has suggested, in "Rights in Collision,"
that an appeal to her relational account of autonomy, combined with the argument from
democracy, can be used to defend the right to engage in even abusive hate speech.
48. Brison, "Speech, Harm, and the Mind-Body Problem," pp. 48-50. See also Dru-
cilla Cornell's discussion of encounters with unavoidable public displays of pornography as
assaults on one's own self conception and on one's imaginary domain: The Imaginary Do-
main: Abortion, Pornography and Sexual Harassment (New York: Routledge, 1995), pp. 147—
158.
49. Cass R. Sunstein, "Social Norms and Social Roles," Columbia Law Review 96
(1996): 911.
50. Ibid.
51. Ibid.
52. Sunstein, "Preferences and Politics," p. 11.
53. Owen Fiss, "Why the State?" Harvard Law Review 100 (1987): 781-794; Frank
Michelman, "Conceptions of Democracy in American Constitutional Argument: The Case
of Pornography Regulation," Tennessee Law Review 56 (1989): 291-319; Cass R. Sunstein,
Democracy and the Problem of Free Speech (New York: Free Press, 1993).
54. Virginia Held, "Access, Enablement, and the First Amendment," Philosophical Di-
mensions of the Constitution, ed. Diana T. Meyers and Kenneth Kipnis (Boulder, Colo.: West-
view, 1988); Judith Lichtenberg, "Foundations and Limits of Freedom of the Press," Philos-
ophy and Public Affairs 16 (1987): 329-355.
55. For discussions of the potentially autonomy-undermining effects of too much
choice resulting from too much information, see Gerald Dworkin, "Is More Choice Better
than Less?" in Theory and Practice of Autonomy, pp. 62-81, and Catriona Mackenzie, "Bod-
ily Autonomy," unpublished manuscript.
56. Judith Butler, Excitable Speech: A Politics of the Performative (New York: Routledge,
1997).
57. Catherine Williams and thousands of others, including the membership of the
NAACP (National Association for the Advancement of Colored People), protested Miriam
Webster's dictionary definition of "nigger" as "a black person—usually offensive." They lob-
bied to substitute a definition that stressed the derogatory and dehumanizing aspects of the
term. National Public Radio, "Morning Edition," 16 March 1998.
298 aujnty and the soi
58. i would apply the following discussion of the autonomy-undermining effects of
group libel to group slander as well. However, it is group libel that has been explicitly con-
sidered, by the Court, to have the autonomy-undermining aspects I discuss below.
'"59. Prosscr, Resla.tem.ent »f Torts, 2nd ed. ( I 977), §559.
60. (bid., p. 156.
61. Ibid., §564/\, pp. 167-168.
62. Delgado, "Words That Wound: A Tort Action for Racial Insults, Epithets and
Name Calling," in Words That Wound, p. 1 10.
63. Bcauhamais.
64. Ibid.
65 578 F.2d i 197 (7th Cir. 1978), ceuiorari denied, 439 U.S. 915 (1978),
66. 376 U.S. 254 (1964).
67. Kcnurh l.asson, "in Defense of Group-Libel Laws or Why the First Amendment
Should Not Protect Nazis," Human Rights Annual, vol. 2 ( 1 985), p. 2.98. Lasson notes that
in four of these states (Connecticut, Massachusetts, Montana, and Nevada), "the gravamen
of the offense is holding up to ridicule, hatred, or contempt of any group or class of people
because of their race, color, or religion. J'hc Illinois statute, changed from that which was
upheld m Beauharnais, specifically requires that: the offensive speech be provocative of a
breach of the peace" (p. 2.98).
68. Berlin, " Jwo Concepts of Liberty," p. 1 55. (1 rccogni/e the irony of citing Berlin
in this context.)
69. ''Freedom of Expression."
70. Joel Fciuberg, ''Limns to the Freedom of Expression," in Philosophy of Law, 5th cd.,
ed. Joel Feinbcrg and Flyman Gross (Belmont, CaL: Wadswotth, 1995).
71. 1'einberg argues, however, that if we allow the category of group defamation as un-
protected speech, we have to ban any vaguely insulting speech about any group. But Fein-
berg, along with other liberals, would presumably not be prepared to run a similar argument
against the justness of affirmative action programs, in that case, they are prepared to ac-
knowledge both differences among different discriminated against minority groups and d i f -
ferences in the degree' of discrimination suffered by them. (One must acknowledge these im-
portant differences to counter the objection that affirmative action programs are unfair to
those minorities, say, Irish males, not favored by them.) But if one has this way of distin-
guishing among different sorts of discrimination, one surely has ways of distinguishing
among different kinds of damaging falsehoodstf«<afdifferent degrees of damage. Such an en-
terprise will take into account the present vulnerability of the defamed group, as well as their
ability or inability to remedy the damage with more speech.
72. It is important to note the history of stereotyping this claim reinforces. Such sen-
safiouaiistic stereotyping of, for example, black males as rapists, may be seen to be an instance
of irrational belief formation based on what has been called the "availability error." A. Tver-
sky and D. Kahneman, "Availability: A Heuristic for Judging Frequency and Probability,"
Cognitive Psychology 2 (1973): 207-232. See also Antonio R. Damasio's discussion of the
availability error in Descartes' Error: Emotion, Reason, and the Human Brain (New York: Avon,
1994): 191 192. Damasio considers such failures of rationality to be "not just due to a pri-
mary calculation weakness, bur also due to the influence of biological drives such as obedi-
ence, conformity, the desire to preserve self-esteem, which arc often manifest as emotions and
feelings" (p. 191).
73. Heauhtirndts,
74. Feinbetg, The Moral Limits of the Criminal Law. Vol. 2: Offense to Others (New
York: Oxford University Press, 1985), p. 158.
 Relational Autonomy and Freedom of Expression 299

75. Ibid.
76. "Democracy and Defamation: Control of Group Libel," Columbia Law Review 42
(May 1942): 727-780, 772.
77. Ann Garry, "Pornography and Respect for Women," Social Theory and Practice 4
(1978): 395-421; see also the articles in Price We Pay.
78. John Locke, An Essay Concerning Human Understanding, ed. A. D. Woozley (New
York: New American Library, 1974), pp. 259-267, 368-369; George Berkeley, The Princi-
ples of Human Knowledge (London: Collins, 1962), pp. 50-59.
79. Such a picture would more easily be seen by whites in a white-dominant culture as
a picture of women in general if it appeared to be a picture of a white woman. The depicted
race of the woman would typically not be viewed as salient and, thus, would not be a dis-
traction.
80. Joseph Raz, "Multiculturalism: A Liberal Perspective," Dissent, Winter 1994, pp.
67-79.
81. For a discussion of the ways in which expressions of racism can issue from the un-
conscious and not be the result of conscious, autonomous choice, see Charles R. Lawrence
III, "The Id, the Ego, and Equal Protection: Reckoning with Unconscious Racism," Stanford
Law Review W (January 1987): 317-388.
82. Mari J. Matsuda, "Public Response to Racist Speech: Considering the Victim's
Story," in Matsuda et al., Words That Wound, p. 26.
83. Conversation with Kendall Thomas, 16 March 1998, Berkeley, California. Thomas
does not endorse the view that hate speech regulation is categorically forbidden under the
U.S. Constitution. He cautions, however, that although regulating hate speech may be con-
stitutionally permissible, it may not be politically desirable. I agree with him that pragmatic
considerations may lead us to refrain from regulating even harmful hate speech. For exam-
ple, in recent years, some violators of campus hate speech codes have been elevated to the sta-
tus of First Amendment heroes, with the effect that the harms of their speech, as well as those
of the environment that fostered it, have not been adequately addressed.
This page intentionally left blank
 INDEX

Abbott, Ann, 279n.51 postmodern conceptions of, 28n.28

Abortion, 96, 101, 233n.l9, 256n.35. rational, 97, 122
£«• i«Zso Contraception; Pregnancy; social/relational dimensions of, 4, 6, 7,
Motherhood; Reproduction 10, 17,22-26,58-68,75,79,
Abrams, Fredrick, 206n.37 80-82, 85-89, 95-107, 113-14,
Addelson, Kathryn, 256n.29 118-22,139-44, 151-62, 252
Advocacy, 7, 25, 181, 182, 184-85, 187, Agent-regret, 140, 149n.36
188-98, 199,201,291 Alcoff, Linda, 177n.27, 178n.38, 201,
Agency 205n.34
atomistic/individualistic, 3, 6, 181-83, Alexander, M. Jacqui, 203n.6
196-97, 204n.lO Alienation, 35, 76, 90n.l8, 91n.27, 142,
autonomous, 13, 21, 22, 54, 57, 62-68, I49n.38, 155, 163, 177n.33
97, 139-40, 184,259 Anderson, Elizabeth, 97-98, 100, 102,
foucauldian theories of, 10, 11 106-8, 110n.6
free, 15, 19, 20, 29n.36, 87, 90n.9, Anderson, Gwen, 256n.44
227n.20 Anderson, Joel, 50n.34
moral, 4, 5, 73-74, 79-83, 85-89, Andrews, Lori, 256n.43
93n.50, 112, 152, 158-62, 183, 184, Anscome, G.E.M., 205n.28
187 Antony, Louise, 8, 49n.21
political, 4, 5, 73, 152, 158-62, 168, Appiah, Anthony, 157, 177n.26, 179n.45
183 Arendt, Hannah, 119-121, 123n.20

301
302 index
Aristotle, 17 moral, 37, 5In.37, 63, 158-62, 282,
Asch, Adrienne, 254n.20 295n.21
Atherton, Margaret, 49n.21 normative competence, 19-2,1, 89,
Autonomy 107-9, 11 I n . 51
atomistic, 6, 7, 3D, 40, 52, 183, 188, political, 158-62, 216
198,2.16 as procedural independence, 1 6, 60,
as authenticity, 17, 18, 54, 151 -63, 94
166-75 procedural theories of, 13-21, 22, 23,
as capability, 26, 280, 282 -86, 292-93, 2.4, 40, 41, 49n. 18, 5()n.31, 53, 56,
297n.47 59-61,64,94-107,215, 227,
competency, 13, 17-19, 22, 26, 39, 40, 232, 263. See also content-neutral
44, 53-58, 64, 12,4, 144, 145, 154, conceptions
166, 167, 172-75,226-32, programmatic, 18, 227, 255n.29
234n.4l, 240, 260-63, 267, relational, 4, 8, 10, 13, 17, 21-26,
2,83-84, 2,86 27n.l, 36, 40-47, 49n,23, 181, 189,
content- neutral conceptions of, 13, 190, 196, 203, 223 32, 233n.3,
18-19, 28n.33, 30n.57, 94, 95, 263, 237-53, 259-60, 268, 271-74,
284. Sec also procedural theories of 280, 282-86, 292-93, 297n.47
critical reflection in, 6, 13 21, 25, as self-sufficiency, 5, 6, 8, 9, 25, 35, 39,
29n.38, 37, 40 43,45,47,48, 52, 59,73,94,95, 152, 183, 185,
5 1 n.4-2, 53--56, 60-68, 75, 76, 190, 194-96,214
9In.27, 9:5, 98, 100-7, 124, and socialization, 4, 13, 15-24, 37, 39,
133-49, 147n.l9, 148n.27, 164-65, 40, 50, 54-58, 65, 95, 97, (06,
262, 284 I l()n.5, 124, 142, 144, lSOn.44, 157,
episodic, 18, 22,7, 255n.29 214, 219, 222, 225-26, 228-29,
episternic, 6, 25, 185-84, 1 9 1 - 9 2 , 198, 234n.41,284-85
20 5 n. 24 and social/personal relationships, 6-9,
feminist critiques of, 3--12, 2 1 , 35-36, 17, 21-23, 25, 35 -48, 50, 55, 57,
52-53,57, 152, 196-97,249 60-61,67-68,95, 124, 139,
group vs. individual, 5 1 n.50 14.3-44, I47n.l9, I49n.38, 164,
hierarchical theories of, 14—16, 19, 218, 222, 239-53, 256n.29, 260,
29n.38, I47nn.19,2.1, 169-72 283-85
hyperbolic, 6, 8, 25, 152, 181-5, substantive conceptions of, 19-21, 22,
194-8, 200, 202 24, 40, 95, 97, 98, 107-9, 110n.7,
as independence, 6, 8, 9, 39, 52, 14-2, I 1 1 n.5 1, 263, 264, 284, 285. See also
152, 195, 22,2. See also substantive normative competence
independence as substantive independence, 6, 9, 40,
individualistic conceptions of, 3, 8, 2,5, 59-60, 94
38, 50, 52, 59, 60, 181, 183-85,
189-90, 196-98, 202, 216, 237, Babbitt, Susan, I77n.27, 275, 276nn.5,10
238,240-42,246,252 Bacon, Francis, 185, 205n.23
as informed consent, 5, 25, 213 19, Badhwar, Neera Kapur, 5 In.49
224-26, 231-32, 237-39, 267 Baier, Annette, 7, 27n.9, 50n.28, 58,
integration accounts of, 12, 15, 17,2.1, 70n.32, I48n.34, 149n.37, 265, 283
24, 25, 29nn.36,38, 30n.46, 55, Baker, C. Edwin, 284, 293n.2, 296n.24
103-6, 125, 133-35, 139, 140-42, Baldwin, Nicola, 253n.2
I46n.16, 147n.l9, I48nn.26,27, Ball, David, 255n.25
30-31, I49nn.35,38, 150n.44, 154, Bar On, Bat-Ami, 207nn.51,70
168-75, ISOim.54,56, 227 Barclay, Linda, 23
legal, 216 Barker, Pat, 115-17, 146n.l6, I49n.38
 Index 303

Barrels, Diane, 258n.55 Brownmiller, Susan, 281

Bartky, Sandra Lee, 89n.6, 224, 277n.l7 Brungcr, Fern, 257n.53, 258n.55
Bartletc, Katharine, I49n.42, 176n.8 Bulkin, Elly, 49n.l4
Bassin, Donna, 179n.45 Burman, Erica, 204n.6
Battaglia, Debbora, 204n.6 Burt, Sandra, 206n.43
Bauer, Joanne, 296 Bushnell, Dana, 27n.6, 50n.30, 89n.6
Bayatpour, M., 278n.31 Buss, Sarah, 110n.7, 296n.32
Beauchamp, Thomas, 215-16, 237-41, Butler, Judith, 28nn.25,26,28, 176nn.l4,20,
276n.2 179n.45, 287, 288
Beauvoir, Simone de, 164
Beneficence, principle of, 215-16, 237 Cadava, Eduardo, 122n.5
Benhabib, Seyla, 27n.2, 28n.28, 57, 67, Calhoun, Cheshire, 89n.6, 104, 105,
69nn.l9,20, 113, 123n.20, 175n.4 180n.54
Benjamin, Jessica, 28n.23, 50n.27 Campbell, Sue, 273
Bell, Daniel, 296 Candib, Lucy, 200, 254n.lO
Bell, Kim, 49n. 16 Caplan, Arthur, 258n.55
Benn, Stanley, 29n.35, 55 Card, Claudia, 89n.6, 177n.27, 184, 194,
Benson, Paul, 15-16, 19, 20-21, 23, 202, 203n.4
28nn.33,34, 31n.59, 101-2, 107-9, Care See Ethics of care
109n.3, 110n.7, llln.51, 124, Cartesianism, 11, 185
I49n.39, 150n.44, 263, 266, Castoriadis, Cornelius, 202, 208n.88,
276n.lO, 277n.20 209n.89
Berkeley, George, 291 Certeau, Michel de, 200
Berlin, Isaiah, 289, 295n.20 Chang, Jung, 150n.46
Bernal, Ellen M., 206n.38 Childress, James, 215-16, 237-41, 276n.2
Bernauer, James, 123n.l7 Chodorow, Nancy, 9, 10, 50n.27, 176n.l9,
Biehl, Janet, 208n.79 183
Bioethics, 4, 5, 25, 213-32, 233n.l9, Choice
237-40, 252, 260 autonomous, 22, 24, 26, 37, 49n.l8,
feminist, 213-14, 218-23, 232 53-56, 60-61, 158, 217-23, 231,
feminist critiques of, 25, 214, 218, 223, 234n.38, 237, 251, 267, 271,
238 283-84
principlism in, 215-16, 221, 237-39, communitarian vs. liberal, 62 — 68
246, 254n.8 general conceptions of, 5, 213-16,
See also Health care, ethics 218-19,223-26
Black, Max, 145 rational, 5, 142, 214-15, 220-23,
Blackball, Leslie, 254n. 13 226-28, 231, 234n.38, 251
Blakeslee, Jill A., 234n.51 responsibility for, 75-76, 80, 82, 185
Bobrow, M.,257n.51 social influences on, 95-109, 164-65,
Bok, Sisela, 279n.48 219-21, 223-26, 261-62, 269,
Bordo, Susan, 49n.20 272, 286-87, 292
Borg, Susan, 178n.34 voluntary, 5, 142, 217, 225, 226, 238,
Bowden, Peta, 191 271
Bowen, Jan, 234n.37, 235n.54 See also Options; Preferences; Rational
Brison, Susan, 12, 27n.4, 37, 51n.4l, 82, Choice Theory
141, I46n.l6, I48nn.26,29,33, Christman, John, 16, 27n.6, 28n.33,
I49nn.35,38, 233n.2, 234n.52, 239, 29n.35, 48n.9, 50n.30, 90nn.9,17,
248, 294n.lO, 295n.l7 100, 102, 106, 108, 109nn.2,3,
Brothers, Doris, 272 110n.l2, 284, 293n.2, 296n.36
Brown, Justin Pugh, 178n.39 Cixous, Helene, 181, 198
304 hide1.
Clarke, Angus, 255n.25, 256n.32, 257n.51, Culler, Jonathon, 122n.5
258n.54 Cuomo, Chris, 208n.77
Class, 141, 183, 194, 195, 220, 223, 239, Curtis, David Ames, 208n.87
251,286
as determinant of identity, 4, 153-57, Damasio, Antonio, 298n.72
159-62, 164-65 Dan-Cohen, Meir, 297n.43
See also Identity, class Darwall, Stephen, 90n.ll
Clouscr, K. D., 254n.6 Davion, Victoria, 207n.51
Coady, C.A.J., 205n.25 Davis, Angela, 93n.54
Code, Lorraine, 6-8, 24, 25, 28n.2.6, De Laurctis, Teresa, 175n.3
49nn. 19,20, 50n. 26, 52, 59, 70n.28, Delgado, Richard, 288, 294nn.7,9
89n.5, 93n.43, 203n.3, 205n.28, Deliberation
206nn.34,35,43, 296n.33 epistemic, 193, 197
Coercion, 37, 40, 48n.9, 75, 76, 87, 194, moral, 191, 249, 252
217, 221, 224, 238, 262, 271, 274, practical, 4, 18, 24, 75, 124-26,
276n.3 133-40, 157, 174, 180n,56, 227,
Collins, Patricia Hill, 37 263, 268, 269
Communitarianisrn, 4, 5, 23, 38, 52, 55, De Man, Paul, 115-16
57, 69n.8, 255n.23 Dependency, relations of, 7-10, 25, 39,
Communitarian critiques of liberalism, 45,57-58,68,95, I48n.34,
27n.lO, 38, 53, 61-68, 183 191, 195,220,223,225,230,
Community 249, 283. See also Connection;
of choice, 23, 68 Interdependence
communitarian conceptions of, 60 — 68 Devall, Bill, 208n.80
of origin, 38, 64, 159 Derrick Jacques, 115-16, 118, 119,
moral, 75, 78, 81, 83, 84, 108, 109 121
relationships and traditions, 6-7, 41-4-7, Determinism
53, 57, 67, 104, 112, 119-20, 139, metaphysical, 4, 13, 15, 29n.38, 53,
141, I48n.34, 155, 163, 165, 194, 90n.l2, 169
238,251, 283,286, 288 social, 53-58, 61-62, 66-67
Compatibilism, 13, 16, 90n.l2 Dcwey, John, 77
Conditioning. See Autonomy, and Diamond, Irene, 208n.79
socialization Dillon, Robin, 20, 90n.ll, 263, 266,
Connection, relations of, 8-10, 35, 52, 276n.lO, 277nn. 17,20
95, 112-14, I47n.l9, I49n.38, Dirks, Nicholas, 178n.38
221-22, 239, 240-42, 248. See also Disability, 191, 227, 229, 244, 245, 249,
Dependency; Interdependence 251,269,270
Conley, Verena, 197, 199, 200 Disavowal, l47-48n.26, 170
Consciousness raising, 73, 89n.2 Dispositions, 127, 129-36, 141, 143,
Contraception, 96-99, 102-5, 108. I46n.l5,228,230
See also Abortion; Motherhood; Dissociation, 134-35, l47-48n.26, 169,
Pregnancy; Reproduction 177n.28
Contract, relations of, 59, 112, 113, 120, Dodds, Susan, 25, 26, 27n.3, 30n.49,
213, 219, 238, 254 234n.51
Cosmetic surgery, 30n.49, 142, 223, 269, Domination, 10, 11, 83, 144, 153,
270, 277-78n.26 155-57, 159-62, 168, 196, 198,
Cornell, Drucilla, 28n.25, 67, 69n.l9, 202, 260
297n.48 Dominelli, Lena, 278n.35
Covinton, S. S., 268n.29 Donchin, Anne, 25, 26, I48n.28, 233n.3
Crenshaw, Kimberle, 11-12, I49n.42, Downs, W. R., 278n.28
154-55, 160, 163 Drake, H., 257n.51
 Index 305

Dworkin, Gerald, 15, 16, 29n.35, 50n.31, Ethnicity, 4, 43, 153-57, 159-62,
51nn.37,42,63, 69n.lO, 109n.2, 227, 164-65, 223, 238, 281, 286. See also
284, 297n.55 Identity, ethnic
Dworkin, Ronald, 27n.3, 134, l47n.24, Experience, epistemic significance of, 185,
293n.2, 295n.20 187-89, 191-94, 196
Dwyer, James, 254n.22, 255n.23 Exploitation, 46, 60, 67, 73, 220-22

Ecology, 197-203 Faden, Ruth, 277n.25

Deep, 201, 208n.80 Faludi, Susan, 48n.5
Eistenstein, Hester, 27n.l7, 28n.23 Farrell, John, 203n.5
Eley, Geoff, 178n.38 Federman, Daniel, 277n,25
Ellison, Ralph, 72 Feinberg, Joel, 27n.3, 29n.35, 50n.33, 101,
Elshtain, Jean Bethke, 121 119, 290-91, 298n.71
Elster, Jon, 285 Femininity, ideal/norms of, 8— 10, 16, 20,
Embodiment, 4, 21, 113, 185, 218-20, 36-39, 46-47, 59, 74-76, 78, 80,
256n.35 95-109, 145, 152, 164-66, 217, 223
Emotions Feminism
as associated with femininity, 73, 49n.21, and communitarianism, 23, 65 — 68,
106, 224 69n.8
in autonomy, 13, 17, 18, 22, 37, 48n.8, cultural, 25,218,221-22
93n.46, 173-74, 234n.4l eco, 200
as characteristic of agents, 21, 37 liberal, 218-20, 222
in deliberation, 124, 126, 128-38, radical, 218, 220-22, 224
140-44, I47n.20, 167 See also Autonomy, feminist critiques
and reason, 38, 49n.21, 84, 106, 124, of; Bioethics, feminist; Bioethics,
224, 298 feminist critiques of; Epistemology,
Epistemology, 6, 113, 184-97, 210 feminist; Ethics, feminist; Individual-
feminist, 113, 162, 187, 189 ism, feminist critiques of
epistemic authority, 186, 190-94 Feminist philosophy, 35-36, 38, 40, 94,
epistemic communities, 184, 193 112-13, 125, I45n.5
epistemic credibility, 185, 187, 193, Feminist theory, 3-12, 18, 28n.28, 52, 59,
265 72
epistemic experts, 194, 196, 197 Ferguson, Ann, 207n.51
epistemic practices, 25, 184, 190-94, Ferguson, Kathy, 207nn.63,70
199 Fine, B. A., 257n.48
epistemic responsibility, 186, 189, 193, Firestone, Shulamith, 254n.21
194, 195, 197, 199, 201, 202, 203 Fischer, John Martin, 76-78, 83-85,
Epstein, Richard A., 293n.2 90nn.20,21,93n.45
Erlich, Howard J.,294n.7 Fischer, Karla, 5In.39
Equality Fiss, Owen, 287
gender, 88 Fivush, Robyn, 49n.22
moral, 80, 81, 188 Flanagan, Owen, 147n.l8, 177n.25
legal/political, 56, 125, 188, 194, 282, Flax, Jane, 28nn.25,26,28, 196
285, 288 Flinter, Francis, 255n.25
Ethics, 52, 201,215,221 Foucault, Michel, 10, 11, 183, 204n.l3,
feminist, 73-74, 81-86, 89n.4, 90n.l2, 205n.32, 206n.47
112 Foulks, James, 278n.27
of care, 8-10, 73, 89n.4, 94, 182, 218, Frankfurt, Harry, 14, 24, 29nn.35,36,38,
221-22, 255n.23 48n.8, 50n.28, 54, 105, llln.39,
See also Bioethics; Health care ethics; 134, 147n.22, 154, 169-72,
Moral theory 179n.50, 233n.4, 284, 295n.23
306 inde
Frascr, Nancy, 178n.38 Gill, Gillian C., I45n.5
i;razcr, Elizabeth, 69nn.8,19 Gilmati, Charlotte Perkins, 89n.9,
Freewill, 4, 10, 14, 'A, 55, 169 9!)nn.9,10
Freedom, 22, 56, 73, 92n.38, 106, 144, Glover, (onalhon, 30n.47, 5 In.47, 294n.6,
183, 186, 203n.5, 222, 247 296n.29
of choice, 218-20, 222 Golotnbok, Susan, 49n.22
of expression, 2,3, 26, 43, 280- 82, 286, Gondoli, I). M., 27811.28
292, 293n.2 Goodin, Robert, 5In.49
See also Speech, free Gordon, Colin, 145n.5, 20.5n.32
French, Peter, 93n.45, 1 10n.7 Gordon, Lewis, 204n.6
Freud, Sigmutid, 10, 1 1 , 14, l45ti.5, Cordon, Linda, 196, 2()7n.59, 279n.57
147n.2l, 156, 182 (iovier, "Irudy, 20, 1 75n.4, 265, 277n. 1 1..
Pricker, Elizabeth, 205n.25 278n.35
Fried, Charles, 293n.2 Grant, Judith, 279n.58
Friedman, Marilyn, 14, 15, 17,23, Graylin, Willard, 184, 194
27nn.l,6,12,16, 51nn.43,48, 55, 59, Green, Josephine, 257n.52
60, 64, 68, 69n.7, 70nn.2! ,24,42, Green wait, Kent, 293n.2
71n.4-8, 95, 104, I47n.19, lSOn.44, Griffin, Susan, 208n.79
17511.2, 246, 247, 256n.35, 296n.25 {Griffiths, Motwenna, 28n.24, 48n.! 1,
Friendship, 6, 9, 24, 41, 68, 112 18 I48n.26, 205n,28
Frye, Marilyn, 276n.7 Grimshaw, Jean, 2,8n,2/l, I48n.26, 203n.3
Fulda, Flans Friedrich, 48n,8 Guilt, 1 19, 120
Fullarton, Jane, 256n.43 Gunman, Amy, 176n,18, 177n.2.6, 179n,45

Gardiner, j, 1,., 28n.2.5 Habermas, Jurgen, S()n.34

Garry, Ann, 29911.77 Hanen, Marsha, 254n. 10
Gaugin, Paul, 35, 4} Haraway, Donna, 190, 199, 201, 205n.2.4,
Gauthicr, David, 70n.27 20711.52,
Geller, Gail, 254n,20 Harding, Sandra, 190, 201
Gender, 43, 239, 240, 241, 286 Flardwig, John, 248-49
as determinant of identity, 4, 153-57, Harm, 248, 272, 287-92, 2 9 5 n . l 6
159-62, 164- -65, 175n.6 as constraint on autonomy, 218-19, 2,21
inequality, 9, 45-46, 56, 225, 2.49 52 group, 154-55, 160, 163, 170,281-82,
norms/stereotypes, 36-39, 49n.21, 74, 287-92
7 5 , 8 1 , 9 8 - 1 0 9 , 125, 1 4 2 - 4 4 , to others, 2.48, 272
164 - 6 6 , 1 9 1 , 2 0 1 , 2 1 9 , 2 6 1 , 264, Harper, Peter, 255n.25, 256n.4l, 257n.47
269, 270, 272, 2.86 Flartsock, Nancy, 205n.34
oppression, 3, 72-73, 99-100 Hastrup, Kirsten, 204n.6
roles, 65, 1 9 6 , 2 5 0 - 5 1 Fiealth care, 25, 213-25, 229-30, 237,
socialization, 16, 18, 19, 37- 39, 56, 65, 2.59, 61-63, 266-71, 274, 276
72, 98, 107, 220 decision-making, 23, 2.5, 2,17-32,
See also Identity, gender; Internalization; 236-253, 240, 244, 248, 250-53,
Oppression; Stereotypes 267
Genetics, 25, 2.36-38, 240, 241-53 ethics, 218, 220, 223, 228, 232, 259,
genetic disorders, 236, 240, 243, 24-4, 276n.2
245, 248-5!, 256u.43, 257n.46 policy, 219, 230, 232, 237, 251
Cert, B., 254n.6 Flekman, Susan, 28nn.25--28
Gibbard, Allan, 49n.20 Heller, Thomas, 50n.3.3, 204n.9
Gilligan, Carol, 9, 27n.l7, 48n.4, 70n.29, Hegel, G.W.F., 62, I45n.5
94,221 Held, Virginia, 9, 10, 70nn.28 29, 73,
 Index 307

112, 234n.25, 254n.lO, 255n.23, personal, 4, 82, 145, 153, 155, 156, 158,
287, 296n.33 162-63, 177n.32, 179n.42
Helm, Bennett, 48n.8 racial, 141, 143, 153-57, 159-62,
Henig, Robin Marantz, 192, 206n.44 164-65, 175n.6, 179n.41, 241
Herb, Alice, 206n.37 sexual, 104, 153-57, 159-62, 164-65,
Hervik, Peter, 204n.6 175n.6, 220
Hester, R. K., 278n.27 social/cultural dimensions of, 4, 8, 17,
Heteronomy, 41, 43, 151, 161, 180n.51, 41,52-3,55,61-67,82, 117-18,
183, 186, 188, 195,215 139-44, 151-68, 177n.26
Heterosexuality, 36, 65, 157, 161, 164, 220 social group, 11,12, 24, 42, 119-20,
Hiley, David, 203n.5 141, 143, 153-70, 174, 176n.7,
Hill, Thomas E., Jr., 30n.53, 50n.33 177nn.26,27,33, 178n.38,
Hippocratic oath, 215 179-80n.51,238
Hoagland, Sarah Lucia, 28n.l9, 52, 73, See also Class; Gender; Ethnicity; Race;
89n.6, 203n.4 Sex
Hobbes, Thomas, 57, 70n.27 Identification, 14-16, 19, 21, 76, 105,
Holford, S., 278n.31 128, 131-34, 140, 155, 169-70, 224
Holmes-Seidle, M., 257n.51 sympathetic, 24, 114-15, 117-20
Holton, Richard, 265-66 Imagery
Holtzman, Neil, 256n.43 mental, 125-32, 136-39, 143-44
Honey, Margaret, 179n.45 cultural/metaphors, 38, 125, 137, 139,
Honneth, Axel, 203n.5, 204nn.7,8 143-44, I45n.5, 153-56
Hood, Colleen, 278n.43 Imagination, 20, 21, 24, 114, 117, 118,
Hood, Leroy, 255n.27 121, 122, 124-28, 132-33,
Horstmann, Rolf-Peter, 48n.8 136-40, 143-44, 166, 168
Houston, Barbara, 89n.6, 93n.4 philosophical, 112-14, 121, 151-52
Howard, M.O.,279n.51 cultural/social imaginary, 24, 25, 125,
Humanism, 182, 198 126, 139, 143-44, 182-84, 186,
Hume, David, I49n.37 187, 192-94, 198, 199, 202, 203
Hunter, Lynette, 188 Impartiality, 52, 161, 177n.31, 188, 189,
198
Identity Injustice, 41, 120, 156, 159-61, 170, 184,
class, 141, 153-62, 164-65, 175n.6, 241, 248, 253. See also Justice
183 Individualism
collective, 114 abstract/atomistic, 7, 52, 89, 185
essential properties and, 7, 8, 27n.l4 epistemic, 189-90, 197,202
ethnic, 42, 141, 143, 153-57, 159-62, feminist critiques of, 6-8, 56-60, 120,
164-65, 175n.6, 241 181-83, 189-90, 197-98,216
gender, 4, 39, 125, 143-44, 153-57, metaphysical, 7-8, 27n.lO
159-62, 164-65, 175n.6, 176n.l9, normative, 6, 25, 59-60, 196, 202,
179n.4l,220, 224 204n.lO, 289
individual, 8, 11, 12, 17, 24, 37, 55, Individuality, conceptions of, 5, 24, 112-14,
63-68, 114, 117-18, 120, 133-35, 117, 119, 120, 204n.lO, 241
140-44, I49n.35, 152-63, 169, Informed Consent, 5, 25, 213-19,
177n.26, 179n.44, 241 224-25,237-39,267
intersectional, 11, 12, 24, 30n.46, 141, Integrity
I49n.42, 151, 153-75, 176n.24, of relationships, 60, 61
177n.27, 179n.44 moral/personal, 73, 83, 89n.6, 105, 161,
metaphysical, 7, 8, 9, 14, 29 180n.54, 197,202
moral, 73 psychological, 248, 249, 288
308 indx
Interdependence, relations of, 6, 9, 25, 52, Kipnis, Kenneth, 297n,54
57,68, 112-13, 118,203,221-22, Kittay, Eva Feder, 70n.29, I45n.5
236 — 53. See also Connection; Knowledge, 185-98, 201-3, 263, 264,
Dependency 267-68
Interest groups, 193, 195 politics of, 185, 186, 192, 193, 196,
Interests, Patients' Best, 190-91, 213, 217, 206n.43, 208n.72
242-43,267 situated, 201, 205n.24
Internalization, 16, 19, 20, 74-76, 80, 87, subjugated, 189, 192, 202, 205n.32
98,99, 101, 106-9, 153, 158, 159, Kohen, J., 278n.29
163, 165, 170, 261, 275. See also Kornblith, Hilary, 206n.36
Gender; Oppression; Stereotypes Korsgaard, Christine, 92n.42, I46n.l0,
Irigaray, Luce, I45n.5 I48n.32
Itzen, Catherine, 294n.9 Koss, Mary, 207n.51
Kuhn, Thomas, 206n.46
Jaggar, Alison, 27n.l3, 49n.20, 89n.5 Kuczewski, Mark, 233n.l4
Jardine, Alice, 27n.l7, 28n.2.3 Kymlicka, Will, 27n.2, 55-56, 70n.43
Jecker, Nancy S., 206n.37
Jennings, Bruce, 184, 194 Lacan, Jacques, 153
Jenson, J. M., 279n.51 Lacey, Nicola, 69nn.8,19
Johnson, Barbara, 203n.2 Laidlaw, Foni Ann, 278n.35
Jones, Karen, 27n.3, 30n.49, 265 Lane, Ann, 90n.lO
Judgment Langer, Ellen, 178n.39
autonomous, 152 Larrabee, Mary Jeanne, 233n.23
moral, 76, 119, 160, 161, 162 Larson, Jeffrey, 178n.34
reflective/evaluative, 56, 81, 87, 128-9, Lasson, Kenneth, 289, 294n.l5
134-37, 140, 172, 262-64, 269, Lasker, Judith, 178n.34
272-73, 275, 277n.l2, 279n.57, Lawrence, Charles R., Ill, 294n.lO,
289 299n.81
Justice , 62, 183, 215, 216, 221, 237, 238, Lederer, Laura, 294nn.7,9
243 Legal Theory, 282, 289, 294n.l 5
social, 152, 156, 160,215 Le Doeuff, Michele, I45n.5
See also Injustice Lehrer, Keither, 277n.l2
Le Pichon, Yann, 48n.2
Kahneman, D., 298n.72 Lepine, Diane, 278n.35
Kamenka, Eugene, 29n.35 Le Roy, 258n,55
Kamuf, Peggy, 122n.5 Lesbianism, 38, 104, 157, 220, 264, 269,
Kant, 183, 184, 186, 198, 204n.!3 288
Kantianism, 5, 10, 1 1, 20, 5()n.33, 52, Liberalism, 4-5, 52, 55-56, 73, 156,
92n.42, 141, 182, 183, 184, 188, 177n.31, 182, 203, 216, 282, 286,
198,215,216,246 289, 298n.71. See also Communi-
Kaplin, Mcryle, 179n.45 tarian critiques of liberalism
Keil, A., 278n.28 Libertarianism, 5, 286, 289
Keller, Evelyn Fox, 9, 10, 28n.23, Liberty, 287, 289
50nn.27,33 negative, 5, 282, 295n.2()
Kennedy, Rosanne, I49n.42, 176n.8 personal, 216,219, 237
Kevles, Daniel, 255n.27 positive, 295n.20
Kim, Jaegwon, 206n.36 Lichtenberg, Judith, 287
King, Deborah, 175n.3, 176n.21 Lindlcy, Richard, 11 On. 13
King, Kathleen, 206n.44 Lindsay, Cecilc, 122n.5
 Index 309

Lippman, Abby, 257n.53, 258nn.54-55 69nn.10,11, 70n.29, 89n.4, 102, 107,

Lloyd, Genevieve, 24, 38, l45n.5, I46n.ll, 109n.5, llln.34, 124, 133,144,
I49n.34, 234n.31 145n.2, I46n.l6, I47n.l9, 149n.36,
Locke, John, 291 150n.45, 175n.24, 177n.29, 203n.6,
Lomasky, Loren, 38, 49n.l8 215, 226-32, 233n.4, 234nn.38,4l,
Longino, Helen, 48n.7, 206n.34 235n.52, 254nn.l6,18, 255n.29, 262,
Lovibond, Sabina, 184 284, 293n.2, 296n.25, 297nn.47,54
Lorde, Audre, 47 Michelman, Frank, 287
Lugones, Maria, 12, 89n.6, 155, 163, Michaels, Anne, I46n.l6
175n.3, 179n.45, 197 Mill, John Stuart, 185, 215, 216, 282, 289,
Luker, Kristin, 95-109 295n.22
Miller, B. A., 278n.28
Maclntyre, Alasdair, 49n,17, 53, 67 Miller, Brenda, 272
Mackenzie, Catriona, 24, 110n.7, 233n.l9, Miller, W. R., 278n.27
256n.35, 297n.55 Minow, Martha, 177n.30, 195, 196, 239
MacKinnon, Alison, 5In.44 Mohanty, Chandra Talpade, 203n.6,
MacKinnon, Catharine, 89n.2, 294n.9 208n.85
Mahowald, Mary, 256n.42 Moody-Adams, Michelc, 203n.4
Malmo, Cheryl, 278n.35 Moral competence, 23, 75-81, 83, 89.
Malterud, Kirsti, 206n.45 See also Autonomy, normative
Mangham, Colin, 278n.43 competence
Mann, Patricia, 176n.20 Moral luck, 35, 48n.3, l49n.36
Marshall, Leo, 205n.32 Moral theory, 4, 6, 52-53, 59-60, 73,
Marfatia, Lavanya, 257n.48 112-13, 162, 188,216
Marteau, T. M., 255n.25, 256n.30, Moral psychology, 4, 5, 12-22, 73, 162,
257nn.47,51 215
Masculinity, ideals of, 3, 8, 9, 35-39, Morgan, Kathryn, 30n.49,92n.43
49n.l6, 60, 94, 145,201,249 Morgen, Sandi, 204n. 11
Mastroianna, Anna, 277n.25 Motherhood/maternity
Matsuda, Mari, 292, 294n.lO norms of, 64-65, 99, 108-9, 142, 157,
May, William, 206n.37, 278n.35 163-65, 178n.37, 220-21,
McClure, Kirstie, 155, 163 261-62
McGuire, Don, 278n.43 surrogated, 142, 218, 219
McLeod, Carolyn, 25, 26, llln.44 See also Abortion; Contraception;
McMurrin, Sterling, 91n.29 Pregnancy; Reproduction
Medicine, 190-94, 219-25, 251-52, Motulsky, Arno, 256n.43
267, 269-70. See also Health care; Mouffe, Chantal, 155, 163, 176n.ll
Paternalism, Medical; Physician- Mullin.Any, I48n.30, 176n.7
Patient Relationship Multiculturalism, 197, 292
Memory, 21, 115-16, 118, 122, 143,
l48-9n.34 Naess.Arne, 208n.80
event/experiential, 126, 127, 130-31, Nagel, Thomas, 187, 293n.2
136-37, 146n.l5 Nedlesky, Jennifer, 9, 10, 26n.l, 50n.33,
traumatic, 146n,16, 148n.26 52, 70n.28, 93n.46, 95, 175n.4, 239,
Mepham, John, 205n.32 296n.25
Merchant, Carolyn, 208n.79 Nelson, Hilde Lindemann, 205n,30, 249
Meyers, Diana Tietjens, 12, 13, 17-19, 22, Nelson, James Lindemann, 249
24, 27n.l, 28n.25, 29n.38, 30n.46, Nidditch, P. H., I49n.37
39,48n.lO,50n.33, 54, 55, 56, Nielsen, Kai,254n. 10
310 indx
Nietzsche, Fricdricli, 11, 1 82 Pettit, Philip, 30n,53, 11 In.51
Noddmgs, Nel, 48n.4, 221, 234n,25 Physician-patient relationship, 4, 190-94,
Nonmaleficence, principle of, 21 5, 237 199, 213, 220, 224, 226, 252
No/ick, Robert, 57 contractual models of, 2.13, 238, 254
Nussbaum, Martha, 3()n.47, 5 1 n.4'7, Sec also Contract
18()n,56, 283, 294n.6, 296n.29 Piper, Adrian, 177n.2,5
Plant, Judith, 20811.79
Oakley, Ann, 258n,56 Plato, I45n.5
Object-relations theory, 183 Plumwood, Val, 197, 201, 2()7n.69
Objectivity, 188, 189, 193,201,224 Political theory, 4, 52-53, 59-60, 73, 1 12,
objectivism, 185, 187 113, 182, 188,216,282
Okin, Susan Mollcr, 5 In.45, 57 Point of view, 125, 127-36, 138-40, 14-3,
Oppression 145, 146n.i5, 147n.2.1, I48n.2.7
as characteristic of social structures/ Pornography, 4, 281-82, 290-91, 294n.9,
relationships, 12, 46, 5In.50, 60, 29711.48
67-68, 85, 88, 92n.33, 95, 106, 158, Porter, Catherine, 204n.l3
182 Postcolonialism, 181, 182, 198, 200
theories of, 4, 181, 276n,4, 277n. 17 Postmodernism, 28n.28, 158, 163
as undermining of autonomy, 10, 13, and critiques of autonomy, 1 0 — 1 2 ,
15-22,24-26,95-109, 124, 141, 25
143-44, 152, 162-63, 184,203, Porter, Elizabeth, 177ti.2,7, 20.5n.34
218-26, 238, 259-61, 263-76 Power
Sec also Gender; Internali/ation; Stereo- foucauldian theories of, 10, 11
types and knowledge, 192-94, 199-202
Oppressive socialization. See Oppression; political, 39
Autonomy, and socialization Power, relations of, 26, 39, 195, 216, 218,
Options, 61, 97, 2.23-24, 236, 260, 267 220, 240, 2,49, 256n.29, 264, 267
limited, 25, 224, 237, 239, 253, 261, within families, 2.37, 241, 247-48
287-88 Pratt, Minnie Bruce, 38
significant, 26, .152, 284-86, 292 Preferences, 5, 2,4, 26, 217, 228, 23 1, 247,
See also Choice; Preferences 297n.43
Orenstcin, Gloria Fcman, 208n.79 autonomous/authentic, 19, 20, 94, 95,
Ortner, Sherry, 178n.38 97, 100, 103,284,287
Oshana, Marina, 5()n.32 preference formation, 15-18, 2,2, 94, 96,
Overall, Christine, 233n.22 100 7, 144, 261, 284-87
See also Choice; Options; Rational
Parfit, Derek, l45n.l(), l46n.P5 Choice Theory
Paternalism, 194, 196, 2.19, 244, 251 Pregnancy, 96-103, 108, 109, 164,
medical, 194, 213 14, 223 24, 253, 218-19, 178n.37, 233n.l9, 241,
267, 268, 27711.24 243-44, 250-52, 256n.35, 261. See
Patriarchy, 37, 38, 46, 152, 1 86, 217, 2.20, also Abortion; Contraception; Mother-
221,223,241 hood; Reproduction
Paul, Denis, 205n.2.8 Privacy, right to, 237, 247, 289
Pearsall, Marilyn, 175n.3 Privilege, social, 200, 224, 252, 253, 260
Peelc, Stanton, 279n.45 and autonomy, 44, 152, 195
Personhood, 5, 6, 7, 58, 1 14, 129, 133, and identity, 153, 157, 160, 161-62,
172, 222, 269, 285 168, 179n.51
second persons, 6, 7, 58, I48n.34, 283, and knowledge, 192, 194
285, 287 Psychoanalysis, 9, 10, 125
See also Selfhood; Subjectivity; Respect Pnnalfs- Morejon, Diana, 257n.48
 index 311
311

Purdy, Laura, 27n.3, 30n.49, 218-19, Reproductive technology, 4, 30n.49,

254n.ll 218-21, 223, 257n.45, 261-62
Respect
Quine, W.V.O., 190 for autonomy, 216, 219-20, 223-24,
226-28, 231-33, 237-41, 243,
Race, 43, 183, 195, 220, 223, 239, 281, 246-48, 251-53, 262, 271, 287,
286,288-91,299n.79 292
as determinant of identity, 4, 153-57, for cultural traditions, 163, 292
159-62, 164-65 for the other, 116
theory, critical, 153, 176n.l6, 179n.4l, for persons, 90n.ll, 141, 295n.21
198 See also Self-respect
See also Identity, racial Responsibility
Racism, 11, 38, 67, 153, 154, 157, 196, collective, 113-14, 117-22, 195, 268
287, 290-92 epistemic, 186, 189, 193, 194, 195, 197,
and antiracism, 11, 157, 181, 182, 199, 199,201-3
275 family, 45-46, 57, 160, 171, 249-51
Radden, Jennifer, 204n.6 moral, 19,21-25,72-89,
Radin, Margaret, 30n.49 91nn.22,24,26,28-31,
Rapp, Rayna, 257n.48 92nn.35,37,42, 93n.45,49,50,
Rational Choice Theory, 5, 97, 226. See also llln.51, 113-22, 154, 159, 160,
Choice; Options; Preferences 240, 246, 248, 250, 252, 256n.35
Rationality, 38, 207n.69, 216, 234n.38 relational/social dimensions of, 74—76,
and autonomy, 96, 11 On. 13, 224 79-89, 92n.31, 93n.49, 113-22,
as characteristic of agents, 21, 37 I49n.34
in decision-making/deliberation, 96-97, social relationships and, 73, 79-80,
228, 286 83-88, 92n.42, 113-22, 249-52
deliberative, 63, 226, 234n.38 Rhetoric, 188, 193, 194, 195
minimal, 103 Richards, David A., 293n.2
See also Reason Richards, Martin, 255n.25, 256nn.30,32,
Ravizza, Mark, 76-78, 83-85, 257nn.47,49
90nn.20,21,93n.45 Ricoeur, Paul, I45n.5, I46nn.10,15
Rawls, John, 5, 18, 30n.47, 57, 59, 62-64, Riesman, David, 291
70nn.28,43, HOnn.13,16, 183, 184, Rights, 5, 6, 26, 59, 121, 156, 181, 198,
204n.l6, 215, 233n.4 200, 237, 248, 280, 295n.21
Raymond, Janice, 220 Robinson, A., 257n.51
Raz, Joseph, 31n.60, 50n.33, 285, 292, Robinson, Gillian, 209n.89
296n.38 Rogers, Lisa Kay, 178n.34
Reactive attitudes, 79, 82, 92n.31 Rooney, Phyllis, 145n.5
Reason Rorty, Amelie, l46-47n.!8, 177n.25
and autonomy, 19-20, 41, 97, 183, Rorty, Richard, 41
185-87 Rose, Mary, 50n.39
and emotions, 38, 49n.21, 106 Rosenberg, LeoT., 206n.37
Reasons-responsiveness, 74-77, 80, 83, 84, Ross, W. D.,215
87, 90n.20 Rossi, Alice, 205n,37
Redish, Martin, 284, 293n.2, 296n.24 Rowland, Robyn, 220-21
Reflective equilibrium, 104, 133, 134, 136, Ruddick, Sara, 38
140, I46n.l8, 149n.36 Rundell, John, 209n.89
Reproduction, 178n.37, 218-21, 249-51,
261 — 62. See also Abortion; Contra- Sandel, Michael, 27n.2, 49n.l7, 50n.35,
ception; Pregnancy; Motherhood 53, 62-68, 204n.8
312 indx

Sandilands, Catriona, 208n.84 Self-determination, 5, 13, 36, 55, 152, 182,

Sassoon, Siegfried, 1 17 183, 203n.5, 239, 240, 247, 249,
Saussure, Ferdinand, 182 252, 256n.35
Scanlori, Thomas, 9in.29, 289, 293n.2, Self-direction, 17-19, 102, 156-58, 165,
295-96n.2,3 175, 180n.56, 227-28, 231, 284
Scarry, Elaine, 191 Self-discovery, 17-19, 102, 159, 168,
Scheman, Naomi, 175n.7, 179n.51 172-74, 227, 263, 284
Schmitt, Richard, 204n.ll Self-doubt, 74-78, 87, 91n.27
Schneewind, J.B., 50n.33 Self-esteem, 8, 87, 107, 108, 141, 159, 273,
Schoeman, Ferdinand, 29nn.36,38, 9()n.l9, 279n,54, 286, 298
91 n.30, I47n.22 Self-interest, 6, 36, 59, 61, 195, 222
Schuckit, Mark, 278n.27 Self-knowledge, 24, 55, 63, 102-3, 129,
Schultz, Margaret, 30n.49 1.31, 138-40, 149nn.35,36, 156,
Scott, Joan, 17611.14 157-62, 167, 170, 172-75,264
Selby-Bigge, L. A,, 149n.37 Self-realization, 17-19, 21, 25, 30n.47, 35,
Selfhood 152,282
authentic, 17, 18, 24, 54, 102, 151-75 Self-respect, 4, 18, 21, 22, 75, 80, 90n.ll,
conceptions of, 3, 5, 7, 9, 1 1 3 — 1 4 , 93n.52, 263, 266, 269, 277n.21, 283
116-22, 156, 167, 179n.44, 182-8.3 Self-trust, 8, 21, 22, 26, 88, 95, 97, 107,
hierarchical accounts of, .14 — 16, 259, 261-75, 277nn. 12,21, 279n.47
29n.38, I47n.l9, I48n.27 Self-worth, 16, 20, 21, 22, 23, 72, 75, 78,
narrative accounts of, 36-38, 82, 80-82, 84-88, 92n.33, 93n.52, 95,
l48-49n.34, 167 99, 103, 107-9, 140-44, I49n.39,
relational/social dimensions of, 4, 1 0, 21, 263, 266
22-24, 25, 40, 4-1, 52-68, 83, 85, Seller, Anne, 187, 1 89
95, 1 1 2 - 1 3 , 117-22, 139-44, Sen, Amartya, 26, 30n.47, 280, 283, 285,
I48n.34, 149n.38, 158-67, 197, 286
2.38, 239, 247, 248, 265 Sex, 43, 195, 281, 286. See also Identity,
self-other relation, 9-10, 21, 23, sexual
11.3 18, 120-22, 218, 237, 2,48 Sexism, 11, 67, 153, 154, 157, 196, 260,
temporal dimensions of, 24, I 14, 118, 270, 272
121-22, 129-32, 139, I 4 6 n . l l , Sexual abuse, 241, 272-76, 278n.42,
148-9ii.34 279n.57
See also Pcrsonhood; Subjectivity Sexual difference, 10, 125, 143, 145
Self-blame, !78n.39 Sexual orientation, 153-57, 159-62,
Self-conception, 16, 17,22, 107, 125, 126, 164-65, 175n.6, 220, 223, 281, 286
129, 132-40, 142-44, I47n.21, Sexuality, 73,98-109
I48nii.26,27, 153, 156, 159, 162, Shaiiley, Mary, 254n.l5
166, 168, 230, 238, 245, 248, Slier-win, Susan, 25, 26, 11 In.44, 214,
256n.2.9 222-26, 232, 233rm.3, 12,18,
Self-control, reflective, 74-77, 79-81, 83, 234nn.25,32, 240, 254n.8,
84, 86-88, 90n.2l 276nn.6,8, 277n.2.3
Self-deception, 102-3, 136, 161 Shiloh, Shoshana, 257n.50
Self-definition, 17-19,2.4,38, 102, 125, Smart-, J.J.C., I47n.23
127, 129-44, 156-59, 162 68, Smith, Barbara, 49n.l4
170, 172-75, 177.11.27, 178nn.38,40, Smith, Michael, 30n.5.3, 1 1 ln.51
180n.56, 227, 234n.41, 2,48, 284, Social recognition, 80, 88, 125, 139-44,
286 I49n.38
 Index 313

Social welfare, 194-97,228 Thalberg, Irving, 14, 16, I47n.21

Socrates, 120 Theis, E. Charlotte, 206n.38
Solidarity, 24, 60, 62, 113-20, I49n.38, Thomas, Kendall, 292
166, 169, 179n.50 Thomas, Laurence Mordekhai, 5In.50
Soper, Kate, 205n.32 Thompson, Emma, 127
Sommers, Christina Hoff, 193 Thomson, Susan C., 49n.l6
Sosna, Morton, 50n.33, 204n.9 Tiles, Jim, 205n.23
Speech, 280, 281, 286-93, 293n.3, Tiles, Mary, 205n.23
295n.21, 298n.71 Tong, Rosemary, 218, 222, 233n.20
free, 4, 26, 213, 214, 280-82, 287, 289, Tradition. See Community, relationships
290, 292, 293n.2, 297n.47 and traditions
hate, 4, 26, 213, 280-82, 286-92, Tronto, Joan, 51n.49, 203n.4, 222
294nn.9,10,15, 295n.l6, 297n.47, Trust, 6, 23, 97, 189, 192, 197, 248,
299n.83 262-65,268-69,271-74
See also Freedom, of expression Tuana, Nancy, 204n. 11
Spelman, Elizabeth, 89n.6, 175n.3 Tully, Janes, 1230.21
Stereotypes, 167, 278n.34 Tversky, A., 298n.72
cultural/group, 157, 159, 163, 177n.26, Tyler, A., 255n.25
201,238, 291, 292, 298n.72
See also Gender; Internalization; Uniacke, Suzanne, 30n.49
Oppression Utility, 97
Stoljar, Natalie, 21, 23, 27n.l4, 110n.7, Utilitarianism, 215, 216
I49n.43, 296n.31
Stone, Geoffrey, 293n.2 Veevers, J. E., 177n.34
Stratham, Helen, 257n.52 Vig, Elizabeth, 254n.22, 255n.23
Strauss, David, 293n.2
Strawson, Galen, 92n.38 Walker, Margaret Urban, 89n.4, 184,
Subjection, 3, 25, 181, 185, 186, 195, 197, 187, 190, 202, 204n.ll
200, 201, 203 Wallace, Jay, 75-78, 81, 83-84,
Subjectivity, 4, 118, 119, 152, 163, 91nn.27-30, 92nn.35,37
181-83, 196-9,203 Waring, Marilyn, 204n.6
atomistic, 6, 7, 182,221,222 Warren, Mary Anne, 30n.49
enlightenment conceptions of, 11, 73, Warson, Gary, 14, 29n,36, 77, 79,
113, 182-83, 198 91n.24, 92n.32, 93n.49, llln.39,
individualistic, 6, 7, 14, 182-84, 202, 233n.4
204n.lO Weakness of will, 99, 103, 108
relational, 6, 7, 183 Weir, Alison, 175n.4
postmodern theories of, 10 — 11, 182 Weiss, Penny, 51n.48, 67, 69n.8
Subordination, 11, 45, 56, 73, 83, 88, Wellberry, David, 50n.33, 204n.9
152-57, 159-62, 168, 170, 171, Wells, R. D., 278n.31
202, 217, 220 Wendell, Susan, 89n.6, 191, 195
Suleiman, Susan, 179n.45 Wertz, Dorothy, 255n.25
Sunstein, Cass, 286, 287, 293n.2, 296n, West, Cornel, 195
297n.38 Wexler, Nancy, 255n.27
Superson, Anita, 89n.6 Whitbeck, Caroline, 279n.49
Whitford, Margaret, 28n.24, I48n.26,
Taylor, Charles, 57, 296n.30 205n.28
Testimony, 7, 186-88, 191, 194, 273 Williams, Ann, 256n.40
314 indz

Williams, Bernard, 33, 48n,3, 134, wong ruijhid 146-47-18

I45nn,6,7, I47n.23, I49n.3f> wrifte g,h von 20528
Williams, Patricia, 194 wylir alir 208o74
Williams, Raymond, 207n,72
Wilt, Charlotte, 27n. 14, 49n.2 1 yeffe 579 51
Wittgenstein, Ludwig, 182, 187 yiojhfggnhgngjfmhjfhgkah1786
Wolf, Susan, 19- 21, 29n.38, 77, 85, 276n2783227951
9h-).22, 92r,.35, 97, 1 1 l n . 3 1 ,
233nn.l5,18, 239, 252, 234n.8 youbngrojftht28 5569 10
Wollheim, Richard, 125 33, 136,
145mn1014m15 xzhuhgnn17652