THE PHILOSOPHICAL FORUM

Volume XXXI, No. 2, Summer 2000

THE ANTI-SUBJECTIVE HYPOTHESIS:

MICHEL FOUCAULT AND THE DEATH OF THE SUBJECT

AMY ALLEN

The centerpiece of the first volume of Michel Foucault’s History of Sexuality

is the analysis of what Foucault terms the “repressive hypothesis,” the nearly
universal assumption on the part of twentieth-century Westerners that we are the
heirs to a Victorian legacy of sexual repression. The supreme irony of this be-
lief, according to Foucault, is that the whole time that we have been announcing
and denouncing our repressed, Victorian sexuality, discourses about sexuality
have actually proliferated. Paradoxically, as Victorian as we allegedly are, we
cannot stop talking about sex. Much of the analysis of the first volume of the
History of Sexuality consists in an unmasking and debunking of the repressive
hypothesis. This unmasking does not take the simple form of a counter-claim
that we are not, in fact, repressed; rather, Foucault contends that understanding
sexuality solely or even primarily in terms of repression is inaccurate and mis-
leading. As he said in an interview published in 1983, “it is not a question of de-
nying the existence of repression. It’s one of showing that repression is always a
part of a much more complex political strategy regarding sexuality. Things are
not merely repressed.”1 Foucault makes this extremely clear in the introduction
to the History of Sexuality, Volume 1, when he writes:

Let there be no misunderstanding: I do not claim that sex has not been prohibited or barred or
masked or misapprehended since the classical age. . . . I do not maintain that the prohibition of
sex is a ruse. . . . All these negative elements . . . which the repressive hypothesis groups together
in one great central mechanism destined to say no, are doubtless only component parts that have
a local and tactical role to play in a transformation into discourse, a technology of power, and a
will to knowledge that are far from being reducible to the former.2

In short, in Foucault’s view, there is much more to our alleged sexual repression
than initially meets the eye.

113
 AMY ALLEN

In what follows, I will not be concerned with the repressive hypothesis per
se. Instead, I introduce this aspect of Foucault’s work because, just as Foucault
unmasks the repressive hypothesis, I intend to unmask and debunk an assump-
tion that has become commonplace in the critical literature on Foucault, an
assumption that I will term “the anti-subjective hypothesis.” The anti-subjective
hypothesis consists in the belief that the point of Foucault’s archaeological
analyses of discourse and his genealogical analyses of power/knowledge is to
attack, undermine, and eventually eradicate the concept of the human subject.
In short, those who subscribe to the anti-subjective hypothesis believe that
Foucault participates in, even celebrates, the so-called “death of the subject.”
This interpretation of Foucault has had an enormous impact on the reception of
Foucault in the United States, particularly by social and political philosophers.
Insofar as the capacity for being a thinking subject capable of reflecting on and
deliberating about courses of action seems to be a precondition for moral or
political agency, Foucault’s alleged eradication of the subject seems also to
commit him to a denial of the possibility of moral or political agency. Further-
more, insofar as moral agency is generally thought to be a precondition for
moral responsibility, Foucault’s participation in the death of the subject seems
to imply that it is no longer possible to make any claims regarding who is
responsible for unjust social and political arrangements. Indeed, all claims that
a set of social and political arrangements is unjust seem to be out the window as
well, since it is not clear that it makes sense any longer to speak of injustice in a
deterministic world such as Foucault’s is taken to be. Thus, the claim that
Foucault eradicates the subject is often the first step in an argument that serves
to justify a marginal place for Foucault in contemporary social and political
philosophy. Clearly, then, the stakes involved in this reading of Foucault are
quite high.
Interestingly enough, the anti-subjective hypothesis has been accepted by a
wide spectrum of commentators on Foucault, ranging from some of his harshest
critics to some of his most ardent supporters. While his critics bemoan the death
of the subject, his supporters applaud it. Given the virtual consensus among
commentators, one might be tempted to leave well enough alone. However, I
shall argue that a close analysis of the anti-subjective hypothesis and of the ar-
gument advanced in favor of it reveals that there is much more to Foucault’s
alleged eradication of the subject than initially meets the eye. After laying out
the argument that is advanced by supporters of this hypothesis, I shall argue that
this way of reading Foucault should be rejected. I shall suggest in its stead an
alternate interpretation of Foucault’s archaeologies and genealogies that reveals
his work to be an elaboration of the historically, culturally, and socially specific
conditions of possibility for subjectivity, rather than an eradication of this

114
 THE ANTI-SUBJECTIVE HYPOTHESIS

concept. Finally, I shall conclude by sketching out the implications of my rejec-

tion of the anti-subjective hypothesis.

THE ARGUMENT FOR THE ANTI-SUBJECTIVE HYPOTHESIS

Let me begin, then, by presenting the argument put forward by critics of

Foucault in support of the anti-subjective hypothesis. Different versions of this
argument have been made by a wide range of commentators whose work comes
from a variety of philosophical perspectives, including Jürgen Habermas, Axel
Honneth, Thomas McCarthy, Michael Walzer, Charles Taylor, Nancy Hartsock,
and Linda Alcoff.3 For my purposes, the subtle nuances that differentiate these
versions of the critique of the anti-subjective hypothesis are less important than
the general concern that these critics share. Hence, I shall gloss over the particu-
larities of some of the individual arguments in order to reconstruct the general
philosophical worry that motivates these commentators.
Those who endorse the anti-subjective hypothesis argue that Foucault claims,
or that his archaeological and genealogical works imply, that the subject is
merely or nothing more than the effect of discourse and power. They maintain
that such an account of the subject amounts to nothing short of a wholesale repu-
diation of the concept of subjectivity. These critics presuppose or explicitly
endorse something like the following reading of Foucault’s works. In his archae-
ological works, Foucault offers a quasi-structuralist or functionalist analysis of
the rise in the modern era of the human sciences—medicine, psychiatry, sociol-
ogy, anthropology, and so on. The explicit aim of this analysis is to reveal the
theoretical object of the human sciences (man) as a social and cultural construc-
tion. For instance, in The Order of Things, Foucault chronicles the relatively re-
cent invention of man in the late eighteenth or early nineteenth century; this
invention was part and parcel of the emergence of a new kind of scientific and
social-scientific discourse that we have come to call humanism. Because man has
only come into being with the emergence of humanism as a world view, Foucault
claims that “man is an invention of recent date. And one perhaps nearing its
end.”4 The upshot of Foucault’s analysis of humanism, critics claim, is that the
subject itself is nothing more than the effect of the anonymous functioning of a
particular, historically and culturally specific, discourse. The lesson to be drawn
from Foucault’s early work is that, as Axel Honneth puts it, “the human is no
longer the experiential center of a course of action which he encounters and over-
sees, but the arbitrary effect of a network of events out of which he can no longer
make sense and which is produced by the rules of language.”5
In his genealogical works, Foucault’s theoretical focus shifts to include an
analysis of the historically specific non-discursive social practices that have de-
veloped alongside the rise of humanism as a discourse. In these accounts, the

115
 AMY ALLEN

argument for the anti-subjective hypothesis continues, Foucault drives home a

few more nails in the subject’s coffin by claiming that the subject is not only
produced by discursive formations but also generated by a network of anony-
mous power relations that pervade the social body. Foucault terms these inter-
related discursive and non-discursive formations “power/knowledge regimes.”
For instance, in Discipline and Punish, Foucault chronicles the emergence in
the modern era of a distinctive modality of power—disciplinary power—that
creates new, peculiarly modern, subjects. Disciplinary power emerges when
various disciplines (and the dual meaning of the term “discipline” is highly
significant for Foucault’s account) within the human sciences—psychiatry,
education, criminology, the social sciences, and the law—begin to generate
new ways of controlling the citizenry and of punishing those who get out of
line. Disciplinary power creates subjects who are thoroughly subjected—both
from without by the normalizing force of humanistic discourses and practices,
and from within by the self-disciplining impulses that humanism has taught
them to internalize. Examples of such thoroughly subjected subjects include
prisoners who surveil themselves because they know that they might at any mo-
ment be under surveillance, schoolchildren who move quietly from class to
class with the sound of a bell, and factory workers who mechanically perform
one specific, routinized function over and over again for an entire work day.
The upshot of this analysis, supporters of the anti-subjective hypothesis claim,
is that the subject is “nothing but the effect of power relations”;6 it is, as Linda
Alcoff puts it, “causally inefficacious, historically constructed, even a kind of
epiphenomenon of power/knowledge.”7 This reading of Foucault concludes
that, taken together, Foucault’s archaeological and genealogical investigations
imply a wholesale eradication of the concept of human subjectivity.
There are two related implications of this putative elimination of the subject.
First, if Foucault believes that the subject is merely an effect or epiphenomenon
of power/knowledge regimes, then he seems to be committed to an embrace of a
very bleak kind of cultural and social determinism. If there is nothing more to
the subject than the inscription of power/knowledge onto an individual’s body,
then there would seem to be no capacity in the subject for reflection upon or
resistance to forces operating outside of it; thus, the subject must be wholly de-
termined by such forces. The purported determinism of Foucault’s view is
thought to be problematic in at least two respects. First, it leads Foucault to ig-
nore entirely the kind of reflective self-understandings on which individuals rely
in making sense of their social world.8 That is to say, it completely renders neg-
ligible the first-person perspective of the individual agent as a crucial aspect of
analysis for social and political philosophy. As a result, an overly deterministic
view, such as Foucault’s is alleged to be, is simply methodologically inadequate.
Second, such a deterministic view makes it impossible to give any adequate

116
 THE ANTI-SUBJECTIVE HYPOTHESIS

account of resistance to social relations of domination and subordination. If indi-

viduals were wholly determined from without, then they would have no internal
resources on which to draw in order to combat social domination and normaliza-
tion. If this were true, then prisoners, schoolchildren, and factory workers would
be completely incapable of resisting the disciplinary power imposed upon them.
This kind of a view rightly strikes critics as overly pessimistic and empirically
false. We certainly seem to be able to pick out instances of resistance to domina-
tion, and Foucault himself was known to have encouraged such instances
through his practical, political engagement in the prison reform movement in
France. Moreover, and perhaps more important, if the impossibility of resistance
is indeed an implication of what Foucault says about subjectivity, then his own
philosophical view is in danger of being rendered internally inconsistent, since
he says many times that resistance is not only possible, it is actually coextensive
with the exercise of power.9
The second and related implication of the alleged elimination of the subject is
that the death of the subject implies the death of agency, which in turn makes it
impossible to assign responsibility to anyone for the unjust imbalances of power
that stubbornly persist in the kinds of societies that Foucault analyzed. As Linda
Alcoff puts it, “one important problem with [Foucault’s] view is that if we can-
not posit agency on the part of actors in the historical drama, then it is impossi-
ble to assign responsibility for the state of things to anybody. In other words,
there is no one to blame.”10 Critics of Foucault assume—rightly, I think—that
an effective social and political theory should enable us to escape the trap of
cultural determinism and to assign responsibility and blame for the differential
ability to exercise power and to access social goods that characterizes Western
societies.
Moreover, for at least one critic, Foucault’s eradication of the subject seems
suspiciously convenient, coming as it does at a point in history when many
members of oppressed groups are just beginning to demand to be treated as full
subjects with all of the rights and responsibilities that come along with such a
status. Nancy Hartsock asks, “why is it that just at the moment when so many of
us who have been silenced begin to demand the right to name ourselves, to act
as subjects rather than objects of history, that just then the concept of
subjecthood becomes problematic?”11 In Hartsock’s view, Foucault’s eradica-
tion of the subject not only makes it impossible to construct a viable social and
political theory, it also snatches the ideal of subjectivity out from under those
who have historically been denied its status and benefits and who are at long last
on the verge of attaining that ideal.12
In sum, those who accept the anti-subjective hypothesis contend that Foucault
undermines and even eradicates any notion of the subject because his investiga-
tions into discourse and power maintain that the subject is nothing more than the

117
 AMY ALLEN

effect of anonymously functioning power/knowledge regimes. Such an eradica-

tion of the subject, these critics argue, is methodologically unsound, empirically
false, and normatively suspect.
Foucault’s critics make quite a compelling case and raise problems that, if
they truly are implications of his analysis, may well be insurmountable. Re-
cently, in response to the growing number of social and political theorists who
have traced out the disturbing implications of this reading of Foucault, a number
of his supporters have countered this reading by claiming that there is a viable
account of subjectivity and agency to be found in Foucault’s work. This account,
supporters of Foucault claim, can be found in the works that were published
around the time of his death, which include volumes 2 and 3 of the History of
Sexuality, and assorted interviews and essays on the topic of what Foucault
called “practices of the self.” Those who accept this reading grant Foucault’s
critics that his archaeological and genealogical works lack a satisfactory account
of subjectivity and agency, but, they contend, this deficiency is overcome in his
ethical works, that is, in his late works on the practices of the self in ancient
Greek and Roman cultures. On this way of understanding his oeuvre, after the
publication in 1976 of volume 1 of the History of Sexuality, Foucault turned his
attention away from the way that subjects are produced via the anonymous func-
tioning of power/knowledge regimes and toward the ways in which individuals
actively and reflexively constitute themselves via a particular kind of ethical re-
lation to self (rapport a soi).13 Although Foucault prefers to speak in his late
work of a “self” rather than a “subject,” it seems clear that the notion of a
self-constituting self presupposes some conception of a thinking subject who is
capable of reflecting on what kind of self he or she wants to be and deliberating
about the best way to become that sort of self, and it also presupposes some
conception of an agent who is capable of acting in the world in such a way as to
become the sort of self that he or she wants to be. As Richard Bernstein puts it,

One might think . . . that Foucault is heralding the death of the subject, that he is claiming that
the subject itself is only the result of the effects of power/knowledge regimes, that he completely
undermines and ridcules any and all talk of human agency. There is plenty of textual evidence to
support such claims. But it is also clear, especially in his late writings when he deals with the
question of the self’s relation to itself and the possibility of ‘the man who tries to invent himself’,
that he is not abandoning the idea that ‘we constitute ourselves as subjects acting on others’.14

In other words, Bernstein appeals to the late work on practices of the self as evi-
dence for the fact that Foucault must not have participated in the death of the
subject in his earlier work.
Foucault’s defenders acknowledge that Foucault’s untimely death in 1984
meant that he was unable to complete his project on practices of the self, but,
they claim, the seeds for an adequate account of subjectivity are still to be found

118
 THE ANTI-SUBJECTIVE HYPOTHESIS

in his late works. This account of subjectivity, they maintain, can in turn provide
the necessary corrective for his earlier—admittedly overly deterministic—
accounts of the death of the subject.15 As Thomas Flynn puts it, the analysis in
the late Foucault “fills in a gap in ‘structuralist’ historiography, namely, the ab-
sence of the individual, responsible agent”;16 moreover, it does so without taking
on any of the baggage of the Cartesian, transcendental, and phenomenological
accounts of the subject that Foucault so vehemently criticized in his early work.
Unfortunately, this attempt to respond to the anti-subjective hypothesis does
not help matters much, since those who argue that Foucault’s archaeological and
genealogical works eradicate the subject tend to view his late turn to a study of
practices of the self as, at best, a capitulation to the arguments of his critics and,
at worst, an outright contradiction. Either way, many critics maintain, the turn to
an account of the subject in the late Foucault constitutes a radical and unrecon-
cilable break with his earlier projects.17 As such, it does not help to correct the
wholesale elimination of the subject that they diagnosed in his earlier work. Fur-
thermore, even if the late work could be reconciled with the early work, at least
one critic maintains that the result would not be an integrated and useful account
of the subject; on the contrary, Thomas McCarthy claims, we would merely be
left with an overly deterministic and holistic account cobbled together with an
overly voluntaristic and individualistic account. McCarthy writes:

[In Foucault’s early work,] everything was a function of context, of impersonal forces and fields,
from which there was no escape—the end of man. Now the focus is on ‘those intentional and
voluntary actions by which men not only set themselves rules of conduct but also seek to trans-
form themselves . . . and to make their life into an oeuvre’—with too little regard for social,
political, and economic context. Neither scheme provides an adequate framework for critical
social inquiry.18

Even taking the turn toward the self in the late Foucault into account, we are still
left without a coherent and viable account of subjectivity and, thus, without
workable conceptions of agency, responsibility, and resistance. If the reading I
have been discussing is correct, then the anti-subjective hypothesis still stands,
and Foucault’s view appears to remain in question.
In light of these compelling arguments in favor of the anti-subjective hypothe-
sis, and considering the unsavory implications of such an eradication of the sub-
ject, it appears that the most promising way to defend Foucault is to deny that he
was ever guilty in his archaeological and genealogical works of participating in
the death of the subject. I shall mount such a defense first by questioning one of
the key presuppositions of the argument in favor of the anti-subjective hypothe-
sis. I shall then proceed to offer an alternate reading of Foucault’s philosophical
project, one that takes seriously his own characterization of that project as “a
history of the different modes by which, in our culture, human beings are made

119
 AMY ALLEN

subjects. . . .”19 According to this alternate reading, Foucault’s archaeological

and genealogical works never constituted a full-scale repudiation of the concept
of subjectivity; on the contrary, it was precisely the subject—and not discourse
or power—that was all along the “general theme of [Foucault’s] research.”20
Before I present this alternate reading, let me call into question a crucial and
highly problematic assumption behind the reading of Foucault put forward
by those who accept the anti-subjective hypothesis. As I argued above, these
critics are particularly worried about what they take to be Foucault’s claim that
the subject is merely or nothing more than an effect of discourse and power.
If the subject is merely or nothing more than an effect of power/knowledge
regimes, then clearly individual agency, resistance to oppression, and moral re-
sponsibility for one’s actions are out the window. I maintain, however, that this
worry is based on an overreaction. In fact, Foucault is very careful never (at
least not as far as I am aware) to say that the subject is merely an effect of dis-
course and power. Instead, he says that it is an effect of discourse and power.
For instance, he writes, “the individual . . . is not the vis-à-vis of power; it is, I
believe, one of its prime effects. The individual is an effect of power, and at the
same time, or precisely to the extent of which it is that effect, it is the element of
its articulation.”21 This is a seemingly small but nonetheless extremely signifi-
cant difference. Whereas the former claim would clearly imply a subject wholly
determined by forces outside of him or her, the latter claim does not, I think,
have to have the same deterministic overtones. To suggest just one
counterexample, let me note that children are effects of their parents in this
sense. Children are shaped from birth by the practices, beliefs, and superior
power of their parents. However, no one who actually has children will think for
a second that this statement implies that children have no agency or will but are
instead completely determined and controlled by their parents. The example
may be mundane, but the point behind it is crucial. To say that subjects are the
products of forces that are largely out of their control, as Foucault does, is not to
say that they have no control over anything whatsoever.
But, critics will cry, Foucault spends the vast majority of his time and energy
in his archaeological and genealogical works talking about the outside forces
and precious little time or energy telling us anything at all about individual sub-
jectivity and agency. If he really believed all along that individuals were not de-
termined by power/knowledge but instead played an important role in their
self-constitution, then why did he wait until the early 1980s (all too conveniently
after many people had already criticized him on precisely this point) to say so?
My answer to this question brings me to my alternate reading of Foucault. In
a nutshell, this reading maintains that Foucault’s analyses of power/knowledge
regimes are devoid of references to the concept of subjectivity because they
have to be. And they have to be devoid of such references because precisely the

120
 THE ANTI-SUBJECTIVE HYPOTHESIS

point of these works is to shift subjectivity from the position of that which ex-
plains to the position of that which must be explained, from explanans to
explanandum.

THE SUBJECT AS EXPLANANDUM

If this is a plausible and defensible way of understanding what Foucault was

up to, then it makes perfect sense that he refused to thematize the subject explic-
itly in his early works, choosing instead to concentrate on a discussion of back-
ground discursive and non-discursive practices that make subjectivity possible.
On this reading of Foucault, his argument is that subjectivity should not be con-
ceived as the necessary, fixed point around which all other theoretical explana-
tions rotate, as it has been conceived in Western philosophy since Descartes.
Instead, Foucault maintains that the peculiarly modern, humanist conception of
subjectivity is contingent and that its emergence at this particular point in history
requires explanation. If Foucault is trying to shift the subject from its usual posi-
tion as the explanans in light of which everything else is deciphered to the posi-
tion of the explanandum which must itself be illuminated, then he would have to
begin by setting aside any preconceived notions of the subject. Far from elimi-
nating the concept of subjectivity, however, such a move merely indicates a shift
in the explanatory priority of that concept. Foucault’s aim is to offer an account
of how subjectivity is constituted; the best way to do this is to bracket (and I use
this term advisedly) any notion of the subject as constituent.
Foucault describes his work in precisely this way in an interview published in
1977. He says:

One has to dispense with the constituent subject, to get rid of the subject itself, that’s to say, to
arrive at an analysis which can account for the constitution of the subject within a historical
framework. And this is what I would call genealogy, that is, a form of history which can account
for the constitution of knowledges, discourses, domains of objects, etc., without having to make
reference to a subject which is either transcendental in relation to the field of events or runs in its
empty sameness throughout the course of history.22

Now, this passage has been cited as evidence for the anti-subjective hypothesis
because in it Foucault claims that we have to “get rid of the subject itself.”23
However, I think that this kind of interpretation of this remark is misleading for
two reasons. First, Foucault initially says that what he wants to get rid of is the
“constituent subject,” or what I have called the subject as explanans. As I have
indicated, this seems a reasonable first step in the process of shifting the explan-
atory priority of the concept of subjectivity. Second, although Foucault goes on
to say that we have to “get rid of the subject itself,” the explicitly stated purpose
of doing so is to “arrive at an analysis which can account for the constitution of

121
 AMY ALLEN

the subject within a historical framework.” In other words, Foucault’s aim is not
to get rid of the concept of subjectivity altogether; instead, he sets aside any con-
ception of the subject as constituent in order that he might better understand
how the subject is constituted in a particular way in this particular cultural and
historical milieu.
If we start with this kind of an understanding of Foucault’s philosophical proj-
ect, then we end up with a reading of his archaeological and genealogical works
that is very different from the one offered in support of the anti-subjective hy-
pothesis. Although a full account of all of the details of such a reading is beyond
the scope of this essay, I can offer the following sketch of it. Foucault’s archaeo-
logical works attempt to describe discourses without reference to foundational or
transcendental conceptions of the human subject. This does not mean that the
concept of subjectivity is irrelevant to this project; on the contrary, delineating
the ways in which historically specific discourses make possible particular
modes of subjectivity is precisely the point of that project. Thus, Foucault de-
scribes his project in The Archaeology of Knowledge as belonging “to that field
in which the questions of the human being, consciousness, origin, and the sub-
ject emerge, intersect, mingle, and separate off.”24 Similarly, in the preface to
the English edition of The Order of Things, Foucault emphasizes that although
his analysis starts at the level of discourses rather than at the level of individual
subjects or agents, this should not be taken as a denial of the possible efficacy of
any subject- or agent-centered account. On the contrary, he writes, “it is simply
that I wonder whether such descriptions are themselves enough, whether they do
justice to the immense density . . . of discourse.”25 Although he does argue
against the possibility of a particular sort of subject-centered dis-
course—namely, a humanist account which views the subject as founda-
tional—he makes it clear that his mode of investigation does not constitute “a
rejection of any other possible approach. Discourse in general . . . is so complex
a reality that we not only can, but should, approach it at different levels and with
different methods.”26 Here Foucault indicates that his aim is not to take the con-
cept of the subject out of philosophical investigation altogether; rather, his aim
is to turn philosophical investigation on the concept of subjectivity itself.
Furthermore, although Foucault’s investigation of discourse proceeds without
reference to a subject, the point of this investigation is to “grasp the subject’s
points of insertion, modes of functioning, and system of dependencies.”27 The
analysis addresses the following questions:

how, under what conditions, and in what forms can something like a subject appear in the order
of discourse? What place can it occupy in each type of discourse, what functions can it assume,

122
 THE ANTI-SUBJECTIVE HYPOTHESIS

and by obeying what rules? In short, it is a matter of depriving the subject (or its substitute) of its
role as originator, and of analyzing the subject as a variable and complex function of discourse.28

Again, Foucault indicates here that the goal of his analysis is to resist the temp-
tation to explain discourse in terms of the constituent subject, and instead to
explain the subject as constituted through discourse, to shift the subject from the
explanans to the explanandum of philosophical inquiry. Such a shift allows
Foucault to examine the ways in which discursive practices constitute a histori-
cally, socially, and culturally specific, modern mode of subjectivity.
Although knowledge and discourse are the primary locus of Foucault’s early,
archaeological investigations, in his genealogical works, the locus shifts to in-
clude power alongside discourse as that which serves as a constitutive condition
for modern subjectivity. Perhaps the most striking aspect of Foucault’s analysis
of power is his rejection of the belief that power is solely or even primarily re-
pressive. Foucault insists that power could not possibly be effective if it only
functioned by saying “no.” As he writes, “if power were never anything but re-
pressive, if it never did anything but say no, do you really think one would be
brought to obey it?”29 His analysis of the repressive hypothesis, which I alluded
to above, illustrates this point. Far from being squelched by the repressive power
of moralistic, Victorian attitudes, discourses on sexuality and sexual practices in
our era have actually multiplied. Power does not function in the domain of sexu-
ality merely or even primarily by repressing, prohibiting, censuring, and restrict-
ing; it incites, produces, provokes, and induces; and it prohibits by producing
just as it produces by prohibiting.
For Foucault, as we have already seen, one of the key effects of power’s pro-
ductivity is the subject. However, Foucault’s investigations into the operation of
power are not attempts to write the subject out of philosophical analysis; once
again, they are inquiries into the conditions of possibility for modern subjectivity.
Thus, he characterizes his genealogies as attempts to demonstrate “how it is that
subjects are gradually, progressively, really and materially constituted through a
multiplicity of organisms, forces, energies, materials, desires, thoughts, etc. [His
genealogies] try to grasp subjection in its material instance as a constitution of
subjects.”30 Foucault’s account of disciplinary power in Discipline and Punish
illuminates this process. This account traces the process by which individuals are
subjected to normalizing disciplinary practices and thereby transformed into a
peculiarly modern kind of subject. Individuals are subject to disciplinary power,
which is exercised over them and subtly and insidiously constrains their choices,
desires, and actions, and, at the same time, they are made into subjects by disci-
plinary power, which creates various subject-positions and incites individuals to
take them up. In this way, power both enables the constitution of subjects and
constrains the subject so constituted. In and through its operation, power

123
 AMY ALLEN

constructs modern subjectivity. Foucault’s accounts of power/knowledge do not,

then, eliminate the subject as a crucial concept in philosophical analysis, as his
critics have maintained; instead, they attempt to account for the emergence of the
concept of subjectivity by tracing the trajectory of its construction.
In an essay on Kant’s “Was Ist Aufklärung?” published posthumously in
1984, Foucault summarizes his archaeological and genealogical works and pre-
sents them as attempts to lay out discourse, power, and sexuality as social forces
that construct the modern subject. He differentiates his way of approaching the
constitution and self-constitution of the human subject from Kant’s transcenden-
tal account. The passage is worth quoting at length:

criticism is no longer going to be practiced in the search for formal structures with universal
value, but rather as a historical investigation into the events that have led us to constitute our-
selves and to recognize ourselves as subjects of what we are doing, thinking, saying. In that
sense, criticism is not transcendental . . .: it . . . seek[s] to treat the instances of discourse that
articulate what we think, say and do as so many historical events. And . . . it will not deduce
from the form of what we are what it is impossible for us to do and to know; but it will separate
out, from the contingency that has made us what we are, the possibility of no longer being,
doing, or thinking what we are, do, or think.31

It is interesting that in this passage Foucault characterizes his version of criti-

cism as not transcendental, but only in a very precise sense; his version of cri-
tique is not transcendental because it is fully historicized. Critique involves
laying out the historically, culturally, and socially variable “events that have led
us to constitute ourselves and to recognize ourselves as subjects of what we are
doing, thinking, saying.” In other words, it involves laying out the historically,
socially, and culturally specific conditions of possibility for subjectivity. Thus,
Foucaultian critique is not transcendental in this sense, though it is transcenden-
tal, or at least quasi-transcendental, in another sense, namely, the sense that it
lays out a set of conditions without which there would not be subjects as we cur-
rently conceive of them. In other words, it seeks to specify what Foucault once
upon a time called the “historical a priori”32 of subjectivity in the modern era.
In this passage, Foucault also indicates the payoff of his shift in the explana-
tory priority of the concept of subjectivity. Once Foucault shifts the subject from
explanans to explanandum, the modern formation of subjectivity is revealed as
contingent rather than necessary. For this reason, it is now possible to envision
that subjectivity could be constructed in different—potentially more liberating—
ways. In short, it is possible for prisoners, schoolchildren, factory workers, in-
deed for all of us to be other than we are. From the perspective of those who ac-
cept the anti-subjective hypothesis, this rather optimistic, even emancipatory,
characterization of Foucault’s project seems like too little too late. Because this
way of understanding Foucault takes the elimination of the subject to be the point

124
 THE ANTI-SUBJECTIVE HYPOTHESIS

of his archaeologies and genealogies, his talk in “What Is Enlightenment?” of the

practices that “have led us to constitute ourselves and to recognize ourselves as
subjects of what we are doing, thinking, saying” (emphasis added) seems baldly
contradictory. From the perspective of my reading of Foucault, however, this dis-
cussion is much less problematic. I have suggested that Foucault’s archaeological
and genealogical works should not be understood as attempts to eradicate the
concept of subjectivity; on the contrary, they are attempts to specify the discur-
sive and non-discursive practices that make a particular, and particularly modern,
conception of subjectivity possible; that is to say, they are attempts to specify the
social, cultural, and historical conditions of possibility of modern subjectivity.
Foucault writes the subject out of his analyses of discourse and power simply be-
cause he wants to shift the subject from that which explains to that which must be
explained. And the end result of this shift is a conception of the subject that hints
at the possibility of new and potentially more liberating modes of subjectivity.33
If we accept the interpretation of Foucault that I am proposing, it becomes possi-
ble to make sense of the transition to a discussion of practices of the self in
“What Is Enlightenment?” and his other late works. After Foucault has carefully
laid out the ways in which subjectivity is constructed via discursive and
non-discursive practices, he turns rather naturally to an analysis of how it is that
individuals come to take up, occupy, or inhabit particular subject-positions. From
the perspective of this reading of Foucault, this transition can be seen as a logical
progression in his philosophical project, rather than as the quixotic, contradictory,
incoherent, and ultimately unconvincing move that many critics think it to be.
The reading of Foucault that I am recommending casts him as, in Ian
Hacking’s words, “a remarkably able Kantian.”34 Of course, as any real Kant
scholar will be quick to point out, Foucault’s philosophical project preserves at
most the spirit, not the letter, of Kantian critical philosophy. Foucault himself is
happy to concede this point when he maintains that “the thread that may connect
us with the Enlightenment is not faithfulness to doctrinal elements, but rather the
permanent reactivation of an attitude—that is, of a philosophical ethos that could
be described as a permanent critique of our historical era.”35 Foucault transforms
Kant’s attempt to discover the transcendental conditions of possibility of experi-
ence into the search for the historical, contingent, and hence, mutable, constitu-
tive conditions of possibility of subjectivity. Significantly, this is not a move that
Foucault only makes toward the end of his life and career; as I have shown, he
starts down this path almost at the very beginning of his philosophical project.
Far from eliminating the subject, as the anti-subjective hypothesis suggests,
Foucault writes its history in the modern era.36
Let me emphasize, however that to say that Foucault is interested in laying
out the historically a priori conditions of possibility of subjectivity is not to say
that he is “merely” doing history as opposed to philosophy. Some of Foucault’s

125
 AMY ALLEN

admirers have presented his project in this way, in an attempt to defend it

against critics who accept the anti-subjective hypothesis. For instance, Mario
Moussa has argued that

critics assume that Foucault, as a historian, makes claims about the human subject that must be
taken as objective in the crudest sense: that, for example, if Foucault assembles a genealogy ac-
cording to which subjectivity is a product of repressive self-forming technologies, then the
human being is nothing more than the product of those technologies. But very few historians,
even the most traditional, would understand the activity of writing history in this way; history
never says all there is to say.37

Here, in attempting to defuse the same sort of objection that I have been dis-
cussing, Moussa actually grants the objection but then suggests that it doesn’t
matter because Foucault, like any historian, had to pick and choose what to put
into and what to leave out of the construction of narrative. On this line of argu-
ment, we might say that it so happens that he chose (why? perhaps for reasons
of narrative consistency?) to put into his archaeologies and genealogies a lot of
stuff about discursive structures and power relations and nothing about the role
that individual subjects play in the construction of such structures, but that this
doesn’t necessarily mean that he thought that individuals did not play a role in
that construction.
I think that there may be something to what Moussa suggests, but I have two
worries about this strategy for responding to critics who accept the anti-
subjective hypothesis. The first is that this defense of Foucault does not really
meet the full force of the original objection, which is not just that Foucault
doesn’t talk enough about the concept of subjectivity but that his archaeologi-
cal and genealogical works seem to deny that this concept has any value at all,
to suggest that there is no such thing as subjectivity. I think that the reading
that I have proposed, which interprets Foucault’s archaeologies and genealo-
gies as offering historically and culturally specific conditions of possibility for
subjectivity, helps to explain why this isn’t the case. My second worry is that,
although I agree that Foucault was a very historically minded philosopher, he
was first and foremost a philosopher; analogously, I think that the reason he
brackets the concept of subjectivity in his archaeologies and genealogies is
first and foremost a philosophical rather than a historical one. The reading that
I have proposed also gives us a way to make sense of the philosophical ratio-
nale for this methodological move.

CONCLUSION

By way of a conclusion, let me indicate some of the most important implica-

tions of my rejection of the anti-subjective hypothesis and my defense of an

126
 THE ANTI-SUBJECTIVE HYPOTHESIS

alternate reading of Foucault. First, for Foucault scholars, this way of reading
Foucault holds out the possibility of a new understanding of Foucault’s philo-
sophical project, one which emphasizes the continuity between the three periods
—archaeology, genealogy, and ethics—of his work. Now, the desire to find con-
tinuity in Foucault’s philosophical project may seem quite un-Foucaultian, but
recall that Foucault himself indicated toward the end of this life that he thought
there was such continuity to his work, and he claimed that the continuity could
be found in the fact that “it is not power, but the subject, which is the general
theme of my research.”38 The interpretation of Foucault that I have proposed al-
lows us to make sense of this remark, insofar as it allows us to see the point of
his archaeologies and genealogies as the elaboration of historically a priori con-
ditions of possibility for subjectivity, and his late work as a complementary—
rather than contradictory—account of the role that individuals play in their own
self-constitution.
A second implication of my argument is that it opens up new avenues of
thought with respect to several prominent debates that have surrounded
Foucault’s work in recent years, including the Foucault/Habermas debate and
the debate over the use and abuse of Foucault’s philosophy for feminist theory.
In each of these debates, Foucault’s alleged embrace of the death of the subject
and the implications thereof have been major stumbling blocks to an effective
exchange of ideas. If Foucault’s archaeological and genealogical works can be
read as an elaboration of the background conditions of possibility of subjectivity
rather than as a wholesale elimination of the subject itself, then the prospects for
a reconciliation between Foucault’s view of subjectivity and Habermas’s
intersubjective and dialectical account of individuation through socialization
seem much brighter.39 Although I would not dream of denying that significant
differences between Foucault’s and Habermas’s projects remain, a rapproche-
ment between Foucault and Habermas on the question of the intersubjective
constitution of subjectivity has the potential to reinvigorate the somewhat stag-
nant Foucault/Habermas debate. Foucault and Habermas have offered two of the
most subtle, interesting, and important philosophical analyses of society to be
produced in the latter half of the twentieth century, and yet, in my view, their
work has not been brought into serious and sustained dialogue. Similarly, the
reading of Foucault that I have proposed has the potential to advance the linger-
ing debate over the usefulness of Foucault’s work for feminist theory. Many
feminists, including Alcoff and Hartsock, have cited Foucault’s eradication of
the subject as a serious limitation of the usefulness of his work for feminist the-
ory. But as Nancy Fraser has suggested, “nothing in principle precludes that sub-
jects are both culturally constructed and capable of critique.”40 It is just that
feminists have yet to offer an account of subjectivity that emphasizes both the
way in which subjects are culturally constructed through relations of power and

127
 AMY ALLEN

discourse and the critical capacities that subjects nonetheless (or, as a result)
have. I am not suggesting that the interpretation of Foucault I have offered here
provides, by itself, such an account; but I am suggesting that it provides a fruit-
ful starting point for the construction of such an account, since a greater rap-
prochement between poststructuralist accounts of the cultural construction of the
subject such as Foucault’s and critical-theoretical accounts of the capability of
subjects of critique such as Habermas’s seems like a particularly promising way
of arriving at such an account. The important point is that, on my reading of
Foucault, there is much less standing in the way of such a rapprochement than
has previously been thought.
Finally, the reading of Foucault that I have suggested here has the potential to
open up new ways of thinking about what we might call the structure/agency
problem in social and political philosophy. Philosophers and social and political
theorists have long struggled with the following paradox: on the one hand, if our
theories emphasize the autonomy, agency, and freedom of individuals, then they
run the risk of being blind to the massive impact that social, political, cultural,
and discursive structures have on the very formation of us as individuals; if, on
the other hand, our theories focus on such structures, then they run the risk of
painting an overly deterministic picture of the role that structures play in consti-
tuting individuals, thereby implicitly denying or at the very least undermining
the possibility of individual subjectivity, agency, and freedom. The challenge
that this paradox poses for philosophers is the necessity of finding a way of
thinking about the dialectical interrelationship between social, political, linguis-
tic, and cultural structures, on the one hand, and individual subjects/agents, on
the other. The reading of Foucault that I have suggested here makes it possible
to see his work as an especially productive starting point for a reconsideration of
this paradox because, on my account, far from coming down on the structure
side of this divide, Foucault’s work actually provides some of the theoretical and
conceptual resources necessary for working out the interrelationship between
structure and agency.

Dartmouth College

NOTES

An earlier version of this essay was presented at the Eastern Division Meeting of the American
Philosophical Association in Philadelphia; I am grateful to participants at that meeting for their ques-
tions and comments. Special thanks also to Linda Alcoff, Thomas McCarthy, and members of the
Department of Philosophy at Dartmouth College for their helpful suggestions on earlier versions of
this essay.

128
 THE ANTI-SUBJECTIVE HYPOTHESIS

1 Michel Foucault, “The Minimalist Self,” in Michel Foucault: Politics, Philosophy, Culture, ed.
Lawrence D. Kritzman (New York: Routledge, 1988), 9.
2 Michel Foucault, The History of Sexuality, Vol. 1: An Introduction, trans. Robert Hurley (New
York: Vintage, 1978), 12.
3 See Jürgen Habermas, The Philosophical Discourse of Modernity, trans. Frederick G. Lawrence
(Cambridge, Mass.: MIT Press, 1987), chaps. 9, 10; Axel Honneth, The Critique of Power:
Reflective Stages in a Critical Social Theory, trans. Kenneth Baynes (Cambridge, Mass.: MIT
Press, 1991), chaps. 4–6; Thomas McCarthy, “The Critique of Impure Reason: Foucault and the
Frankfurt School,” in McCarthy, Ideals and Illusions: On Deconstruction and Reconstruction
in Contemporary Critical Theory (Cambridge, Mass.: MIT Press, 1991); Michael Walzer, “The
Politics of Michel Foucault,” in Foucault: A Critical Reader, ed. David Hoy (Oxford: Blackwell,
1986); Charles Taylor, “Foucault on Freedom and Truth,” in Foucault, ed. Hoy; Nancy Hartsock,
“Foucault on Power: A Theory for Women?” in Feminism/Postmodernism, ed. Linda Nicholson
(New York: Routledge, 1990); and Linda Alcoff, “Feminist Politics and Foucault: The Limits to
a Collaboration,” in Crises in Continental Philosophy, ed. Arlene Dallery and Charles Scott
(Albany, N.Y.: State University of New York Press, 1990).
4 Foucault, The Order of Things: An Archaeology of the Human Sciences (New York: Vintage,
1970), 387.
5 Honneth, Critique of Power, 112.
6 McCarthy, “The Critique of Impure Reason,” 59.
7 Alcoff, “Feminist Politics and Foucault,” 71.
8 On this point, see McCarthy, “The Critique of Impure Reason,” 56–58, and Honneth, Critique of
Power, chaps. 4–6.
9 For example, see Foucault, “Power and Sex,” in Politics, Philosophy, Culture, 122.
10 Alcoff, “Feminist Politics and Foucault,” 76.
11 Hartsock, “Foucault on Power,” 163.
12 On this point, see also William Connolly, “Taylor, Foucault, and Otherness,” Political Theory,
13 (1985): 365–76.
13 Foucault is not claiming here that the subject literally invents itself (for clearly this would pre-
suppose a subject who is doing the inventing); instead, he claims that the subject defines itself in
terms of “patterns that he finds in his culture and which are proposed, suggested and imposed on
him by his culture, his society, and his social group.” Foucault, “The Ethic of Care for the Self as
a Practice of Freedom,” in The Final Foucault, ed. James Bernauer and David Rasmussen (Cam-
bridge, Mass.: MIT Press, 1988), 11. On this point, see James Johnson, “Communication, Criti-
cism, and the Postmodern Consensus: An Unfashionable Interpretation of Michel Foucault,”
Political Theory, 25 (1997): 578, n. 20.
14 Richard Bernstein, The New Constellation: The Ethical-Political Horizons of Modernity/
Postmodernity (Cambridge, Mass.: MIT Press, 1992), 154.
15 For variations of this argument, see Jana Sawicki, “Foucault, Feminism, and Questions of Iden-
tity,” in The Cambridge Companion to Foucault, ed. Gary Gutting (Cambridge, U.K.: Cam-
bridge University Press, 1994); Thomas Flynn, “Truth and Subjectivation in the Later Foucault,”
Journal of Philosophy, 82 (1985): 531–40; and Lois McNay, Foucault and Feminism: Power,
Gender, and the Self (Boston: Northeastern University Press, 1992).
16 Flynn, “Truth and Subjectivation in the Later Foucault,” 538.
17 For this argument, see Peter Dews, “The Return of the Subject in Late Foucault,” Radical Philos-
ophy, 51 (1989): 37–41. Bernstein acknowledges this problem as well, when he notes that
Foucault’s notion of ethics “presupposes the notion of an ethical or moral agent that can be free
and that can ‘master’ itself. But Foucault not only fails to explicate this sense of agency, his

129
 AMY ALLEN

genealogical analyses seem effectively to undermine any talk of agency which is not a precipitate
of power/knowledge regimes” (Bernstein, The New Constellation, 164).
18 McCarthy, “The Critique of Impure Reason,” 74.
19 Foucault, “Afterword: The Subject and Power,” in Michel Foucault: Beyond Structuralism and
Hermeneutics, ed. Hubert Dreyfus and Paul Rabinow, 2nd ed. (Chicago: University of Chicago
Press, 1983), 208.
20 Ibid., 209.
21 Foucault, “Two Lectures,” in Power/Knowledge: Selected Interviews and Other Writings 1972–
1977, ed. Colin Gordon (New York: Pantheon, 1980), 98.
22 Foucault, “Truth and Power,” in Power/Knowledge, 117.
23 See, for example, Alcoff’s reading of this passage in “Feminist Politics and Foucault,” 71.
24 Michel Foucault, The Archaeology of Knowledge, trans. A. M. Sheridan Smith (New York: Pan-
theon, 1972), 16.
25 Foucault, The Order of Things, xiii.
26 Ibid., xiv, emphasis added.
27 Michel Foucault, “What Is an Author?” in The Foucault Reader, ed. Paul Rabinow (New York:
Pantheon, 1984), 118.
28 Ibid. See also Foucault, The Order of Things, xiv.
29 Foucault, “Truth and Power,” in Power/Knowledge, 119.
30 Foucault, “Two Lectures,” in Power/Knowledge, 97.
31 Foucault, “What Is Enlightenment?” in The Foucault Reader, 45–46.
32 See Foucault, The Order of Things, for a discussion of the historical a priori.
33 By making the argument that Foucault’s work contains a conception of human subjectivity—
indeed, that the conception of the constituted subject is the focus of Foucault’s work—I do not
mean to argue that this conception is a fully adequate one. In fact, I think it is not. In my view,
even in his late work, Foucault does not give a fully adequate account of the capacity of the
subject for critical agency. Such an account has been offered by Jürgen Habermas. Up to now,
however, the acceptance of the anti-subjective hypothesis has been a major stumbling block to an
effective exchange between Foucaultians and Habermasians; my hope is that debunking the
anti-subjective hypothesis can serve as a necessary first step on the road to a more fruitful
rapprochement between these two positions.
34 Ian Hacking, “Self-Improvement,” in Foucault, ed. Hoy, 238.
35 Foucault, “What Is Enlightenment?” in The Foucault Reader, 42.
36 Derrida makes a similar sort of claim about Foucault in an interview with Jean-Luc Nancy: “we
would appear to have a history of subjectivity that, in spite of certain massive declarations of the
effacement of the figure of man, certainly never consisted in ‘liquidating’ the Subject” (quoted in
Peter Dews, “The Truth of the Subject,” in Deconstructive Subjectivities, ed. Peter Dews and
Simon Critchley (Albany, N.Y.: State University of New York Press, 1996), 149.)
37 Mario Moussa, “Foucault and the Problem of Agency: Or, Toward a Practical Philosophy,” in
Ethics and Danger: Essays on Heidegger and Continental Thought, ed. Arleen Dallery and
Charles Scott (Albany, N.Y.: State University of New York Press, 1992).
38 Foucault, “Afterword,” in Michel Foucault: Beyond Structuralism and Hermeneutics, ed. Hubert
Dreyfus and Paul Rabinow, 209.
39 For this account, see Jürgen Habermas, “Individuation through Socialization: On George Herbert
Mead’s Theory of Subjectivity,” in Habermas, Postmetaphysical Thinking, trans. William Mark
Hohengarten (Cambridge, Mass.: MIT Press, 1992).
40 Nancy Fraser, “False Antitheses: A Response to Seyla Benhabib and Judith Butler,” in Seyla
Benhabib, Judith Butler, Drucilla Cornell, and Nancy Fraser, Feminist Contentions: A Philo-
sophical Exchange, ed. Linda Nicholson (New York: Routledge, 1995), 67.

130