Harry Binswanger

HOW WE KNOW
Epistemology
on an
Objectivist
Foundation

TOF Publications
2014 New York
How We Know

Epistemology on
an Objectivist
Foundation
 Klaus Nordby’s cover illustration
capitalizes upon an analogy mentioned in the book:
gaining conceptual knowledge is like assembling
a jigsaw puzzle; both involve integrating particular
items, according to relationships among
their features, to form an intelligible whole.
In the cover illustration, the puzzle pieces
are analogs of the perceptual concretes integrated by a concept.
Each puzzle piece has a surface pattern; when the pieces are fitted together
correctly, the patterns join to form subtle circles.

Nordby has taken the analogy further: as indicated by the pieces already joined,
the completely assembled puzzle would show that the individual circles form
a grid — a higher-level pattern, in analogy to the all-important higher-level
integrations of conceptual cognition.

Note that the topmost, loose piece does not fit into the nearby slot.
Placing it there would contradict, not complete, the partial circles
showing on the pieces that are already in place. But if one inspects
further, one will find the place where that piece does fit:
the waiting space in which appears the author’s name.

Edition 1.4

Updates and ongoing discussion: www.how-we-know.com

Copyright © 2014 TOF Publications, Inc. All rights reserved.

ISBN-13: 978–0–9856406–1–3

Cover design, book design & typesetting by Klaus Nordby.

Produced with Adobe InDesign CS6. Typeset in Adobe Minion Pro

& Adobe Brioso Pro, both designed by Robert Slimbach.
 CONTENTS

Acknowledgements 11
Preface 13
Abbreviations 19

1
Foundations 21
The First Axiom: Existence 22
The Second Axiom: Consciousness 26
Axiomatic Concepts 28
Consciousness: Four Fundamentals 30
1. Consciousness has an object and a subject 30
2. Existence has primacy over consciousness 32
3. Consciousness is an active process 35
4. Consciousness is a biological faculty 37
Consciousness as Irreducible 42
The Causal Efficacy of Consciousness 49
The Validity of Introspection 54
 H OW W E K N OW 6

2
Perception 57
Perception as Axiomatic 58
Perception vs. Sensation 59
Sensationalism 64
Direct Realism 68
Perception as Inerrant 72
Perception vs. Conceptual Identification 73
Form vs. Object 78
Perception vs. Hallucination 84
A Final Definition of Perception 85
Three General Conceptions of Consciousness 86
1. Objectivism: consciousness as identification 86
2. Naïve Realism: consciousness as reproduction 86
3. Representationalism: consciousness as self-consciousness 88
Appearance and Reality 93

3
Concept-formation 97
Three Theories of Concept-Formation 101
The Realist Theory 101
The Failure of Realism 104
The Nominalist Pseudo-Theory 107
The Objectivist Theory 110
Unit 117
Measurement 119
Integration 123
Unit-Economy 129
 C ontents Contents 7

4
Higher-Level Concepts 137
The Formation of Higher-Level Concepts 139
Wider Integrations 139
Subdivisions 142
Conceptual Hierarchy 143
Abstract vs. Concrete 149
Concepts of Characteristics 152
Concepts of Consciousness 162
Axiomatic Concepts 167

5
Propositions 171
Classificatory Propositions 173
Descriptive Propositions 175
The Role of Grammar 178
Negative Propositions 179
The Analytic-Synthetic Dichotomy 181
Concepts vs. Propositions 183
The Cognitive Function of a Proposition 186

6
Logic: Theory 191
The Three Laws of Logic 192
Logic and the Identity of Consciousness 195
Context 198
Hierarchy 205
The Spiral Process of Knowledge 208
 H OW W E K N OW 8

7
Logic: Practice 213
Logic and Concepts 214
The Function of Definitions 214
The Rules of Proper Definition 215
1. The rule of genus and differentia 216
2. The rule of reference 219
3. The rule of scope 220
4. The rule of fundamentality 222
5. The rule of unit-economy 224
Definitions as Contextual 225
Valid vs. Invalid Concepts 227
1. Concepts that Lack Units 228
2. Concepts that Misclassify Units 229
Rand’s Razor 232
False division 235
False integration 236
Logic and Propositions 239
Propositions as trans-temporal 240
Defective propositions 241
1. Propositions using invalid concepts 242
2. Invalid combination of concepts 243
Truth 251
Logic and Inference 254
Deduction 254
Induction 255
Logic in a Nutshell 259

8
Proof and Certainty 261
Proof 261
Certainty 271
The Law of Rationality vs. The Arbitrary 278
The Ad Ignorantiam Fallacy 285
The Burden of Proof Principle 286
What is Objectivity? 292
A Trichotomy in the History of Thought 296
The Objective Status of Concepts 296
 C ontents 9

9
Principles 299
Fundamentality 300
What Identifying Fundamentals Accomplishes 303
Principles as Fundamental Generalizations 305
The Need for Principles 306
Principles as Cognitive Bridges 307
Principles as Simplifiers 310
Principles as Contextual 311
Principles as Absolutes 312
“Just One Contradiction” 314
An Example: Individual Rights 315
Philosophy and Principles 317

10
Free Will 321
Focus 324
Rationality 326
The Reality-Orientation 328
Motives and Focus 329
The Choice of Action 331
An Extended Concretization 334
Volition and Character 338
Validation of the Theory 344
Determinism: The Denial of Free Will 345
Causality and Volition 347
Focus and “The Problem of Agency” 352
Volition as Axiomatic 355
Determinism as Self-Refuting 355
Free Will and Social Environment 359
Free Will and the Ego 360
 H OW W E K N OW 10

11
Overview 363
Concepts Condense Percepts 365
Two Primacies 366
Concept-Formation in Science 369
1. “Inertia” 370
2. “Natural selection” 370
3. “Germ” 370
Bottom-up vs. Top-down Theories 371
The Prose Principle 376
Some History 381
The Kantian Reversal 385
Summary 390

Bibliography 391
Index 397
 ACKNOWLEDGEMENTS

M y unique and overriding acknowledgment is to Ayn Rand.

Her achievements, and the inspiration provided in her novels, made
possible not only this book but also my career in philosophy, and have
shaped the course of my life.
My education in Objectivism, in philosophy in general, and in the
art of objective writing, was aided immensely over decades by Leonard
Peikoff. He, however, has seen no drafts of this work beyond a first draft of
Chapter 1, many years ago, and he may well disagree with several of the
positions I defend herein.
I am very happy to thank Gregory Salmieri for several extremely helpful
conversations on epistemology, John Ridpath for his insightful comments
on a draft of the book, and Tara Smith for organizing and leading a highly
productive workshop on writing projects. Allan Gotthelf, now painfully lost
to us, was always encouraging and enlightening.
Lee Pierson has earned my enduring gratitude for prodding me to study
J. J. Gibson’s invaluable work on perception and for calling to my attention
elements of sensationalism and representationalism infecting my prior views.
Travis Norsen responded patiently and astutely to my questions on physics.
A very special acknowledgment goes to Shrikant Rangnekar, who met
with me weekly for more than two years, becoming my de facto coach on this
 H OW W E K N OW 12

project. He read over each section that I drafted, made very perceptive and
philosophically informed comments on each page, and in many other ways
acted to shepherd the work, which had been stalled out, to its completion.
Finding a good subtitle was an especially difficult task because it needed
to convey that I am writing as an Objectivist, but not writing on Objectivism,
while also avoiding any suggestion that the book’s content forms part
of Objectivism (Objectivism, being specifically Ayn Rand’s philosophy, is
limited to what she wrote or endorsed). It was Tore Boeckmann who came
up with “Epistemology on an Objectivist Foundation,” a subtitle which is
both smoothly professional and as accurate as a short phrase can be. (The
longer story is that ideas originated by Ayn Rand supply far more than the
founda­tions of this work, as the many passages quoted from her should make
clear.) He also asked me penetrating questions at a very early stage in the
writing, questions that helped me recognize a wrong premise I had accepted,
which was trapping me in a false alternative.
In like manner, I profited greatly from a question that Dina Federman
asked in a graduate seminar I taught in 2002 at the University of Texas. After
I had presented to the class my view of how cognition works, Dr. Federman
raised her hand to ask me: “How is that different from Plato?” The question
startled me, but over time I came to realize that I had indeed fallen prey to
a Platonic error, which led me to get back on the proper, Aristotelian, path.
Several volunteers generously acted as “beta-testers” by reading through
a late draft. They caught many errors and asked for clarification of more than
a few badly written sentences. The book is thus much improved thanks to the
efforts of: Michael Berliner, Tore Boeckmann, Tom Bowden, Chuck Butler,
Bobby Sandler, Dan Sullivan, and Edwin Thompson.
Tym Parsons gave the book a superb, professional line-editing. Stephanie
Bond did a no-holds-barred proofreading of two late iterations of the book,
plus she generously reviewed the virtually final version, as did Richard Witt.
Allison Kunze, in an amazing display of tenacity, skillfully met the arduous
challenge of indexing this densely interwoven material.
Klaus Nordby designed and composed the book, applying his exacting stan­-
dards to every esthetic and typographical issue, large and small. He prepared
the photographs and graphics, conceived and produced the ingenious cover
design (about which, see the copyright page). And on the side, he provided
some philosophically astute copy-editing, repeatedly finding passages that
were unclear or misleading; rewriting these passages always paid dividends.
Finally, to my wife, Jean, I owe more, intellectually and emotionally, than
I know how to express.
 PREFACE

M ankind has existed for 400,000 years. But 395,000 of

those years were consumed by the Stone Age. The factor that freed
men from endless toil and early death, the root cause of the elevated level
of existence we now take for granted, is one precious value: knowledge.
The painfully acquired knowledge of how to master nature, how to organize
social existence, and how to understand himself is what enabled man to rise
from the cave to the skyscraper, from warring clans to a global economy,
from an average lifespan of less than 30 years to one approaching 80.
Though mankind has risen from the cave, things have not been going
well for us lately. The serene confidence of the Age of Reason has given way
to a cultural atmosphere of depression and anxiety — especially among the
intellectuals, who have become convinced that life is “fear and trembling,
sickness unto death.” The art that speaks to modern intellectuals is typified
by Edvard Munch’s painting The Scream and by literature that trumpets the
futility of all human endeavor and celebrates unintelligibility. In 1998, a panel
of literary scholars and authors was asked to pick the top one hundred
English-language novels of the twentieth century. Here is a small taste of the
novel they rated as number one, James Joyce’s Ulysses:
 H OW W E K N OW 14

Im sure thats the way down the monkeys go under the sea to
Africa when they die the ships out far like chips that was the Malta
boat passing yes the sea and the sky you could do what you liked

What explains the critical acclaim for this puerile, subjectivistic chaos
— in an era when technology’s disciplined, structured logic is putting
smartphones into the hands of people around the globe and landing remote-
controlled vehicles on Mars? What explains the wider malaise of our culture?
The two-word answer is: bad epistemology. Epistemology, the theory of
knowledge, is the branch of philosophy that defines the nature, means, and
standards of knowledge. Epistemology deals with the crucial questions:
What is knowledge? How is it acquired? How is it validated? Since knowledge
is man’s means of dealing with reality, a man attempting to function on an
irrational epistemology is unequipped to deal with reality, dooming him-
self to doubt, confusion, and failure. Post-Renaissance philosophers, from
Descartes to Hume to Kant, have spun out ever worse theories of know-
ledge, and the intellectuals are the social group most directly and intensely
affected by philosophical theory. No effective antidote to the epistemological
poison has appeared, so the paradoxical situation described by Ayn Rand half
a century ago rings true today:

If we look at modern intellectuals, we are confronted with the

grotesque spectacle of such characteristics as militant uncer-
tainty, crusading cynicism, dogmatic agnosticism, boastful self-
abasement and self-righteous depravity — in an atmosphere
of guilt, of panic, of despair, of boredom and of all-pervasive
evasion. [FNI, 11]

Our technological success has come from a dedication to reason and logic,
but reason and logic have been distorted or openly attacked by mainstream
epistemologists for the last 200 years, ever since Kant’s theory of knowledge
gained dominance in the intellectual world. Establishment epistemology has
carried to its logical conclusion Kant’s claim that reason cannot know reality.
The result has been two schools of thought, one that accepts reason while
ignoring reality, and one that accepts reality while denying reason.
Rationalism is the school that scorns sensory perception and con-
structs intellectual castles in the air. Empiricism is the school that scorns
abstractions and demands that men hold their minds down to the animal
level of unconceptualized, unintegrated sensing. Rationalism ultimately
 P reface 15

degenerates into mysticism, as in its ancient father: Plato. Empiricism

ultimately degenerates into skepticism, as in its modern father: Hume.
The mystics hold that knowledge can be acquired without any sensory or
rational means; knowledge is said to come from “intuition” or “revelation,”
which washes over us and to which we need only surrender. The skeptics,
observing that men disagree — even about allegedly “revealed truth” — throw
up their hands and announce that there is no truth, that any claim to know-
ledge is proof of dogmatism, and that we are doomed to perpetual doubt.1
In the words of a former chairman of the UCLA philosophy department,
“There are no answers. Be brave and face up to it.”2
Both the mystic and skeptic schools fly in the face of human history.
In the one thousand years ruled by the mystical view, from the fall of Rome
to the end of the medieval era, reliance on alleged revelations and religious
authorities led not to cognitive progress but to stagnation. On the other hand,
since the rebirth of reason in the Renaissance, fueled by the rediscovery of
Aristotle’s works, a vast body of painfully won scientific knowledge — know-
ledge, not mere opinion — has produced our magnificent technological
achievements. The broad record of human history shows that knowledge
is achievable, but only by reason, applied to observational data.
Nonetheless, mysticism and skepticism have lived on, zombie-like, due
to the success of the Kantian attack on reason. That attack has drawn its
power from the errors and concessions in the theories of reason’s defenders
(e.g., John Locke). Lacking a clear, uncompromised understanding of what
reason is and how it operates, epistemology has succumbed to the Kantian
onslaught, leaving men to face the lethal false alternative of mysticism versus
skepticism.
The advocates of reason have been unable to answer the crucial question:
what makes a cognitive choice valid or invalid? Since God or nature
doesn’t tell us how to proceed in our thinking, what standard can we use to
guide our thought processes?
Contrary to the foggy notions of a non-judgmental age, there is a right
and a wrong direction to take — if grasping the facts of reality is one’s goal.
The right direction means the one suited to cognitive success; any deviation

1 In colloquial usage, “skepticism” often means merely a cautious, “show me” attitude,
but in philosophy, “skepticism” means the idea that knowledge is impossible, that man
knows nothing. Of course, that would mean that no one could know that skepticism
itself was true. On the self-refuting nature of skepticism, see Chapter 1, Chapter 5,
and Chapter 10.
2 Donald Kalish, Time Magazine, Jan. 7, 1966, p. 24.
 H OW W E K N OW 16

from that direction is wrong — wrong in relation to that goal, wrong in terms
of the unwaivable requirements of acquiring knowledge of objective reality.
Whether a man wants to know the sum of two plus three, the method of
forging metals, or the principles of a proper political system, to reach the cor-
rect answer he must follow a definite series of steps. But the steps one takes in
pursuing knowledge are not set by instinct, genes, or culture. The course of
a thought-process is up to the thinker to choose (see Chapter 10).
Understanding how knowledge is acquired and validated enables one
to bring the cognitive quest under one’s conscious control and direction,
equipping him to succeed in acquiring knowledge, to avoid whole categories
of error, and to reach objective certainty in his conclusions.
On a wider, cultural scale, the need for a rational epistemology could
not be more urgent. Western civilization itself is now under attack by the
revived mysticism of Christian and Islamic fundamentalism and by the new
skepticism of multiculturalism and postmodernism. The mystics say that
science is wrong — false in its conclusions and blasphemous in its contraven-
tion of the Bible or the Koran. The skeptics say that science is neither right
nor wrong — that truth, falsehood, good, and evil are baseless “constructs”
imposed by a “patriarchal power-structure.”
An open, progressing, benevolent future requires a theory of knowledge
that rejects the false dichotomy that sustains both mysticism and skepticism:
the dichotomy of Empiricism vs. Rationalism. What is required is a theory
that upholds both sensory perception and logic, a theory that shows how
abstract, conceptual knowledge derives in a logical fashion from perceptual
observation.
That theory has been provided by Ayn Rand, especially in her work
Introduction to Objectivist Epistemology. Rand’s definition of “reason” sets
the context for integrating perception and logic: “Reason is the faculty that
identifies and integrates the material provided by man’s senses.” [VOS, 22]
At the base of Rand’s view of reason is her new theory of how abstractions,
i.e., concepts, are formed from perceptual observation. Concepts are the
tools of reason, and it is by means of concepts that man stores and accesses
his knowledge.
The present work makes extensive use of Rand’s Objectivist epistemol-
ogy, as I understand it after fifty years of professional study and teaching.
To a modest degree, I elaborate on and build upon Rand’s system, but
my extensions, even if valid, do not constitute part of the Objectivist
philosophy, which is limited to what Rand wrote, plus those articles by
others that were published under her editorship. (The definitive secondary
 P reface 17

source on Objectivism is Leonard Peikoff ’s consummate work, Objectivism:

The Philosophy of Ayn Rand.)
In this book I do not assume any prior familiarity with the details of
philosophy nor with Objectivism. This book is addressed to the intelligent
layman, assuming he has a definite interest in understanding how we know.
The organization of this book follows one of the cardinal principles of
Objectivist epistemology: knowledge is hierarchical. Chapter by chapter,
I trace the development of progressively more advanced forms of knowledge,
from its base in the axioms of all knowledge, through the fundamental role
of sensory perception, to the formation and use of concepts, through more
abstract concepts, to propositions, and inference — first from the standpoint
of what knowledge is, then from the standpoint of the means of validating it.
After this hierarchical progression, I devote a chapter to Rand’s revolutionary
identification that “man is a being of volitional consciousness” — i.e., that
free will consists in one’s sovereign control over the operation of his own
mind. A concluding overview contrasts the right (“bottom-up”) and wrong
(“top-down”) theories of how we know.
My perspective is causal and biological. Knowledge is an achievement,
one reached by employing certain necessary means, and its purpose is to aid
men in the task of survival.
Knowledge is a product of the wider faculty: consciousness. If one adopts
the causal-biological perspective on consciousness, and applies it to each
of the different functions and levels of awareness, one can gain a crucial,
even life-altering, understanding of the mind and its cognitive needs.
The misunderstandings of consciousness that have wreaked havoc on
the history of philosophy, making philosophy appear irrelevant to daily life,
all stem from taking consciousness to be non-causal and non-biological — or
even, in the latest aberration, non-existent. But consciousness exists, and it
functions according to its nature. Refusing to recognize its existence and its
identity makes men mysterious to themselves. It turns men, in Rand’s graphic
phrase, into “prisoners inside their own skulls.”
To gain self-understanding, one must understand the essence of the self:
one’s mind.
H OW W E K N OW 18
 ABBREVIATIONS

AS Atlas Shrugged, Ayn Rand

CUI Capitalism: The Unknown Ideal, Ayn Rand
FNI For the New Intellectual, Ayn Rand
ITOE Introduction to Objectivist Epistemology, Ayn Rand,
2nd ed. (1990), ed. by Harry Binswanger
JAR Journals of Ayn Rand, ed. by David Harriman
OPAR Objectivism: The Philosophy of Ayn Rand, Leonard Peikoff
PWNI Philosophy: Who Needs It, Ayn Rand
RM The Romantic Manifesto, Ayn Rand
RP Return of the Primitive, Ayn Rand, ed. by Peter Schwartz
VOR The Voice of Reason, Ayn Rand, ed. by Leonard Peikoff
VOS The Virtue of Selfishness, Ayn Rand
H OW W E K N OW 20
 FOUNDATIONS
1
C onsider some examples of knowledge, from the primitive
to the highly advanced. A dog knows where it buried a bone; a baby
knows its mother; a savage knows how to hunt; a student knows the multi-
plication table; a physicist knows the laws of motion.
What do all these examples of knowing have in common? A retained
awareness of some fact of reality. From the dog that retains its awareness
of the bone’s location, to the physicist who has a retained awareness of the
laws of motion, information is stored and can be reactivated — whether the
reactivation is automatic, triggered by sensory cues, as it is for the dog, or is
volitional, as it is when a man asks himself questions in order to bring stored
material back to mind.
Knowledge is not a transient state of awareness, as in viewing the pass-
ing scene from the window of an automobile, but a stable and enduring
mental product — information that you possess, facts that you have gotten
hold of, grasped.
Ayn Rand’s characterization of knowledge summarizes this, and states the
basic means by which knowledge is acquired: Knowledge is “a mental grasp
of a fact(s) of reality, reached either by perceptual observation or by a process
of reason based on perceptual observation.” [ITOE, 35]
 H OW W E K N OW • 1 : F oundations 22

The First Axiom: Existence

Knowledge is of facts of reality, i.e., aspects of existence. The basis and
starting point of all knowledge is the fact that there is a world to be known.
Or, in Rand’s indelible statement, “Existence exists.” [AS, 1015]
The existence of things is perceived directly: we see things, hear things,
feel them, smell them, and taste them. That there is something is perceptually
given; it is not learned by inference from other facts (which themselves
would have to exist). “Existence exists” is a formulation of what is self-
evident. “Self-evident” means: available to direct awareness.
All knowledge, whether perceptual or intellectual, is of something, some-
thing that exists. Any claim to knowledge is a claim to know that something
is the case, that some state of affairs exists.
Accordingly, “Existence exists” is not a derivative or restricted truth but
an axiom: a fundamental, primary, self-evident truth implicitly contained in
all knowledge.
Axioms cannot be proved. This is not a weakness or subjectivity lurking
in them. Axioms are better than proved: they are self-evident. “Existence
exists” does not need to be proved; it is directly perceived. Just open your
eyes, and you know all there is to know about the reality of reality. There is
an unlimited amount to be learned about what exists, the forms and varieties
and aspects of existents, but nothing further to be learned about the fact that
existence exists, nothing beyond what is contained in your first awareness at
the start of your life: “it is.”
Some people demand that axioms be proved. But such a demand fails
to grasp what proof is. “Proof ” is an advanced, not a primary, concept.
It depends upon the prior concept of “existence,” and on an immense body
of other knowledge. Young children and savages have no concept of proof.
All ideas do have to be shown to be valid. But “validation” is a wider idea
than “proof.” There are, broadly, two forms of validation: by proof and by
direct perception.
Proof is a process of inference — deductive or inductive inference. In either
form, inference is a process of moving in thought from something known to
something else logically related to it. An inference is made from something,
not from nothing. Consequently, there must be a starting point. The starting
point of any valid chain of proofs, however long, is the information given
in direct awareness — i.e., the self-evident.
If you see footprints in the sand and conclude that someone has walked
by, that conclusion is reached by inference. But your seeing of the footprints
 T he F irst A x iom : E x istence 23

constitutes direct, non-inferential perception; the presence of those shapes

in the sand is self-evident to you.
As Aristotle observed, it is illogical to hold that absolutely everything has
to be proved. Proof is indispensable when direct observation is not available.
But proof is neither necessary nor possible in regard to the basic information
on which all knowledge is based: perceptual data. As important as proof is,
it is the secondary, not the primary, means of validating ideas. The primary
means is direct awareness.
Self-evidencies, directly perceived facts, are what make proof possible.
To state the point in an extreme form: proof is what we resort to when some-
thing is not self-evident.
And let us ask: why does proof prove? What makes it “work”? Proof
establishes an idea by connecting it to the directly perceived, the self-evident.
To demand, therefore, a proof of the self-evident is an absurd reversal.
Many philosophers dismiss the idea of self-evidency as arbitrary or
subjective. Since I will argue that consciousness and many facts about
consciousness are self-evident, it is important to establish firmly the idea of
self-evidency from the outset.
Although many things have been falsely claimed to be self-evident, in all
such cases, the error lies with what has been taken to be self-evident, not
with self-evidency as such. Again, “self-evident” means: available to direct
awareness. The self-evident is that which makes itself evident by being
directly observed, rather than by being inferred from something else.
“Self-evident” is not a synonym for “obvious.” To one who has learned
arithmetic, it is obvious that two plus two is four, but that truth is not self-
evident; it is inferred by a process of comparison and counting. But that the
page you are reading exists is not an inference: it is self-evident.
The data of sensory perception are self-evident, but the conceptual
interpretation of that data, and inferences drawn from it, are not self-evident.
They must be validated by reducing them back to the self-evident.
The opponents of self-evidency will tell you that in the medieval era
it was self-evident that the world is flat, but we now know it is round. But the
medievals were not able to perceive the shape of the planet. What is given
in perception is a very small portion of the Earth’s surface, and all one can
say about what is given in perception is that the curvature is less than the eye
can detect. The expanse of ocean or prairie one can see is indeed flat — to the
standard of precision given in visual perception. To take a position on what
the shape of the Earth is beyond what perception can reveal is to engage in
either inference or blind guessing. Only astronauts in space are able to see
 H OW W E K N OW • 1 : F oundations 24

enough of the world to have a perceptual experience of its shape, and of

course they see it as round.
Another stock example falls just as easily: the revolution of the Earth
around the sun. It is not self-evident that the sun goes around the Earth, nor
that the Earth goes around the sun. What is self-evident is that the sun and
the Earth are in relative motion. It is not self-evident which frame of refer-
ence, that of the Earth or that of the sun, is the proper one to use in science.
Such alleged counterexamples to self-evidency have nothing in common
with “Existence exists,” which states a self-evident fact, a fact we confront in
every waking moment, from our first acts of awareness to our last.
Although the truth of “Existence exists” does not have to be proved, one
does have to prove that, within the class of truths, it has the special status
of an axiom. Axioms are a special subclass of self-evident truths. Axioms
are self-evidencies that express a primary fact at the base of all knowledge.
Only a very few self-evident truths are axioms. A statement like “The grass
in front of me is green,” though self-evident, is not fundamental and is not
contained in all subsequent claims to knowledge.
The axiomatic status of “Existence exists” has already been demonstrated
by showing that it is an irreducible primary, presupposed by and implicitly
contained in all knowledge. But there is also a specific test of axiomaticity.
Because all knowledge depends on it, an axiom is cognitively inescapable.
As Ayn Rand observes, an axiom is “a statement necessarily contained in
all others, whether any particular speaker chooses to identify it or not.”
[AS, 1040] One cannot think at all without assuming the truth of a genuine axiom.
An axiom is absolutely fundamental. Even the attempt to deny it tacitly
counts on its truth.
For instance, in Ancient Greece, one Sophist announced: “Nothing exists.”
But if nothing exists, then his statement does not exist. He has to assume that
his statement exists in the very act of denying that anything exists. And he
has to assume the existence of much more: of himself, of his understanding
of the meaning of the words he uttered, and of all the learning he had to
accomplish from infancy onward in order to reach the day when he could
make a fool of himself by uttering that self-refuting statement.
This test for axiomaticity is called “re-affirmation through denial,”3 because
the speaker has to implicitly re-affirm the axiom in his attempt to deny it.
“An axiom is a proposition that defeats its opponents by the fact that they
have to accept it and use it in the process of any attempt to deny it.” [AS, 1040]

3 See, for instance, Blanshard, 1939, Vol. 2, 252.

 T he F irst A x iom : E x istence 25

Again, “re-affirmation through denial” does not prove an axiom to be true:

axioms are perceptual self-evidencies, not something proved by anything else.
“Re-affirmation through denial” tests only whether a given statement,
already recognized to be true, is implicitly contained in all others, and thus
has the status, within the class of truths, of an axiom. (In contrast, although
a patch of grass is self-evidently green, saying “This grass is not green” does
not in any way re-affirm its greenness.)
We must be very clear on what “Existence exists” means. It does not mean
that there is some state or property of things constituting their existence.
It does not say, “Things have the property of being” — there is no such prop-
erty. It says that there are existents — things that exist. To posit a property of
“beinghood” characterizing existents is to imply that there are non-existents
which lack this property. But there are no non-existents: what is, is; what is
not, is not. As Leonard Peikoff puts it, “ ‘Existence’ here is a collective noun,
denoting the sum of existents.” [OPAR, 4] “Existence exists” is equivalent to
“All that which is, is.”
In fact, “Existence exists” simply puts into the form of a proposition the
axiomatic concept “existence.” To state “Everything that is, is” is simply
to understand the concept “is.” Or, equivalently, it is to have grasped the
concept “everything.”
Just as one cannot make existence into a property of existents, one can-
not analyze the fact of existence in any way. It is a primary, irreducible fact.
One can analyze a non-axiomatic fact into its components or aspects.
But that process presupposes that those components or aspects exist.
One can analyze lightning into electric discharges, or plants and animals
into cells, but then the electric discharges or cells have to exist. What would
one analyze existence into? That which does not exist?
Thus the ponderous, pretentious silliness of those philosophers who try
to ask: “Why is there something rather than nothing?” I say “try to ask,”
because there is no actual question here, only words having the linguistic
form of a question. Whatever answer one would try to give to this pseudo-
question — call it the X-factor — the X-factor would have to exist, leaving
the pseudo-question unanswered.
A simpler form of the same error can be found in the theist’s pseudo-
question: “What caused the universe?” Here, again, there is no actual question,
because to ask for a cause is to ask for a cause that exists, and the universe is:
all that which exists. When the theist uses this pseudo-question to “prove” the
 H OW W E K N OW • 1 : F oundations 26

existence of a God, it does not occur to him that, by his logic, he must then
ask: “What caused God?”
We may ask for a causal explanation of any particular thing within existence,
but it is nonsensical to ask for a cause of exis­tence as such.
Existence, the universe, is a self-sufficient primary. There is nothing that
causes there to be something rather than nothing; there is nothing prior
to existence, beneath existence, or outside of existence. Existence exists,
and only existence exists. What is not, is not.
Nor can there be an origin of the universe or an end to the universe.
Particular arrangements of elements within the universe — arrangements
constituting ourselves, planets, even galaxies — can come and go. But the uni-
verse — the entirety of that which is — neither came into being (from what?)
nor can go out of existence. There was no time at which nothing existed;
the universe is not in time or space; time and space are relationships among
things within the universe.
In other words, existence exists. Period.

The Second Axiom: Consciousness

The axiom of existence is the base of all our knowledge. But it is not the only
axiom. The second axiom, implied in one’s awareness of existence, is that one
is conscious.
Consciousness is the faculty of awareness. Since knowledge is a retained
awareness, to set the context for investigating knowledge, I need to discuss
the axiom of consciousness in considerable detail.
An axiom is a truth that is cognitively primary, self-evident, and stands
at the basis of knowledge. It is easy to show that “I am conscious” possesses
each of those four characteristics of axiomaticity.
1. The fact that one is conscious is a truth; indeed consciousness, like
existence, is a foundation of truth: truth pertains to a certain relationship
between consciousness and reality. External existents, apart from any rela-
tionship they have to consciousness, are not “true” or “false” — they just
are; truth and falsehood pertain to something mental — an idea, statement,
proposition — in its relation to the external world.
2. The fact that one is conscious is a cognitive primary: for newborns,
or even in the womb, cognition begins with being conscious of some-
thing — of a pressure, a temperature, etc. — and there is no cognition prior
to being conscious.
 T he Second A x iom : C onsciousness 27

3. The fact that one is conscious is self-evident: it is directly experienced,

not inferred. To be sure, one’s direct awareness of one’s consciousness is not
sensory, as one’s direct awareness of existence is. One does not see or
touch or taste one’s awareness. But, from a very young age, one is directly,
non-inferentially aware that in sense-perception, and later in more complex
mental activities, one is aware. Self-awareness is a given for man.
4. The fact that one is conscious is the base of all knowledge: knowledge
is a phenomenon of consciousness. To know by means of unconsciousness is
a contradiction in terms. Roses and rocks do not have knowledge.
“I am conscious” passes the test of axiomaticity: re-affirmation through
denial. “I am not conscious” is a self-refuting statement. Non-conscious
entities exist, but they are not conscious of being non-conscious. Any attempt
to consider the possibility of one’s not being conscious presupposes that
one is conscious to consider it. Stones do not consider, question, believe,
or doubt anything. There are no stone philosophers to assert or deny that
they are conscious.
Again, passing the test of “re-affirmation through denial” does not prove
that one is conscious — one already knows that (and proof itself is an
action of consciousness). Rather, the implicit re-affirmation of conscious-
ness in any attempt to deny it establishes the epistemological status of “I am
conscious” — it shows that this is not a limited or derived truth, but an axiom.
Just as there are thinkers who, explicitly or implicitly, attempt to deny
existence, so there are those who attempt to deny consciousness. Often the
rejection of consciousness is explicit, as it was in the case of the ­Behaviorist
psychologists (John Watson4 and B. F. Skinner being the arch-examples).
If Martian philosophers existed, there would be no self-contradiction in
their denying that Earthlings are conscious; but any behaviorist who admits
to membership in the human race is contradicting himself by claiming
to be aware that man is unconscious. The fact that behaviorists consider
themselves to be scientists merely compounds the contradiction: science is
a highly advanced product of consciousness. (A comprehensive autopsy of all
forms of materialism — the doctrine that consciousness does not exist — will
come at the end of this chapter.)
Consciousness, unlike existence, is a property: “Consciousness is an
attribute of certain living entities, but it is not an attribute of a given state
of awareness, it is that state.” [ITOE, 56] Just as existence is not something
distinguishable from, added to, or underlying the various things that exist,

4 “[C]onsciousness is neither a definite nor a usable concept. . . . [B]elief in the existence of

consciousness goes back to the ancient days of superstition and magic.” [Watson, 4]
 H OW W E K N OW • 1 : F oundations 28

so consciousness is not something distinguishable from, added to, or under-

lying the various states of awareness. Just as a chair’s existence is not some-
thing distinguishable from, added to, or underlying the chair, so the aware-
ness provided by seeing the chair is not something distinguishable from,
added to, or underlying the seeing. To be a chair is to exist; to see a chair is
to be aware of it.
One cannot analyze the state of awareness as such. It is a primary,
irreducible fact. There are no “consciousness atoms” to analyze the state of
consciousness into. (The belief that awareness can be analyzed into neural
states or processes will be refuted a little later.)
It is the state, not the faculty, of consciousness that is irreducible and
unanalyzable; the faculty of awareness has a physical side — the nervous
system — which, of course, can be analyzed physically and physiologically.
But what it is to be aware of something, rather than being unaware, is a
primary fact that cannot be analyzed.
The varieties and forms of states of awareness can be distinguished —
as can the varieties and forms of existence. One can distinguish pain from
pleasure, and both from memory or imagination — just as one can distin-
guish, in regard to existence, tables from rocks, and both from molecules
or galaxies. But one cannot ask “what is it to exist?” nor “what is it to be
aware of something?” Existence and consciousness are irreducible primaries.

Axiomatic Concepts
Ayn Rand discovered that beneath axiomatic propositions lie axiomatic
concepts. “Axioms are usually considered to be propositions identifying
a fundamental, self-evident truth. But explicit propositions as such are not
primaries: they are made of concepts.” [ITOE, 55]
A proposition is a statement, such as “Cats are animals.” A proposition
applies some predicate (e.g., “animal”) to some subject (e.g., cats). Thus, the
general formula for a proposition is “S is P.” To form or grasp a proposition,
one has to know what the S and the P refer to — i.e., one has to have the
concept of “S” and the concept of “P.” If a child does not have the concept
“cat” and the concept “animal,” he cannot form or understand the proposi-
tion “Cats are animals.” Propositions presuppose concepts. Axiomatic
propositions presuppose axiomatic concepts, with the implication Rand
draws: “The base of man’s knowledge — of all other concepts, all axioms,
propo­sitions and thought — consists of axiomatic concepts.” [ITOE, 55]
 A x iomatic C oncepts 29

In particular, the base of the axiomatic proposition “Existence exists”

is the axiomatic concept of “existence.” The base of the axiomatic proposition,
“I am conscious” is the axiomatic concept of “consciousness.”
The identification that axiomatic concepts, not axiomatic propositions,
are the base of all knowledge resolves a long-standing dilemma about
axioms. This dilemma can be put as follows. On the one hand, an axiom is
a primary — something that one must know before one knows anything
else. On the other hand, knowledge of the axioms as propositions is in fact
gained very late in one’s development, if ever. Clearly, the infant in its crib
is not thinking to itself: “Existence exists,” or any such sentence, before it
has language. So if axioms are propositions, sentences, or thoughts, they are
not available to a beginning knower. And, in fact, the vast majority of men
have lived their whole lives never having heard or thought such propositions.
So, if we need the knowledge of axioms in order to go on to further know-
ledge, how can that be reconciled with the preceding?
This dilemma dissolves once one recognizes that the required knowledge
is not propositional knowledge. Propositions connect two separate items,
a subject and a predicate that is distinct from it, as in “Dogs have tails.”
But axiomatic propositions, as Rand notes, merely restate the fact that its
subject term refers to. “Existence exists” merely puts into propositional form
a single, unanalyzable fact — the fact that something is there, it exists.5
The fact of existence is grasped not through a proposition but perceptually.
In the infant’s first perception and in every succeeding one, it is aware
of something that exists. Later knowledge merely names what is grasped,
wordlessly, from the beginning. As Rand puts it, the infant from the outset
has the axiomatic concept implicitly. It knows exactly what the philosopher
knows when he uses the sophisticated term “existence.” There is nothing
more to be learned about the basic fact that things are. As far as the fact
of existence goes, once you’ve seen one existent, you’ve seen them all.
Axiomatic concepts, held implicitly, are thus the base of cognitive develop-
ment. Rand writes:

The building-block of man’s knowledge is the concept of an

“existent” — of something that exists, be it a thing, an attribute or

5 Likewise, “Consciousness is conscious” and “A is A” are restatements of the (implicit)

concepts of “consciousness” and “identity.”
 H OW W E K N OW • 1 : F oundations 30

an action. Since it is a concept, man cannot grasp it explicitly until

he has reached the conceptual stage. But it is implicit in every
percept (to perceive a thing is to perceive that it exists) and man
grasps it implicitly on the perceptual level — i.e., he grasps the
constituents of the concept “existent,” the data which are later
to be integrated by that concept. It is this implicit knowledge that
permits his consciousness to develop further.6 [ITOE, 5–6]

Consciousness: Four Fundamentals

Of the properties of consciousness, four are fundamental and undeniable.7

1. Consciousness has an object and a subject.

2. Existence has primacy over consciousness.
3. Consciousness is an active process.
4. Consciousness is a biological faculty.

I will discuss each in turn.

1. Consciousness has an object and a subject

A state of consciousness is an awareness of something, by some organism.
Consciousness is an activity involving a relation of a subject to an object.
Consciousness implies an object, that of which one is aware. One can’t see
without seeing something; one can’t think without thinking about something.
For one to be aware, there must be the what, the object that one is aware of.

Some object, i.e., some content, is involved in every state of

awareness. . . . Awareness is awareness of something. A content-
less state of consciousness is a contradiction in terms. [ITOE, 29]

6 There are other axiomatic concepts besides “existence” and “consciousness.” Notably,
the concept of “identity” is axiomatic, and is expressed propositionally in the statement:
“A thing is what it is,” or “A is A.” Rand describes the concept of “identity” as a corollary
of the concept of “existence.”
7 The first three of these facts are self-evident; the biological nature of consciousness
is a fact that is reached by inductive inference.
 C onsciousness : F our F undamentals 31

What if one imagines something that doesn’t exist, such as a golden

mountain? There is still content: one is imagining something, not nothing.
One is imagining a golden mountain, not a nothing; to imagine a golden
mountain is to imagine that, not something else.
The materials used in imagination come from one’s past perception of
reality, here of gold and of mountains. Imagination is the ability to mentally
combine and rearrange such materials. Imagination, in stark contrast to
perception, is under direct volitional control: you can visualize whatever you
like, but you can see only what is there to be seen.
If we liken imagination to painting something in the mind, the pigments
have to exist or else there can be no painting; the materials of imagination
have to come from memory of things taken in from reality. One cannot
construct a fantasy out of sensory qualities that one has never experienced.
We cannot, for instance, visualize ultraviolet, since its wavelength is outside
the range of human vision. Men blind from birth do not have a visual form
of imagination; they do not even have visual dreams.8
As Rand states, “Directly or indirectly, every phenomenon of conscious-
ness is derived from one’s awareness of the external world.” [ITOE, 29] All
consciousness is consciousness of something.
Consciousness also involves a subject, which is the man or animal that is
conscious.
Consider how one forms the concept “consciousness.” One forms the
concept by reflecting on one’s own mental actions, to integrate one’s own
seeing, hearing, thinking, remembering, imagining, etc.; then one includes
the similar conscious activities of other men and of the higher animals.
(One cannot experience or directly apprehend any consciousness other than
one’s own.)
Philosophers sometimes ask, following David Hume, “How do I become
aware of the self? Introspection gives me only concrete mental acts, not a self.”
The answer is: it is you introspecting, you observing your consciousness
of something.9 It is not the case that one discovers that consciousness in general
exists and then wonders whether it is attached to a person. One first becomes

8 Imagined content is usually compared to perceptual content, but it would seem to be

more closely related to the content of memory. Imagining an absent friend’s face seems
to have just the same inner quality or “feel” as remembering it, which is in line with the
point that imagination is the ability to rearrange stored perceptual data.
9 In addition, a direct experience of the self comes with one’s exercise of free will.
The difference between what one does and what merely happens to one is an experience
of the self, the ego, as the cause of one’s actions. (See Chapter 10)
 H OW W E K N OW • 1 : F oundations 32

aware of one’s own consciousness and then infers that other people and the
higher animals are also conscious.10
Rand’s summarizing statement is:

Existence exists — and the act of grasping that statement implies

two corollary axioms: that something exists [the object] which
one perceives and that one exists [the subject] possessing
consciousness, consciousness being the faculty of perceiving that
which exists. If nothing exists, there can be no consciousness:
a consciousness with nothing to be conscious of is a contra­diction
in terms. [AS, 1015]

2. Existence has primacy over consciousness

The fact that consciousness has an object means that consciousness cannot
be self-contained. Consciousness is inherently something that points outside
itself, to something else: its object.11
This leads to the principle Ayn Rand named “the primacy of existence”:
existence has primacy over consciousness. Consciousness is a secondary
phenomenon: in order for an organism to be conscious of something, that
something first has to exist.
The primacy of existence is the recognition that existence is independent
of consciousness; things exist, and are what they are, whether or not any
organism is conscious of them. By the same token, consciousness is depen-
dent on existence: consciousness has to have an object rather than being
purely self-contained.
(The contrary position, “the primacy of consciousness,” reverses this order,
treating consciousness as the self-sufficient primary, relegating existence
to the status of a derivative of consciousness — as in the idea that God wished
the universe into being.)
Epistemologically, the primacy of existence is the recognition that
existence must be known before consciousness is known, and that existence
is known by extrospection, not introspection. [See PWNI, 24–25]

10 The basis of inferring that other people and the higher animals are conscious is, obviously,
their anatomical and behavioral similarities to oneself.
11 The case of introspection, which involves consciousness of consciousness, will be
discussed shortly.
 C onsciousness : F our F undamentals 33

From one’s first grasp in early childhood of any action of one’s conscious-
ness, one learns that existence, the object of awareness, is independent of that
awareness. Peikoff writes:

From the outset, consciousness presents itself as something

specific — as a faculty of perceiving an object, not of creating or
changing it. For instance, a child may hate the food set in front of
him and refuse even to look at it. But his inner state does not erase
his dinner. Leaving aside physical action, the food is impervious;
it is unaffected by a process of consciousness as such. [OPAR, 18]

One must (implicitly) accept the primacy of existence in order to grasp

any concept of consciousness — e.g., “seeing,” or “thinking.” Such con­­cepts
require distinguishing between one’s awareness and the object of which one is
aware — e.g., between the act of seeing and what is seen — and the only means
of doing so is to notice what depends on us and what exists independently
(e.g., by observing what happens on closing and re-opening one’s eyes).
The primacy of existence is fully consistent with the fact that self-­
consciousness exists: a consciousness can turn back on itself to make its own
activities into an object. One can certainly be conscious of one’s conscious-
ness, through introspection. But introspection presupposes extrospection.12
One can introspect only after one has perceived things; until then, one is
not conscious, so there is nothing to introspect. Consciousness precedes
self-consciousness.

A consciousness conscious of nothing but itself is a contradiction

in terms: before it could identify itself as consciousness, it had to
be conscious of something. [AS, 1015]

To “identify itself as consciousness” requires making the distinction

between subject and object, between self and the world, which presupposes
that there is a world.
The opposite view, the primacy of consciousness, was injected into
post-Renaissance philosophy by Descartes. Although he recognized that

12 The introspected object, e.g., a thought, must have had its own object. A thought must be
a thought of something. (The object of thought does not have to be an external existent:
one can think about a dream one had; but any dream is of something, and its content is
ultimately derived from perception of reality.) No matter how many levels of “conscious-
ness of consciousness of . . .” one adds, the final “of ” requires an object.
 H OW W E K N OW • 1 : F oundations 34

consciousness must have an object, he asserted the possibility that this object
might itself be mental, not external. He took the existence of his conscious-
ness to be axiomatic, but the existence of existence to be non-axiomatic,
problematical. He asked, in effect, “How do I know that there is a world
outside my mind? What if all that I am ever aware of are experiences inside
my own mind, not external reality?”
But to identify something as “an experience in my mind,” I have to contrast
my mental experiences with something else. Without the contrast between
the internal and the external, “internal” loses its meaning. “Everything
is internal” is an incoherent statement, one that contains an implicit
contradiction. “Everything is only in my mind” likewise renders “my mind”
meaningless. It is only the contrast between existence and consciousness
that makes the concept “consciousness” possible.
The logical fallacy exemplified here was first identified by Ayn Rand.
She called it “the fallacy of the stolen concept.” [AS, 1039]A thorough treatment
of the stolen concept fallacy will have to wait until Chapter 7, but it is worth
pausing here to give a preliminary account.
The stolen concept fallacy consists of a certain kind of violation of the
hierarchy of concepts. Concepts have to be formed in a certain order,
and their meaningful use depends on not violating that order. A child’s first
concepts, such as “dog,” are formed from sense-perception; then more
advanced concepts, such as “animal” and “pet,” are formed on the basis of
the prior concepts, creating a hierarchy, in which some concepts depend on
others. For instance, a child cannot grasp “pet” before he grasps “animal.”
(I am referring to the grasp of a concept, not merely the uttering of a word.)
But suppose a demented philosopher announces: “Pets exist, but animals
do not.” In violating the necessary hierarchy of concepts, his statement wipes
itself out. Obviously, if there are no animals, there are no pets, since a pet
is “any domesticated or tamed animal that is kept as a favorite and cared for
affectionately.” [Random House College Dictionary, 1980]
On the other hand, if one says only: “There are no such things as pets,” one
has made a false statement but has not “stolen” any concept. The concept-
stealing here occurs when one attempts to retain the concept “pet” while
denying the hierarchically prior concept “animal.” Doing that “steals” the
concept “pet” — i.e., uses “pet” without any logical right to do so.
The uniquely perverse nature of the fallacy of the stolen concept, in the
form it usually occurs, is its attempt to use a concept in the very act of
negating that concept’s own base — thereby sawing off the cognitive branch
one is sitting on. Consider the statement: “It has been proved mathematically
 C onsciousness : F our F undamentals 35

that there are no such things as numbers.” If there are no numbers, there
is no science of mathematics and no such thing as a mathematical proof:
one can grasp “mathematics” only on the basis of first grasping “one,” “two,”
and other numerical concepts.
All the versions of the primacy of consciousness that litter the history
of philosophy “steal” the concept of “consciousness,” or some particular
concept pertaining to consciousness. For, as Rand observes, “It is only in
relation to the external world that the various actions of a consciousness
can be experienced, grasped, defined or communicated.” [ITOE, 29] To grasp
“consciousness,” one must distinguish actions of consciousness from their
objects — i.e., from things that exist independently. When a Cartesian says,
“Maybe nothing exists outside my mind,” he has “stolen” the concept “mind,”
depriving it of any meaning, just as if he had said, “Maybe the entire universe
is indoors,” an utterance that renders “indoors” meaningless.
The recognition that existence is independent of consciousness, and that
consciousness is the faculty of perceiving it, is the base of all further discus-
sion of consciousness — in fact, of all further discussion, period.

3. Consciousness is an active process

So far, I have noted that consciousness is an organism’s awareness of
something that exists. The next point is that consciousness is an activity,
an ongoing, continuous process of interaction with the world. This marks
a fundamental re-orientation, rejecting the static view of consciousness, the
view of Plato and the mystics. Rand gives a clear statement of the right view:

Awareness is not a passive state, but an active process. On the

lower levels of awareness, a complex neurological process is
required to enable man to experience a sensation and to inte-
grate sensations into percepts; that process is automatic and
non-volitional: man is aware of its results, but not of the process
itself. On the higher, conceptual level, the process is psychological,
conscious and volitional. In either case, awareness is achieved and
maintained by continuous action. [ITOE, 29]

Let’s look at this point on both the conceptual and the sensory levels.
The briefest introspection shows that conceptual awareness involves action.
Thinking is an activity, and it is not possible to freeze a thought and hold it
still. Perform this experiment. Think, “Triangles have three sides,” and try to
 H OW W E K N OW • 1 : F oundations 36

hold your mind locked on that single thought. It cannot be done; your mind
automatically starts moving, with an implicit “now what?” and “so what?”
You can look more closely at what you mean by triangles having three sides,
you can picture a succession of triangles, you can repeat the same words
over and over (though doing that soon turns them into meaningless sounds).
What you cannot do is stop “the stream of consciousness,” as William James
phrased it. In regard to the conceptual level of awareness, we are aware by
direct introspection that consciousness is an activity. And, of course, neuro-
physiological action underlies conceptual processes.
In contrast, sense-perception may seem static. As you look at the page
you are reading, it is simply there, motionless, for as long as you look.
From science, however, we know that your nervous system has to be engaged
in constant physiological action in order for you to have the seemingly static
­perception. But even pre-scientifically, what is evident to anyone is that
perception requires the active use of our senses to explore the world, gather
infor­mation, and make discriminations. Perception is not the passive regis-
tration of momentary input. In viewing this book page, or any other object,
your relationship to it has to change if your awareness of it is to continue.
A constant background odor ceases to be experienced after a while; a con-
stant low-level hum in the background drops out of awareness (though one
can introduce a change by an act of attention, in which case it pops back into
awareness).
Consciousness requires contrast, change, difference. Consciousness is
a difference-detector.13 The primary function of consciousness is to differen­
tiate, which is an active process.
The fact that consciousness is active does not contradict the primacy of
existence (nor imply that consciousness is somehow invalid). Rand’s impor-
tant aphorism makes the necessary distinction: “Consciousness is meta-
physically passive, but epistemologically active.”14 That is, consciousness
does not create or alter its object (consciousness is passive, metaphysically),
but awareness is achieved by an active process (consciousness is active,
epistemologically).

13 This is not meant as a definition of consciousness, but as a statement about its biological
function and terms of operation. “Detect” here means “have discriminated awareness
of,” so “detector” cannot be used to define consciousness. In fact, consciousness, as an
axiomatic concept, can be defined only ostensively, by providing examples, such as:
your seeing this page, your thinking about these ideas, your remembering what the
previous paragraph said, your feelings toward your friends, and so on.
14 Reported by Leonard Peikoff (in personal conversation).
 C onsciousness : F our F undamentals 37

4. Consciousness is a biological faculty

Consciousness is not only an action but also a living action. Living action is
goal-directed. [Binswanger, 1990 & 1992] An organism’s actions are adapted to
securing its survival. Consciousness, like the heartbeat, is a biological activity
that evolved because it promotes survival. But few philosophers in history
have regarded consciousness that way.
The Judeo-Christian thinkers regard “things of the spirit” and “things of
the flesh” as opposites, as belonging to different realities. Consciousness is an
implant of the supernatural in man, religionists proclaim, and consciousness
offers man, in their view, nothing but an intractable conflict with his “all too
human” body.
Even Plato was attracted to this view: he describes the body as the tomb
of the soul. But to Aristotle, this was nonsense: he recognized that the
body is a living body and that the soul (i.e., consciousness) is the “form”
(“entelechy”) of the body — i.e., an expression of bodily powers. Death, he
recognized, ends the life of body and soul. Human consciousness is an activ­
ity of a person, involving his body’s interaction with the external world.
The living body is not the tomb of anything; it is the enabler of consciousness.
And when a person dies, what had been his body becomes a corpse.
Where does the soul go when you die? To the same place as your heartbeat.
Conscious activities, whether sensory or conceptual, have, like the heart-
beat, a biological function. Man has eyes for the same reason he has a heart:
to sustain his life; vision is an adaptive, biological, life-sustaining capacity.
The same is true for the other sense modalities: each provides a man with
life-sustaining information about the world.
And the same is true of the faculty of reason. The mind, the reasoning
intellect, is a vital organ. A biologist could not understand the heart if he
did not know its biological function, and a philosopher cannot understand
reason, or any other faculty of consciousness, if he ignores the biological
function of that faculty.
The heart serves the organism’s survival by circulating the blood. In what
way does consciousness serve survival? What does sight, for example, do for
sighted animals?
First, note what makes an animal an animal. In simplest terms, the
distinguishing characteristics of animals are the faculties of locomotion
and consciousness. In contrast to a plant, an animal perceives the world
and moves itself through the world. But the deeper issue concerns how the
animal makes its living — how it gets nourishment.
 H OW W E K N OW • 1 : F oundations 38

Plants synthesize their own nutrients; animals feed on plants or on other

animals which, ultimately, have fed on plants. Animal life depends on plants
having photosynthesized the basic nutrients that animals need.
Putting it as simply as possible: animals eat. If they eat plants, the food
supply within reach is soon exhausted, so they must move around: they graze.
If they eat other animals, they need to catch their prey, which again means
they must move around: they hunt. Whether they graze or hunt, animals
need to find their food. Consciousness is their means of doing so.
(There is an exception to animal motility, and a telling one: There are some
sea-animals, such as oysters, that do not move themselves through their
environments during most of their lifespans, since their food floats to them.
And oysters have no eyes.)15
In general, animals have to move to get food; consciousness enables them
to locate their food. It also enables them to avoid being eaten, but food is the
fundamental: life is not fundamentally the avoidance of death but the gaining
of the materials for self-sustenance. Consciousness does also enable animals
to attain other goals, e.g., to find mates for reproduction, but getting food
is the fundamental.
That’s the simplified overview. Now let’s look in more detail.
Consciousness does several things, each of which contributes to the
organism’s survival.
1. Consciousness enables the animal to integrate all the various parts of
its body to pursue its overall goal in relation to the perceived environment
as a whole. When the lion undertakes the chase, all its muscular activity
is coordinated to that single effort. And the lion chases its prey through
a terrain, not as a simple stimulus-response mechanism. A plant’s parts react
“locally” — the leaf may curl to preserve heat on a cold day, but a plant cannot
pull up roots and move to a warmer locale.
2. Consciousness enables the animal to bridge space, in the sense that the
animal can respond to distant objects. The lion sees and smells its distant
prey, crouches down, and begins to stalk.
3. Consciousness enables the animal to bridge time, by responding now
and over a span of time to a goal that it will not reach and utilize until later.

15 If you can imagine an oyster with eyes, an oyster that lies motionless on the sea bottom,
passively watching the passing scene, year in and year out, you have the exact opposite
of my view of consciousness. I am not merely saying that eyes would have no survival
value for oysters (and so did not evolve); I am saying that vision without locomotion
would not actually be vision. Vision is connected with and depends upon self-produced
movement (see Chapter 2).
 C onsciousness : F our F undamentals 39

The lion beginning its stalk crouches down now in order to capture its prey
some minutes later.
4. Consciousness enables the animal to guide its actions according to the
continuing changes in its goal and the requirements of reaching it. The lion
uses an integrated perceptual awareness, involving sight, smell, and hearing,
to adjust to the changing position of its prey in the perceived terrain.
5. Consciousness enables the animal to expand the range of its action:
the organism as a whole is sensitive to minute changes in its environment.
The lion’s prey (switching here to its point of view) sees the rustle of a few
stalks of grass, catches the lion’s scent, feels fear, and bolts away. Perceptual
awareness enables the animal to respond not just to separate stimuli but
to the whole situation in the whole environment.
6. Consciousness enables the animal to learn — i.e., to acquire new
­knowledge. The ability to learn greatly reduces the time required for the
adaptation of the organism to its environment. Instead of this adaptation
requiring natural selection over hundreds or thousands of generations,
an animal can learn in seconds what is the survival-significance of a novel
stimulus and can adjust its behavior accordingly. The adjustment is also
swiftly reversible: what formerly meant danger but now is safe can be treated
accordingly.
Here, on the simple animal level, we see that consciousness involves
cognition, evaluation, and the initiation of bodily action. Consciousness is
distinguishable into these three functions (and others), but the three occur
together, as parts of an inseparable whole: the faculty of consciousness.
As philosopher Hans Jonas writes:

Three characteristics distinguish animal from plant life: motility,

perception, emotion. . . .
Fulfillment not yet at hand is the essential condition of desire,
and deferred fulfillment is what desire in turn makes possible.
Thus desire represents the time-aspect of the same situation of
which perception represents the space-aspect. Distance in both
respects is disclosed and bridged: perception presents the object
“not here but over there”; desire presents the goal “not yet but to
come”; motility guided by perception and driven by desire turns
there into here and not yet into now. [Jonas, 101]

Most philosophers in the history of philosophy have downplayed or even

denied one or more of these three functions of consciousness. Even the
 H OW W E K N OW • 1 : F oundations 40

pro-consciousness philosophers have typically regarded consciousness as

nothing but cognition, as if values and action were incidental or dispensable.
But if we are to understand consciousness, we must never lose sight of the
biological fact that consciousness informs the organism about its environ-
ment for the sake of motivating, sustaining, and directing action, the action
its survival requires.

Knowledge, for any conscious organism, is the means of survival;

to a living consciousness, every “is” implies an “ought.” [VOS, 24]

And if we are to understand man’s consciousness, we must bring the

same biological perspective to our consideration of his distinctive attribute:
his rational faculty. Man’s capacity of abstraction and thought is the product
of natural selection. Each genetic variation in the makeup of our anthropoid
ancestors that enhanced their brain-power gave them a survival advantage.
Man’s reasoning mind is a survival instrument, just as his heart and liver
are. The ability to abstract, conceptualize, and think is not only pro-survival,
it is man’s basic means of survival.
Whether devoted to building a hut or measuring the speed of light, human
thought is, in its biological origin and essential function, a tool of survival.
Yes, man can misuse his mind — he can sever the connection of his mind to
reality and drift among imaginary “constructs” of his own devising; but he
can misuse any part of his body, too. The mind, like the body, is an instru-
ment of survival, despite the fact that man does not automatically treat
it as such.
The failure to adopt this biological perspective has crippled philosophy,
preventing man from properly understanding his most vital organ: his mind.
Philosophers have spun out theories that treat the mind as a self-contained
phenomenon, ignoring its roots in and dependence on perception, emotion,
and action in the world. The disasters that stem from ignoring the biological
role of reason will become apparent as this book proceeds.
The nonbiological perspective stands markedly revealed in the common
question: is it possible to develop a computer that can think? My answer is:
before a computer could think, it would have to be able to understand ideas
(concepts); before it could understand ideas, it would have to be able to
perceive the world and to feel emotions, such as pleasure and pain, desire and
fear; before it could perceive and feel emotions, it would have to be alive —
i.e., be engaged in action to sustain itself. We can dismiss notions about
 C onsciousness : F our F undamentals 41

a thinking computer until one is built that is alive — and then it wouldn’t
be a computer but a living organism, a man-made one.
Biologically, seeing is for moving, ideas are for doing, theory is for practice.
A word of caution is needed here. The philosophy known as “Pragmatism”
is merely the other side of the same false alternative: cognition vs. action.
In pseudo-rebellion against divorcing ideas from action, Pragmatists
divorce action from ideas. An idea, they say, is nothing but “a plan of
action.” This is wrong. An idea is for the sake of planning action, but
an idea is cognition, an awareness of some fact of reality. (Pragmatists
are primacy-of-consciousness philosophers; they award primacy not to
existence, but to some undefined jumble of existence and consciousness,
which they call “experience.”)
To make a plan of action, you must know something. For instance, to plan
a plane trip to Detroit, you must know that there are planes, that Detroit
exists, that there are airports, plane tickets, money to buy them with — and
all the facts that newborn infants don’t know — which is why they cannot
form a “plan of action” regarding plane trips, or anything else.
Awareness of reality — cognition — is what makes possible any plans of
action. We have to know the world in order to act successfully in the world.
It in no way denigrates ideas, in no way reduces them to “expedients,”
to remember that they are for the sake of guiding action. It is not that, as the
Pragmatists say, we have to “play it by ear,” “go with the flow,” and engage in
blind groping. We act on the basis of knowledge. And we do so even when we
act on the basis of probability rather than certainty — knowledge of what is
more probable vs. less probable is still knowledge, a very sophisticated form
of knowledge. Those who lack this knowledge are ill-advised to invest in the
stock market.
Much later in the book, we will see that conceptual knowledge builds in
a hierarchy, ascending from the concrete to the abstract. We will see that,
in contrast to conventional wisdom, the more abstract the knowledge, the more
potent it is. At this point, I will only assert that a very abstract form of know-
ledge — knowledge of principles — is the most powerful of all. Principles are,
of course, exactly what Pragmatism rejects. Pragmatism opposes principles
on principle.
Contrary to the claims of Pragmatism, abstract principles, including
the principles of morality, are man’s indispensable guide to coping with
the demands of life and acting successfully in the world.
 H OW W E K N OW • 1 : F oundations 42

Consciousness as Irreducible
Like existence, consciousness is an irreducible primary. One can subdivide
conscious actions, separating different kinds: seeing, for example, is one kind
of conscious activity and hearing is another. Analogously, one can subdivide
existents — e.g., into living and non-living things. But just as one cannot get
beneath the fundamental fact of existence, so one cannot get beneath the
fundamental fact of consciousness. One cannot reduce conscious action,
qua conscious, to something else.
To ask: “What kind of action is consciousness?” is to ask: “What do all
conscious processes have in common that makes them actions of conscious-
ness rather than physical actions?” The only answer is: all these actions
are actions of consciousness; they all involve awareness of something.
And “awareness” is a synonym for “consciousness.”
This irreducibility of consciousness is one reason why “consciousness”
(like “existence”) qualifies as an axiomatic concept.

An axiomatic concept is the identification of a primary fact of

reality, which cannot be analyzed, i.e., reduced to other facts or
broken into component parts. It is implicit in all facts and in all
knowledge. It is the fundamentally given and directly perceived
or experienced, which requires no proof or explanation, but on
which all proofs and explanations rest.
The first and primary axiomatic concepts are “existence,”
“identity” (which is a corollary of “existence”) and “consciousness.”
One can study what exists and how consciousness functions;
but one cannot analyze (or “prove”) existence as such, or con-
sciousness as such. These are irreducible primaries. [ITOE, 55]

“Irreducible” here means: “cannot be analyzed.” If you try to analyze what

it is to be aware, you will soon discover that no analysis is possible. Aside
from giving synonyms, the only terms that are available for your analysis
are much too general. For instance, you might say that to be conscious of
something is to be “in contact” with it. But chairs are in contact with the floor
without awareness — so, what kind of contact is conscious contact? Or, you
might try saying that awareness is a causal response. But the Earth is causally
responding to the sun. So, what makes something a conscious causal response?
Consciousness. That’s all we can say. There is no further analysis.
 C onsciousness as I rreducible 43

Consider, by contrast, some phenomena that are reducible.

We can reduce a whole entity to its parts — e.g., a box consists of its six sides.
That reduction is possible because a physical whole is the sum of its parts.
Or, we can reduce a physical object to the materials out of which it is
made — e.g., the box is made of wood and nails. Again, the reduction is
possible because the object is the materials out of which it is made.
Consciousness is not a candidate for either of these kinds of reduc-
tion. The state or action of awareness does not have parts. Awareness does
have aspects; e.g., your present state of awareness consists of seeing things,
hearing things, feeling things, etc. But, unlike parts, such aspects cannot
be physically separated, as the box can be taken apart. The seeing, hearing,
etc. can be separated only mentally, by a selective focus or abstraction, from
the organic whole that is consciousness.
The state or action of awareness is not composed of some stuff, as the
box is composed of wood and nails. There are no spiritual components that
get put together to form the field of awareness. Nor can consciousness be
analyzed into physical components. An overall conscious activity — e.g.,
a thought process — can be analyzed into its stages or aspects, but not into
any physical events, not even brain events.
Consider a conscious action in as much detail as you can. Do you find
that it is composed of physical constituents? No. Take your reading of these
words. You first see a word or phrase, then process it and “hear” the words
in your mind, and you understand the meaning of those words. There are,
indeed, non-conscious components, such as eye movements and the physi-
cal page-turnings, but there is no way to reduce the seeing, or the internal
“hearing,” or the understanding to one or more physical sub-actions
— not without leaving out the essential, conscious aspect. Seeing the words
involves physical and physiological processes but is not reducible to just the
physical processes.
There are indeed unconscious sub-processes occurring — notably, the
brain processes that underlie reading. Neurons fire in this brain center,
then in that one. But what do these individual brain processes add up to?
A larger-scale brain process. Not to a state of consciousness, not to the seeing
or the understanding. The sum of small-scale physical processes is merely
a large-scale physical process. The state of consciousness is left out.
This raises the question: If consciousness is an action, what is the entity
that acts? There are two possible answers, but neither one will allow for
a reduction of conscious actions to physical actions, existential or neural.
 H OW W E K N OW • 1 : F oundations 44

One can say that the entity that is conscious is the mental entity, the self.
But, clearly, one cannot reduce the self to little sub-selves. The self, the ego,
the “I,” is an indivisible whole. Awareness is an organic unity; it has aspects,
but no component parts. (In this regard, the relation of the aspects of a state
of awareness to the whole awareness is similar to the relation of the attributes
of an entity to the whole entity; an entity is its attributes and a state of aware-
ness is its aspects.)
On a deeper level, one can say that the entity that is conscious is the man,
as a total organism. And the man has parts, physical parts. But the physical
parts are not the components of his state of awareness. We may assume that
when a man is engaged in conscious activity, the physical organ that acts is
his brain.16 However, the parts of the brain and their individual actions are
not parts of his awareness. If you want to describe the aspects of an action
of consciousness as “parts,” it remains true that the “parts” of an action of
consciousness are mental parts, not physical parts — where “mental” means:
pertaining to consciousness.
Again, parts of a brain process add up to a whole brain process, not to
consciousness.
Brain actions are a necessary condition of consciousness. Brain actions
underlie and are involved in the operation of consciousness. But brain
actions are still something different from what they underlie: awareness.
Peikoff writes:

Even if, someday, consciousness were to be explained scientifi-

cally as a product of physical conditions, this would not alter any
observed fact. It would not alter the fact that, given those condi-
tions, the attributes and functions of consciousness are what they
are. [OPAR, 35]

To discover that matter in certain combinations gives rise to the existence

of consciousness would not permit us to equate consciousness with those
combinations of matter or their physical actions. Mental actions are mani-
festly different from brain actions.
(The nearest analogy to the mind-brain relation in this respect might
be the magnetic field produced by an electric current moving in a wire.

16 The tendency to over-isolate the brain from the total organism is plausibly challenged in
Alva Noë’s Action in Perception [Noë, 2004]; I am not committing myself to the position
that the brain alone, rather than the total organism, is sufficient for consciousness;
I am only saying that even on the view that it is, consciousness remains irreducible.
 C onsciousness as I rreducible 45

The electric current and the resulting magnetic field are two different, though
causally related, phenomena. The current produces the magnetic field, but
is not identical with the field. In a similar way, brain actions may produce
awareness, but they are not identical with awareness. Nothing can be exactly
analogous to consciousness, of course, precisely because consciousness is sui
generis. So, we may dismiss the following objection: “How can consciousness,
if not physical, be an exception to everything else in the universe?” The same
objection would apply to whatever distinctions one made — e.g., “How can
spoons have a distinct nature, since spoons would then be an exception to
everything else in the universe?”)
Consciousness exists and matter exists. Each is what it is, and neither
is a form of the other. To be sure, matter is primary; without matter, there
would be no physical world to be conscious of, no organism to be conscious
of it, and no physiological means of being conscious. Nevertheless, matter
and consciousness are two irreducibly different phenomena. And knowledge,
the specific subject of this book, is a phenomenon of consciousness.
Despite this fact, however, most of those writing on consciousness today
assume that consciousness can be reduced to matter. Materialism, the denial
of the existence or causal efficacy of consciousness, holds a virtual monopoly
in philosophy and allied disciplines. Even those few authors who state that
consciousness is irreducible rarely expunge all their materialist premises.
It is worth looking at examples of how materialist premises can linger
undetected in one’s thinking. One contemporary author claims to reject
materialism but then raises what he calls “the hard question” concerning
consciousness: “the question of how and why cognitive functioning is accom-
panied by conscious experience.” [Chalmers, 25] But what else could cogni-
tive functioning be “accompanied by,” other than cognition, in other words,
consciousness? And why does he say “accompanied” — as if when one learns
something, being conscious were incidental to the process? This represents
implicit materialism left over in that author’s thinking.
Impressed by what computers can do, these implicit materialists ask,
in effect, “Why do man’s actions, such as playing chess, happen to be accom-
panied by conscious experiences? When computers play chess, they do so
by purely physical means, so why is human chess-playing ‘accompanied by’
consciousness?”
But there is a mistake in the question: computers do not play chess; they
do not perform any cognitive task. Computers do not even add. A computer
is a physical mechanism in which the flow of electrical current flips switches.
 H OW W E K N OW • 1 : F oundations 46

A computer is essentially a switching device: it turns switches on or off, and

it charges or discharges capacitors.
Strictly speaking, computers do not play games, check spelling, or process
information. All these actions involve consciousness and are performed only
by conscious human beings, even if aided in the task by physical devices such
as a pen, a typewriter, or a computer.
As to the simplest case — addition — one need only realize that mathemat-
ical truths, such as that one plus one equals two, do not exist for computers
and have no effect on them. Nothing is changed if the addition is formulated
in binary: 01 + 01 = 10. This mathematical principle does not exist for the
computer: there are no ones or zeros inside the machine. There are not even
any bits or bytes inside the machine, only high and low voltages.
Men can use a computer to help them add by using the physical state
of the switches to control screen pixels to form patterns that, to us, represent
numbers. Computers themselves no more add than do the old-fashioned,
mechanical adding machines. The adding machines merely turn gears to
rotate wheels with numbers painted on them. Gears and wheels do not add.
Nor do computers. Addition is an action of consciousness.
You can count on your fingers, but fingers cannot count.
Computers cannot “process information,” because information is not
a physical phenomenon. Computers can only combine and shunt electrical
currents. Only electricity, not information, has causal impact on the work-
ings of the computer; information does not exist for the computer.
Of course, there is nothing wrong with saying colloquially that computers
add, process information, and play chess. But in philosophy we have to be
exact: in the strict sense computers only combine currents, throw switches,
charge and discharge capacitors. Computers don’t follow programs, they
simply obey the laws of physics. That’s all that goes on inside them.
If all human beings suddenly vanished from the face of the earth,
but their computers remained running, there would be no information
processing: the computers would merely be combining electric currents
and lighting different pixels on their screens, not processing information or
performing calculations.
Some materialists, relishing the man-as-computer model, have advanced
the slogan: “The brain is the hardware, the mind is the software.” Here, “soft-
ware” is a stolen concept: something can be identified as software only in
relation to the mind. Software, information, symbols, mathematics — none
of these things exist per se in the physical world, apart from a relation to
 C onsciousness as I rreducible 47

consciousness. Just as books contain only patterns of ink, so apart from

man’s mind, software exists only in the form of some physical patterns, such
as the patterns of magnetized iron particles on a hard drive’s disk. Patterns
of ink qualify as words and patterns of iron particles qualify as programs only
in relation to man’s mind.
None of this is to denigrate computers, which are magnificent inven-
tions — inventions of conscious human beings. When the IBM computer
“Deep Blue” beat the world’s top chess player, Garry Kasparov, that was not
a defeat of the human mind but its triumph: it was a victory for the minds
of the men who created and programmed the Deep Blue computer.
There is a similarity between a computer and the brain. The brain, like the
computer, combines and switches electric pulses. But brain processes can
correctly be described as dealing with ideas or information only in relation
to the mind of the person whose brain it is. Apart from the mind that oper-
ates that brain, there are only physical states and physical changes.
Thus, there is simply no such puzzle as: “How is it that man’s calculations
are conscious, but the computer’s calculations aren’t?” One might as well ask:
“Since my television can talk without being conscious, why do men happen to
be conscious when they talk to each other?” Televisions cannot literally talk
and computers cannot literally make computations.
Why do I say that this confusion represents implicit materialism?
Shouldn’t one say, instead, that this school of thought is actually too pro-
consciousness, wrongly importing consciousness into what is purely
physical? However, the belief that is actually operative is not that comput-
ers are conscious, but that human addition, chess-playing, etc. are purely
physical processes. It is not that the materialists believe that computers have
minds (despite the fact that they usually like to state their view in just that way);
rather it is that they eliminate mind altogether. Their inability to see the
difference between man and machine is not an elevation of the machine
but a degradation of man.17
The better of these authors have come far enough from materialism to
accept the existence of consciousness. They recognize that we see, think,
and feel. But, clinging to the materialist definition of cognitive activities,
they see consciousness as “epiphenomenal,” as the proverbial ghost in the
machine, and can find no causal role for it. Most of them even hold that
everything that human beings do could be done by unconscious “zombies,”

17 For an excellent source on this degradation, see Tallis, 2011.

 H OW W E K N OW • 1 : F oundations 48

their science-fiction projection of beings that resemble us exactly in every

physical and behavioral way, but are non-conscious.
Whether the materialism is explicit or implicit, it steals concepts in
profusion — not blushing at statements such as, “What people believe
to be mental is really non-mental” — ignoring that “believe” is a concept
of consciousness. They make pronouncements like, “Consciousness is
a myth”; “Consciousness is an illusion”; “The concept of ‘consciousness’
is invalid” — ignoring that “myth,” “illusion,” “concept,” and “invalid” are
concepts of consciousness.
One philosopher even titled an essay: “A Reason to Doubt the Existence of
Consciousness.” In that title, “reason” and “doubt” leap out as stolen concepts.
The very definition of “materialism” exhibits the stolen concept fallacy.
“Materialism” is the theory that consciousness does not exist. But “theory”
is itself a concept of consciousness, so the definition relies on that which
it claims does not exist.
Materialism is the idea that there are no ideas.
Consciousness is an axiom, and there is no denying axioms: they have
to be used and accepted in any attempt to deny them. “I deny conscious-
ness”? — denial is an action of consciousness. Stones, plants, and computers
cannot engage in denials.
Ironically, materialists — who see themselves as scientific, hard-headed
realists — take their basic premise from the spiritualists. It is the spiritu-
alists’ notion of consciousness as supernatural, not of this earth, “spooky,”
that the materialists adopt as their concept of consciousness. On that basis,
they decide, logically enough, that such a phenomenon does not exist.
Materialists are blind to any view of consciousness other than that of the
spiritualists. But this book asserts a third alternative: a naturalistic, causal,
biological conception of consciousness.
When materialists (properly) reject the spiritualist, mystical conception of
consciousness as “a fragment torn from God,” they assume they have rejected
consciousness. But they have rejected only a straw man.
Supernaturalism must not dictate the terms of this, or any other, question.
The rational, scientific, realistic question is not: “Is there a ghostly tendril
from another world in me?” but rather: “Am I aware of something? Do I see,
hear, think, remember, feel pleasure and pain?”
But then the question answers itself.
 T he C ausal E fficac y of C onsciousness 49

The Causal Efficacy of Consciousness

Biologically, consciousness is not a passive spectator; an organism’s
consciousness controls the actions of its body. It is the efficacy of conscious-
ness in guiding such actions that explains the selection-pressure that favored
its evolutionary development.
The contrary position, the claim that consciousness has no causal efficacy,
holds that consciousness is an “epiphenomenon”: an effect produced by the
brain that does not itself have any effects, does not do anything. The standard
analogy given by epiphenomenalists is that awareness is like the smoke that
comes out of a locomotive: the smoke is an effect, a by-product, which does
not act on the locomotive. In the same way, the actions of your conscious-
ness — your pleasures, pains, fears, hopes, thoughts, plans — are supposed
to be mere by-products of the brain’s action, by-products that are causal
dead ends, impotent to affect anything.
Epiphenomenalists are so opposed to consciousness that they exempt
it from the causality that governs the rest of the universe: consciousness
allegedly has causes but no effects. Note that their own illustrative analogy
fails: the smoke leaving a locomotive does have an effect, however small,
on the locomotive — as it does on the lungs of anyone who breathes it in.
Nothing in the universe acts without having effects, and only a mystical view
of consciousness would allow for making it an exception.
Epiphenomenalism is another case of a self-refuting idea. Anyone who
asserts epiphenomenalism is contradicting himself: in asserting his view,
he assumes that his thoughts are the cause of the sounds coming out of
his mouth. Even asserting epiphenomenalism internally, as an unspoken
thought, contains a contradiction: to think requires memory, and memory
requires retrieving what consciousness has stored physically in the brain.
The conscious storage and retrieval of memories evidences the mind’s ability
to interact with the physical brain.
Epiphenomenalism is actually a form of materialism. Whether one main-
tains that consciousness does not exist or that it does exist but cannot affect
anything, including oneself, the significance is the same.
From a biological perspective, epiphenomenalism represents a denial of
the adaptive value of consciousness. If consciousness has no bodily effects,
then it confers no survival advantage for organisms possessing it. But if
so, how are we to account for the observed facts? How are we to explain
the evolutionary fine-tuning of conscious experiences to fit survival needs?
An epiphenomenalist must assume that it is just a cosmic coincidence that
 H OW W E K N OW • 1 : F oundations 50

the conditions that fulfill bodily needs — eating nutritious food, gaining
shelter and warmth, drinking when dehydrated — happen to produce pleasure,
while damaging physical conditions — a wound, starvation, breaking
a limb — happen to bring pain. If the conscious experiences of pleasure and
pain have no motivational power for the conscious animals, if the actions of
the conscious animals are not affected by their experience of pleasure or pain,
why are pleasure and pain correlated in this fashion with survival needs?
Clearly, there has been a selection-pressure acting in evolution to align
pleasure and pain with actions that promote or impair survival, respectively.
But that selection can occur only if pleasure and pain have effects on the
animal’s (or man’s) behavior. Were pleasure and pain epiphenomena, there
could be no selection operating to prevent animals and men from being
so constituted as to feel excruciating pain when eating and ecstasy upon
breaking a limb.
(People unschooled in evolutionary biology sometimes point to the rare
cases in which pleasure and pain do not correlate with survival. E.g., some
people like the taste of alcohol, which is unhealthy if over-used. But alcohol
is not part of the environment to which man became adapted. And finding
a few exceptions would not help the epiphenomenalist explain the general
correlation of these allegedly impotent conscious experiences with survival.)
This evolutionary explanation of pleasure and pain in terms of their effects
on survival is merely an illustration from science of what we know by direct
introspection: consciousness does something; it has causal efficacy.
Since your consciousness causes your voluntary action, and since the
physiological cause of your action is a process in the brain, it follows that
your consciousness has the power to change the physical state of your brain.
This conclusion may appear to contradict the primacy of existence, since
it means that consciousness alters the state of something in the physical
world — the brain. But that worry is unfounded. The primacy of existence
holds that a state of awareness neither creates nor alters its object. A pencil,
for instance, is not affected by being seen or thought about. But the causal
efficacy of consciousness does not imply otherwise.
When you raise your arm, you don’t do it by somehow making your brain
into the object of your awareness. When you reach for a pencil, the object
of your awareness is the pencil, not your brain. And, of course, seeing and
desiring the pencil does not alter it in the least. Awareness affects not the
object but the subject — i.e., you.
 T he C ausal E fficac y of C onsciousness 51

Your inner decision to reach for the pencil is an act of your consciousness,
based on your awareness of the pencil and your desire to use it. But there
is no effect of your consciousness itself on the pencil. The instru­mental value
of the pencil to your purpose is also an objective fact, one not alterable by
consciousness. There is never a case in which awareness alters its object.
What about the effects of self-observation? When one focuses inwardly
on a feeling, doesn’t that alter the feeling in some cases? No, not in any pri-
macy-of-consciousness way; the inward focus changes what one is attending
to; one’s consciousness is then responding to a somewhat different object
or changed context. This is not a case of consciousness molding the feeling:
the feeling is a process, not an entity. Thus, when one speaks, informally, of
“changing” a feeling, the actual meaning is that one process has ceased and
a different one has begun. Even a seemingly static feeling is a dynamic state,
like a waterfall, fed by ever new “waters” flowing in “the stream of conscious-
ness.” The fact that selecting a different object of attention produces different
feelings in response is an expression of the primacy of existence: the object
attended to causes the response in consciousness.18
The same is true of the effect of introspection on a thought process:
one’s critical scrutiny of a thought process will certainly influence the
subsequent train of thought. But that is not a case of awareness altering the
introspected thought, as if the thought were a static entity: it is a case of
awareness altering the person, the subject, in a way that results in his having
new and different thoughts.
The emotions one has and the thoughts that occur to one are generated
by the brain (i.e., the subconscious). One’s brain is what it is, independent
of one’s awareness of it. The brain is certainly not a creation of consciousness.
The state of your brain can be altered, within definite limits, by your aware-
ness of something in the world. The current state of your brain can even
be altered by recognizing something about your consciousness, as when
self-understanding causes a gradual change in your psychology. But it is not
the case that wishing one had a different type of brain makes one’s brain
change, or that one can escape the effects on one’s brain of what one does
with one’s consciousness. Yes, one’s mind affects one’s brain — in the only way
that the nature of one’s brain permits, given its independent identity.

18 If emotions could be changed by a wish or merely by looking at them, the profession

of psychotherapy would be unnecessary. In reality, to cease feeling a certain way about
a given kind of object, one must change the feeling’s cause: the operative beliefs and
automatized value-judgments. Sometimes simply learning the true nature of the object
is enough to change the emotion (“Oh, that’s only a toy gun”); other times one has
to work over a long period of time to come to grips with and change an automatized
evaluation, as is done in the better systems of psychotherapy.
 H OW W E K N OW • 1 : F oundations 52

For instance, if one wants to return a tennis serve, one has to put forth
the mental effort required to get one’s body in motion in the proper way.
Even here, mere wishing doesn’t make it so. Wanting to perform the proper
actions to return the serve doesn’t mean one will perform them: one has to
focus one’s attention on the right factors and give the right “orders” to oneself.
Nor is ignorance bliss, as the primacy of consciousness would have
it. Ignorance of the consequences on the brain of making a certain choice
does not affect those consequences. A policy of giving in to one’s whims,
for example, has long-term, inescapable consequences on one’s motivational
system — notably, in producing what we call a weak-willed psychology.
Or, on the positive side, a policy of seeking mental clarity, defining
one’s terms, and adhering strictly to logic develops a brain that is efficient,
well-organized, and disciplined.
The point here is that the causal efficacy of a consciousness in regard
to its own brain does not erase the brain’s independent identity nor that of
the object of which one is aware. Thus, the mind’s causal efficacy is actually
an instance of, not an exception to, the primacy of existence.
There is no alternative to accepting the fact that consciousness can causally
affect the physical state of brain processes. For how else are we to describe
the fact that it is my current conscious thoughts and perceptions that are
causing the movements of my fingers on the computer keyboard? (See Schwartz
and Begley, 2003.)
In confronting the causal efficacy of consciousness, materialists raise
several artificial objections and threaten us with straw men. They claim
that there is some problem with a causal interaction between the mental
and the physical, because the two are radically different types of phenomena.
But there is no principle requiring interacting existents to be similar.
This is merely picture-thinking — which means non-thinking. The image
evoked is that of two objects interacting; to interact, they must have a surface
at which they can meet; if they don’t contact each other, if they “miss” each
other, there is no collision, no interface, no interaction. But in fact there is
no requirement that interacting things be similar. “Similarity” is not a place
where two existents meet, and “difference” is not a failure to make contact.
In actual fact, different things — even radically different things — interact.
Rivers interact with rocks, plants interact with sunlight, man interacts with
rainbows, and biotech corporations interact with DNA molecules.
All that is required for A to interact causally with B is that both A and
B be part of the universe — i.e., that both exist, have a definite nature, and act
 T he C ausal E fficac y of C onsciousness 53

accordingly. Both my thoughts and my body exist and have a definite nature;
action follows accordingly. The fact that mind and body are different types of
phenomena provides no grounds for making a puzzle out of their interaction.
As to the materialist’s straw man, he is Descartes. “Cartesian dualism”
is the accusation hurled at anyone who recognizes that consciousness
is real and has real effects. But the error in Cartesian dualism lies in its
reification of consciousness — in making it into a substance, res cogitans,
rather than in merely acknowledging the existence of consciousness and
its causal efficacy.
Consciousness is not an entity, not in the sense that a stone or an organism is.
Consciousness is a faculty of an entity, a man or animal; the operation of
consciousness is a process of that entity.
Actually, Descartes’ mistake here lies, ironically, in his being too affected
by materialism: he attempts to apply to consciousness a category derived
from the physical world: the category of substance or entity. In dividing
reality into res cogitans and res extensa, he forgets that res means “thing,”
a concept that derives from perceiving solid, physical objects. Nothing but
confusion and contradictions can result from likening consciousness to
physical objects.
The materialists are right to reject Descartes’ reification of conscious-
ness, and they are right to reject the related notion of a free-floating “soul.”
But they are wrong to reject the self-evident fact that man is conscious.
Contra materialism, consciousness is undeniable. Contra spiritualism,
consciousness depends on the physical body or nervous system. (Or, more
accurately, consciousness depends on the organism — the person or animal,
as a whole.) Such organisms have two aspects, a body and a ­consciousness,
a physical aspect and a mental one. There is a sense in which one’s
consciousness depends on his physical body, but there is also a sense in
which one’s body depends on his consciousness: a completely unconscious
body, a body in which the faculty has been destroyed or never developed,
is not a functioning body in the sense applicable to that kind of organism.
Aristotle pointed out that it is an equivocation to call a severed human
hand a “hand,” because its essential function is absent. In the same way,
a corpse is not a human body; it is the remains of what was a human body
when there was a person. A person’s body is an abstraction from what is given:
the person. Likewise, a person’s consciousness is an aspect of the whole
person, abstracted out for separate consideration. The entity is the person,
an integrated whole.
 H OW W E K N OW • 1 : F oundations 54

The Validity of Introspection

Implicit in being able to discuss all the preceding is the fact that we, as human
beings, are aware of the actions and products of consciousness — i.e., we are
each self-aware and can introspectively identify our own conscious processes.

Introspection is a process of cognition directed inward —

a process of apprehending one’s own psychological actions in
regard to some existent(s) of the external world, such actions
as thinking, feeling, reminiscing, etc. [ITOE, 29]

The points I have made about consciousness are based on introspection.

It is by reflecting on, conceptualizing, and identifying one’s own mental
actions that one comes to understand the general nature of consciousness.
The denial that we are self-aware is self-refuting: the denial maintains, in
one form or another, that concepts of consciousness are invalid, that they do
not represent an awareness of anything real. Since this claim presupposes
knowledge of the distinction between contents of consciousness and exter-
nal facts, it is just another form of denying the axiom of consciousness and
commits the same stolen concept fallacy: the concept “invalid” pre­supposes,
implicitly or explicitly, the concept “consciousness.” “Invalid” refers to the
status of an idea, a status that is differentiated from that of a valid aware-
ness of fact. But making that distinction presupposes that we have recog-
nized what it is to be aware of facts, which implies an act of self-awareness.
The denial of introspective self-awareness presupposes what is being denied.
I have covered several undeniable facts about the nature of consciousness
as such: consciousness has an object and a subject; existence has primacy
over consciousness; consciousness is an active process, enabled by specific
means; consciousness is a biological faculty; and, the state of awareness
is both irreducible and causally efficacious. (There are also self-evident facts,
covered in later chapters, that pertain specifically to man’s consciousness,
which is conceptual.)
The one-sentence, highly condensed overview is: consciousness is a living
organism’s active process of perceiving reality to acquire the information
required for its survival.
We must reject the all-too-common practice of making arbitrary assump-
tions about what consciousness “ought to be” or how it “ought to function,”
and instead acknowledge these undeniable facts about what consciousness
 T he Validit y of I ntrospection 55

actually is. On that basis, we can go on to develop a proper understanding of

the crucial product of consciousness: knowledge.
The beginning of knowledge, and the ultimate source of all knowledge, is
sensory perception, which is the subject of the next chapter.
H OW W E K N OW 56
 PERCEPTION
2
T he biological function of consciousness is to guide action,
and the basic source of guidance is cognition. A cognitive process is one
devoted to gaining information about reality. Cognitive activities range
from an animal’s perception of the entities in its immediate environment
to man’s complex processes of scientific investigation. However primitive or
advanced, the cognitive functions of consciousness are directed toward pro-
viding awareness of what things are, of their identities.
(Some phenomena of consciousness, such as emotion and imagination,
are not cognitive. E.g., to feel fear is to have an experience,19 not to acquire
information. Fear is a reaction to content acquired by other means. Cognitive
acts — acts of awareness — are the faculty’s base, making possible the rest.)
Sensory perception is an animal’s or man’s primary form of cognitive
contact with the world. Knowledge begins with, develops out of, and is tested
against sensory observation. This point is not self-evident, nor is it the view
of cognition with which mankind began. Perception’s fundamentality was

19 To Gregory Salmieri I owe the idea of using the term “experiences” to cover both
cognitive and noncognitive mental states — e.g., seeing, fearing, and dreaming are all
experi­ences, but fearing and dreaming are not acts of awareness. In the same vein, it is gener­
ally better to use the term “content” instead of “object” in discussing certain noncognitive
states. Dreams have contents, but it is misleading to say they have objects (see pp. 282–83).
 H OW W E K N OW • 2 : P erception 58

first identified by Aristotle, but that identification did not become widely
accepted until almost 1500 years later, after the long night of the anti-senses
Dark and Medieval ages. Even at the dawn of the scientific era, perceptual
observation was attacked and derided. How could men like Copernicus
and Galileo cast aside the revealed word of God? How could they trust
“observations” that were the product of debased bodily senses, or imagine
that their limited, finite intellects, without aid from God, could produce
anything other than confused, conflicting opinions?
Over a span of centuries, through the writings of Thomas Aquinas
(c. 1250), Francis Bacon (1620), and John Locke (1690), the Aristotelian view
won out, and mankind entered the Enlightenment era, the Age of Reason.
But a counter-attack was soon launched by — of all people — philosophers.
Starting with Descartes and bottoming out with Kant, a prominent line
of philosophers peddled a secularized version of the old religious notions.
“I have therefore found it necessary to deny knowledge in order to make room
for faith,” Kant wrote. [Critique, B, xxxi, Kant’s emphasis]
For the open mysticism of the medievals, these philosophers substituted
Rationalism — the idea that the intellect can spin out truths on its own,
without needing sensory data. For the authority of sacred texts, they sub-
stituted the equally baseless notion of innate ideas or innate “categories.”
Instead of attacking the senses as “of the flesh,” they attacked the senses
on other grounds, to be discussed below.
Fully liberating the intellect requires rejecting both open mysticism
and the secularized form of it, which is Rationalism. One must uphold
the efficacy of the unaided individual mind. This means defending both
the senses and reason. The remainder of this book is devoted to doing just
that. I establish two fundamental points: 1) perception is the base of all
knowledge; 2) valid concepts are formed from perception by an objective
process. The present chapter presents a thoroughly naturalistic, biological
view of sensory perception; the remaining chapters present the equivalent
for conceptual activities.

Perception as Axiomatic
Sensory perception is the primary and basic form of cognitive contact
with the world. An organism born entirely without sense organs would be
unconscious. Accordingly, the fact that the senses provide awareness of
 P erception vs . Sensation 59

reality is axiomatic. The issue of the “validity” of the senses does not even
arise: sensory awareness is awareness — which means that it has the status of
a corollary of the axiom of consciousness.
The axiomatic nature of sensory awareness is confirmed by the argument
of re-affirmation through denial, the test of axiomaticity. To make any state-
ment denying the senses, one has to understand the terms the statement
uses — “senses,” “invalid,” etc. But the meaning of these terms is learned,
directly or indirectly, on the basis of perception. Without the senses’ basic
cognitive contact with reality, we could not have any concepts, including
those used to claim that the senses are invalid. Thus, the attack on the senses
constitutes concept-stealing on an unparalleled scale. Without perception,
we would be unconscious, like vegetables; vegetables cannot ponder the
validity of perception.
Because sensory awareness is axiomatic, philosophy, as distinguished from
science, has very little of a positive nature to say about it. Much, however,
has to be said to correct misunderstandings created by wrong philosophic
theories of sense-perception, theories that have led philosophers down
innumerable blind alleys.
The major source of error in this regard comes from confusing perception
with lower or higher levels of awareness — i.e., confusing perception with
sensation or with conceptual cognition.

Perception vs. Sensation

Surveying the range of animal life on the planet, one cannot say with any
confidence where on the evolutionary scale consciousness first appears.
We know that we ourselves are conscious, and it would be bizarre to question
the existence of consciousness in the higher animals, such as dogs and cats.
But what about jellyfish, which have a “neural net,” or flatworms, which
have a primitive brain?20 Perhaps neuroscience will someday provide a better
understanding of the physical factors that give rise to consciousness, and
that understanding will settle the question of which of the lower organisms
are conscious and which are not. But for now such questions remain open
and are for science, not philosophy, to investigate.

20 On what distinguishes a brain from a less developed group of neurons, see Sarnat, 2002.
In another article, Sarnat notes that the flatworm, planaria, is “the simplest living animal
having a body plan of bilateral symmetry and cephalization.” [Sarnat, 1985]
 H OW W E K N OW • 2 : P erception 60

As intriguing as these scientific questions are, they have zero import for
and impact upon philosophy, which is concerned with man’s consciousness.
Non-human consciousness has philosophic significance only insofar as it
illuminates, by contrast, the nature of man’s consciousness.
(The scientist does, however, need the right philosophic base from
which to proceed in studying sensory awareness and for investigating the
mind-brain relationship. Scientists work from a philosophic base, and
much of the research that scientists have done on sense-perception has
been distorted by wrong philosophic premises. For example, the premise of
materialism has led researchers to attempt to reduce perception to brain
events and overt behavior, as if awareness did not exist. Yet these scientists
claim to be aware of the people they study, the data they collect, and the
content of their own theories.)
Among animals with sensory awareness, the simplest possess only the
faculty of sensation.
A “sensation,” as I use that term, is the most primitive form of conscious
response, the response to energy impinging on receptors, not to objects in
a perceived world.
The crayfish, for example, has light-sensitive cells near the end of its tail.
Crayfish need to hide themselves from predators by moving into crevices
or under rocks. By detecting light hitting the end of its tail, the animal can
ensure that not just its head but its whole body is hidden: when its head is
in darkness but its tail is still receiving light, the crayfish will crawl forward.
The crayfish does not see any objects with its tail receptors — the sensory
equipment is too primitive for that — but it responds to light vs. darkness,
and if that response is a conscious one, it is as sensations not percepts that it
experiences the illumination level.
A sensation is a conscious response to stimulation at the receptors, and
that response lasts only as long as the stimulus is applied. A sensation is
thus stimulus-bound: it is a sense or feeling, in response to what is currently
stimulating the receptors.
The higher animals have evolved a much more potent form of awareness:
perception. There are a number of features that distinguish perception from
mere sensations.
1. Perception is awareness of entities — of things (including their
­characteristics). Whereas the crayfish’s tail-spot only discriminates brightness
from darkness, human vision provides man with awareness not of stimuli but
of the objects in the world, the objects that are responsible for the patterns
 P erception vs . Sensation 61

in the light received by the eye. We see trees, dogs, books, clouds — rather
than just discriminating a general level of illumination. Human eyes, like
the crayfish’s tail spot, respond to light, but the human visual s­ystem
is able to detect and exploit patterns in the light. The nature of these
patterns is determined by the layout of the objects that reflected the
light. D­ etecting these patterns enables the visual system to discriminate
entities from each other. Thus, the content of visual perception is a world
of entities. Vision contrasts not light with darkness but a lighter and/or
differently colored thing against the other things in its background.
The same is true of hearing and touch. We hear the actions of things.
Despite some marginal cases, as when one is aware of a background hum
whose location and source are not apprehended, the normal case is hearing
things that make sounds, not just the sounds: a slamming door, a barking
dog, the click of keys on the computer keyboard.
Touch also discriminates entities, unless the conditions of perception are
impoverished. We feel the table, the spoon in our hands, the keyboard under
our fingers — all of which is quite different from simply feeling pressure on
our skin. Even with eyes closed, we can explore by touch the objects within
reach. Touch perception, as opposed to mere feelings on the skin, is an active,
exploratory process, one that presents us with entities.
Taste and smell are more primitive, closer to the level of sensations.
But they occur in a perceptual context: when we bite into a peach and taste it,
we are already aware by sight and touch of the peach as an entity in the world
and of the peach morsel as an entity in our mouth. Similarly, when we smell
something, it is part of a perception of things emitting odors in a perceptual-
level world. (Animals with a keen sense of smell, such as dogs, seem to have
not just smell sensations but some form of perceptual awareness of a scent
trail as an entity.) What perception provides is awareness of entities.
2. A point essential to understanding perception is that perception is spatial;
it presents a world of entities arrayed in space — i.e., in their relative positions.
We do not perceive one isolated entity at a time, but a spread-out world
of entities, each entity being discriminated from the others that are next to it.
Philosophers are apt to take as their example the perception of a single object:
we see “an apple,” for example. But apples do not float in a void; perception is
of a world of spatially arrayed entities (see the two photographs on the next
page, but bear in mind that perception is an ongoing process, not a snapshot)
The three-dimensional spatial array given in perception is what fundamen-
tally distinguishes perception from sensation. It is not merely that perception
 H OW W E K N OW • 2 : P erception 62

(especially vision) gives entities, but also that perception provides the
co-presence of all the entities that the animal can act on or be affected by.
We see in one spread the entire scene of entities.
Contrast discriminating spatially arrayed entities with discriminating the
taste of one flavor element, say cin-
namon, from others in what one is
tasting. Such discrimination does
not rise to the level of perceiving the
cinnamon, precisely because the
cinnamon is not given as spatially
discriminated from the other flavors
that one is also tasting. The percep-
tual world is spatially arrayed.
The space given in perception is
not the abstract space of the geom-
eter, with its three Cartesian axes,
but the relative position of entities.
As psychologist J. J. Gibson stresses, “visual space, unlike abstract geometri-
cal space, is perceived only by virtue of what fills it.” [Gibson 1950, 5]
 P erception vs . Sensation 63

3. Perception gives us awareness of a world of entities extending out in

all directions from “here,” i.e., from oneself. As one moves around in the
perceived world, one’s vantage point, and hence “here,” moves, giving one
a sense of one’s current place in the world. Thus, perception includes at least
some sense of self. And, of course, one perceives one’s own limbs and trunk,
and their spatial relation to the other things in the surroundings.
Additionally, a form of self-awareness is implicit in how the movement of
one’s head affects the scene’s limits at any given moment. Gibson writes:

The head and body of the observer hide the surfaces of the world
that are outside the occluding edges of the field of view. . . . An
observer perceives the position of here relative to the environ-
ment and also his body as being here. His limbs protrude into the
field of view, and even his nose is a sort of protuberance into the
field. [Gibson 1986, 206, 208]

4. Perception is not a momentary, static impression but a continuous pro-

cess over time. In the process of perceiving the world, the animal or man is
an active, exploring observer. He scans his environment, moves through it,
acts on it, and perceives the changes in the world that result.
Psychologists have learned that an animal needs self-produced motion in
order to develop perception. In a classic experiment, Richard Held and Alan
Hein reared kittens under conditions in which their only visual experience
of their environment occurred while they were being passively transported
through it. They did not develop normal perception; kittens raised in the
same conditions but who moved themselves through the same environment
did develop normal perception.21
Perception is both for action and by action: it is an active process, an
ongoing awareness of the entities in the world, achieved by self-moving
animals exploring their environment.
To summarize in a preliminary definition: “Perception” is the ongoing
awareness of entities in their relative positions, gained from actively acquired
sensory inputs.

21 “[V]isual stimulation of the active member (A) of each of 10 pairs of neonatal kittens
was allowed to vary with its locomotory movements while equivalent stimulation of the
second member (P) resulted from passive motion. Subsequent tests of visually guided
paw placement, discrimination on a visual cliff, and the blink response were normal for
A but failing in P.” [Held and Hein, 1963]
 H OW W E K N OW • 2 : P erception 64

Sensationalism
The preceding understanding of perception is radically at odds with most
traditional views. According to most philosophers and psychologists,
perceptions are constructed out of “sensations.” This approach, known as
“sensationalism,” holds that when looking at an apple, we have now, or had
in infancy, separate sensations of color, brightness, roundness, etc; the mind
or brain supposedly puts together those separate sensations into the sight
of the apple.
This notion is completely mistaken. Perception is a unitary phenomenon;
it does not have sensations or anything else as components. It is not the case
that sensations are cognitive “atoms” out of which perception is built up,
whether by the brain or the intellect.22
Do not confuse sensations, which would be states of consciousness, with
physical events at the sensory receptors. E.g., light-absorption by the rods
and cones of the retina is a physical event, not a sensation. Though the
incoming light interacts physically with a whole array of rods and cones
in the retina, the resulting visual perception is continuous, global, unified.
Whatever it is that the brain does in combining, decomposing, or shuffling
the electrochemical outputs from the rods and cones, what results is your
seeing a whole scene. The fact that the physical process begins with light
hitting an array of discrete receptors does not mean that the ensuing state
of awareness has parts. Note the parallel in the case of touch: when you grasp
an object with several fingers, your touch-awareness, prior to any analysis,
is unified, not broken into separate responses from each finger, let alone from
each microscopic receptor in the skin.
A physical object is an organization of atoms. A table is a combination
of atoms. And a certain combination of atoms is a table. But consciousness
is not matter. There are no consciousness “atoms,” no particles of awareness.

22 This marks one of the very rare occasions on which I differ with Ayn Rand; she wrote:
“A percept is a group of sensations automatically retained and integrated by the brain of
a living organism.” Possibly she meant a group not of sensations but of sensory inputs,
since it is hard to see how bits of awareness (sensations) could be integrated by the
brain, which is physical. In the same passage, she speaks of “sensations as components of
percepts.” As explained in the text, I definitely disagree. However, this is not a philosophic
issue, as she notes in the same article: “The knowledge of sensations as components of
percepts is not direct, it is acquired by man much later: it is a scientific, conceptual discov-
ery.” [ITOE, 5] Gibsonians also, with some justification, oppose using the term “percept,”
which connotes snapshots rather than an ongoing, continuous process. (In a few places I do
use the term “percept,” but simply for linguistic grace to go with “concept.”) I am indebted
to Lee Pierson for both of these points and for encouraging me to study Gibson’s works.
 Sensationalism 65

Consequently, it is a mistake to treat sensations as the “atoms” of awareness.

More generally, it is wrong to treat consciousness on the model of matter by
looking for any such “atoms” of awareness.
Perception is not composed of anything; perception is a global, seamless
phenomenon. Even sensory modalities, though distinguishable, are not given
as parts of awareness — not now, and not in earliest infancy. The newborn
infant simply is aware — awake, conscious. Over time, it learns to differentiate
each sense modality from the others — hearing from touch from smell, etc.
But its awareness did not begin with separate or discriminated modalities that
it then had to assemble. Its awareness began as, and remains, a single whole.
Consciousness is not put together from separate experiences. Rather,
what we call “an experience” or (colloquially) “a sensation” is the product of
analysis, something isolated out from one, unified “stream of consciousness.”
Like “sensation,” the term “stimulus” should not be construed on the
atomic model, as if each stimulus caused its own particle of awareness.23
A stimulus does not create an awareness where there was no awareness
before. Rather, when there is a particular change in the ongoing sea of stimu-
lation that the organism is immersed in, the organism becomes aware of that
change, in addition to the things it was previously aware of. A particular
stimulus causes a modification in the field of one’s awareness, not the genesis
of an “atom” of awareness; there are no such “atoms.”
The sensationalist model is one of isolated zings of energy at the receptors
resulting in corresponding pings of sensation in consciousness. But the zing-
ping model contradicts what we plainly experience in perception: a unified,
dynamic, seamless field of awareness.
The zing-ping model takes quiescent non-awareness as the norm, the
default state — as if awareness were a disturbance interrupting nirvana.
In fact, however, our perceptual systems operate in an ever-present, churning
sea of stimulation — variations in light, sound, pressure, force, torque, etc.
What we isolate as “a stimulus” is simply one wavelet singled out mentally
from that choppy, roiling sea.
From the total perceptual experience, it is possible to abstract out a single
dimension of response, such as brightness or softness. The products of such
abstraction are often, but confusingly, referred to as “sensations.” The better
term is “sensory qualities.” In some cases, when authors refer to “sensations,”
they mean these sensory qualities. The danger in using the term “sensations”
is that it suggests that the infant starts life with an experience of disembodied

23 Nor is “stimulus” a good term to use in describing perception, since it suggests passivity
on the part of the perceiving organism.
 H OW W E K N OW • 2 : P erception 66

qualities, which he (or his nervous system) then has to join together to reach
awareness of entities. This is a pervasive and seductive error. Brightness and
the like are abstractions, not the “original stuff ” of awareness, not material
that gets assembled into a constructed whole. In fact, perception of entities is
the given; individual qualities, such as brightness, are what we later analyze
out as a single dimension or variable.
By analogy: we can analyze a force into its vector “components,” but
these are not actual constituents of the force. When a wind is blowing to the
northeast, physics finds it helpful in calculations to analyze that force into
two “components,” a north component and an east component. But in reality,
the northeast wind has no components — it is not a combination of a north-
ward wind with an eastward one. In the same way, your perception of
an apple is not a combination of sensations of color, brightness, texture, etc.
The error in sensationalism is reification: the fallacy of taking an aspect
of a thing, grasped by mental analysis, as if it were an entity capable of
separate existence. The simplest example would be thinking that a coin
is a combination of heads and tails, rather than realizing that heads and tails
are not entities put together to form a coin, but are aspects of the actual
entity, the coin — aspects that we mentally isolate. Likewise, sensory qualities,
like brightness and softness, are aspects of the perceptual whole, which we
mentally isolate but which never existed as separate phenomena.
Analysis is indispensable to cognition, but analysis consists of breaking
down what is given as a whole in perception. Thus, it is a complete inversion
to treat the perceptually given as being a compound of sensory qualities or
sensations — which are, in fact, aspects separated out by analysis.
Take another example: color perception. Although color seems at first
to be a single quality, analysis identifies three factors: hue, saturation, and
brightness. But we do not begin by having separate sensations of each
factor. Nor does our experience of color require that we put together any
such hue-sensations, saturation-sensations, and brightness-sensations.
Only a minuscule percentage of the human race even knows about these
three aspects of color.
Brightness, loudness, pressure, warmth, and the like are the simplest,
irreducible dimensions of a global, multi-modality field of perceptual awareness,
but these dimensions are not elements, components, or “atoms” of aware-
ness put together by either the brain or the mind into a perceptual whole.
Sensationalism must be rejected.
 Sensationalism 67

The term “sensation,” apart from its colloquial use, should be restricted
to the non-perceptual form of awareness experienced by the lower animals,
such as earthworms.24
For the higher animals, those capable of perceiving entities, sensory
qualities are given as inseparable aspects of whole entities, not as their parts.
To isolate in awareness a given color from the entity possessing it, such as
the fire truck’s red color from the fire truck, is a relatively sophisticated feat,
requiring an act of abstraction. In perceiving the fire truck, one sensory
quality, such as its color, is not set off against its other qualities, such as
its size or shape. What perception isolates are not qualities but entities.
The fire truck is automatically discriminated from the other cars and trucks,
the road, the buildings, etc. The spatial discrimination of entities from each
other is given in the perception; it is not the outcome of some higher act
of cognition. Perception presents us with an array of entities, each set off
against the others in a three-dimensional world.
Even the idea that human beings in earliest infancy go through a sensation
stage, during which they are aware not of entities but only of unintegrated
sensory qualities, is a very dubious scientific hypothesis. There can be no
philosophic argument to show that a sensation stage is required in order to
reach the perceptual level, since many animals give every indication of being
at the perceptual level from birth. E.g., a colt will begin running around only
two or three hours after birth, a feat impossible without perception of entities.
For each sense modality — seeing, hearing, touching, smelling, tast-
ing — there are sensory receptors that respond to stimulation from the envi-
ronment. The means of this awareness is the impingement of energy upon
sensory receptors. But we do not perceive the energy or the action of the
receptors; we perceive entities, things. We do not see light; light enables
us to see the objects that reflect or emit it. We do not hear air-vibrations;
the impact of those air-vibrations on structures in the ear enables us to hear
events in the world. And although we do feel the pressure of objects upon
the body, touch perception is not the experience of pressure but of objects
pressing, an awareness enabled by the force of that pressure. Taste and smell
involve the detection of chemicals; however, the perceptual element is the
tasting and smelling of objects, e.g., of a peach.
What are the sensory structures that make the higher animals capable
of perceptual awareness? This is, of course, a scientific question rather than

24 The term “sensation” is used in as many as seven different senses [Efron 1966, 149].
The failure to distinguish among them has led to widespread confusion.
 H OW W E K N OW • 2 : P erception 68

a philosophic one. But a reference to some recent scientific work will be help-
ful in concretizing the nature of perception.

Direct Realism
The science I present here is based on the achievements of one researcher and
theorist: J. J. Gibson. Gibson revolutionized the scientific study of perception
by developing and building upon a proper philosophic base, which he aptly
termed “direct realism”25 — an understanding of perception that explicitly
and thoroughly rejects the error of sensationalism. Gibson’s direct realism
is the base for what follows.26
Gibson found that perceptual awareness stems from the ability of the
nervous system to detect relationships in the incoming stimulation, especially
higher-order relationships. This requires some explanation.
Consciousness is a difference-detector. One soon loses awareness of a
constant odor, pressure, or background hum. Even the seemingly static
vision of a changeless vista is accomplished and maintained only by constant
micro-movements of the eye, movements of which we are introspectively
unaware. In experiments that stabilize the retinal image, it is found that
visual awareness soon vanishes. [Riggs and Ratliff, 1950]
Further evidence: in the so-called Ganzfeld experiments, subjects’ eye
sockets are fitted with half ping-pong balls, making their visual field
entirely uniform, deprived of all optical structure.27 In the absence of any
differences in the incoming visual stimulation, the subjects soon report not
that they see a spread of white, but that they have stopped seeing altogether.
Sometimes they fear that they have gone blind. Even when the ping-pong
balls are dyed a uniform color, the color soon stops being seen. The same
type of phenomenon occurs in snowblindness.28 Change not stasis, differ-
ence not sameness, is what consciousness is adapted to detect.

25 “Indirect realism” is the invalid notion that what we perceive are not objects in the world
but internal products, such as images or “sense data” that “represent” external objects.
This “representationalist” view, a recurrent error, will be discussed in detail near the end
of this chapter.
26 Gibson’s final and best treatment of perception is in The Ecological Approach to Visual
Perception [Gibson, 1986]. Also quite valuable is his earlier work, The Senses Considered
as Perceptual Systems. [Gibson, 1966]
27 I am indebted to Klaus Nordby for stressing the absence of optical structure.
28 Cf. Gibson’s thought-experiment of a man standing in a uniform fog [Gibson, 1986, 53].
 D irect R ealism 69

Animals capable of perception, as opposed to mere sensation, can detect

higher-order differences — in particular, the difference between what changes
and what does not. This higher-order difference enables a perceiver to use
difference-detection to apprehend constancy, sameness, invariance — things
that consciousness, as a difference-detector, is not otherwise sensitive to or
does not highlight. Constancy per se is not normally detected, but constancy
amid change is detected, because it is actually a (higher-order) difference:
the difference between what is changing and what is not. Perception works
precisely by what Gibson calls “extracting the invariants” — i.e., detecting
constant patterns in the stimulation, patterns that stand out because their
constancy makes them different from the items that change.
One of Gibson’s examples of constancy amid change is the flow of the
visual field resulting from moving forward. All points in the visual field flow
radially outward, towards the periphery — except one point, the point toward
which one is moving. The object that one is directly approaching does not
move to either side in the visual field, as do all other objects. The straight-
ahead object enlarges, but stays where it is in the visual field.29

No static images can capture the moving flow, but the three frames above,
created from the photograph on page 62, are to illustrate how moving straight
toward the middle apple causes an outward flow of all other points, blurred
here to indicate their movement. The edges of the bowl move successively
to the frames’ borders, as do all the apples except the apple one is approaching.

29 I am careful to speak of “in the visual field,” because it is also crucial to realize that the
objects moving across one’s visual field (due to one’s own movements) may be stationary
in reality. In that case, the objects are perceived as being stationary. (The term “visual field”
is an imperfect one, but is the best available: it denotes the psychologist’s perspective on
a subject’s means of perception, not the first-person, phenomenological perspective.)
 H OW W E K N OW • 2 : P erception 70

The incoming light contains patterns, and, at a higher level, patterns of

changes in the patterns. The organism’s ability to detect and capitalize upon
these patterns, including higher-level patterns, is what enables the organism
to discriminate entities. The perception of a spatial array of entities requires
multiple levels of pattern-detection in a complex neural process involving
neural integration at multiple levels (retinal, ganglial, and cerebral).30 The
resulting conscious response is not the “ping” of a solitary receptor reacting
to a zing of energy, but the enduring awareness of entities in the world.
In the visual discrimination of entities, a crucial element is the detection
of edges. Edges produce discontinuities in the pattern of stimulation , a jump
from one color to another or from one brightness level to another.
But edges do not maintain a fixed angular relationship with each other:
as one moves around the object (or, as the object moves), there is a shift in
perspective that changes the visual angles among the object’s edges.
Take the case of viewing the top surface of a rectangular table. When one
moves to a closer vantage point, that changes the angles of the light enter-
ing the eye and that changes the top’s shape in the “visual field” (to borrow
a dubious term): the shape begins as trapezoidal and becomes more so, if one
approaches the table from a low viewing angle. When one moves in relation
to a table, one is normally unaware of the perspectival changes, but these
drawings, since they are two-dimensional projections, make the perspectival
differences quite clear.

30 On the role of pattern-detection in cognition, see Hawkins & Blakeslee, 2005.

 D irect R ealism 71

However, there is an invariant order to these perspectival changes; they

are regular, lawful, and continuous. In approaching the table, at no point
does one find the edge-pattern becoming, say, elliptical. Nor does an edge
disappear (unless there is an intervening object, which is then itself seen).
The regular succession of changes in the visual angles is what causes us to
perceive the table’s shape as being constant, not undergoing the kind of
changes shown two-dimensionally in the drawings on the facing page.
It is often asked: “Why don’t we see the table’s shape change as we approach
it, since the image projected on the retina changes shape with the perspective?”
The question assumes that what we perceive is a rendition of the image
on the retina. But perception is not of an image, nor is perception a passive
transmission of receptor events. Perception is an integrated form of awareness,
one that responds to patterns automatically extracted by the brain from the
ongoing flow of sensory input. As a result, we experience the table shape as
being constant.
In the human eye, there is indeed a retinal image, but the eye is not
a camera. The image on the retina is upside down, perhaps fuzzy, constantly
moving, changing; and there are two eyes with two slightly different images.
Instead of using the camera as a model of vision, we should liken it to a radar
dish sweeping the skies.
Moreover, human eyes respond not to points of light, but to patterns.
These patterns exist in space and in time: there is the pattern of adjacent
elements in a scene and there is a pattern over time regarding how those
patterns move and flow, as the eyes move, the head moves, and the person
walks around.
Gibson writes: “We look around, walk up to something interesting
and move around it so as to see it from all sides, and go from one vista to
another. . . . [A] moving point of observation is necessary for any adequate
acquaintance with the environment.” [Gibson, 1986]
Much more could be said about the detection of higher-order spatial and
temporal patterns, but the preceding should be sufficient to indicate how the
perceptual level operates. The philosophic point is this: by detecting higher-
order relationships, the nervous system is able to present us in perception
with a three-dimensional world of entities discriminated from each other.
No assembling of “awareness atoms” is involved or required. Perception
results from the neurophysiological processing of physical inputs, not from an
experience of “sensations,” or of sensory qualities that somehow get combined.
Sensationalism is wrong; direct realism is right.
 H OW W E K N OW • 2 : P erception 72

Perception as Inerrant
The means of perceiving and the means of conceiving are radically
different. In contrast to the conscious and volitional nature of conceptual
activity, the processes producing perception are physical and deterministic.
With eyes open, you see; if a door slams, you hear it; if an object bumps
against your body, you feel it. The automatic nature of perception is not
contravened by your control over your senses and over the focus of your
attention. You indeed can direct your senses to explore this object instead of
that, and you can control, within a given field of perception, what you attend
to. But in either case, you have no control over what you will perceive in
consequence. The content of perception is dictated by what is there to
perceive and attend to, given the physical conditions obtaining.
Since perception is automatic and deterministic, the information provided
by the senses is an absolute. It cannot be evaluated or judged — neither as
correct nor incorrect, neither as valid nor invalid. The content of perception
is necessitated by the objects stimulating one’s senses, the external condi-
tions obtaining (e.g., the level of illumination, for vision), and the automatic
physiological processing done by the nervous system. Moreover, there is no
“higher” standard against which to judge the deliverances of sense. Perceptual
data is the self-evidently given, against which all conceptual and inferential
processes are judged — not the reverse.
To properly describe perception, I call on Ayn Rand’s distinction between
“the metaphysically given” and “the man-made”:

To grasp the axiom that existence exists, means to grasp the

fact that nature, i.e., the universe as a whole, cannot be created
or annihilated, that it cannot come into or go out of existence.
Whether its basic constituent elements are atoms, or subatomic
particles, or some yet undiscovered forms of energy, it is not ruled
by a consciousness or by will or by chance, but by the Law of
Identity. All the countless forms, motions, combinations and
dissolutions of elements within the universe — from a floating
speck of dust to the formation of a galaxy to the emergence of
life — are caused and determined by the identities of the elements
involved. Nature is the metaphysically given — i.e., the nature of
nature is outside the power of any volition. . . . The metaphysically
given cannot be true or false, it simply is. . . . [PWNI, 25, 27]
 P erception as I nerrant 73

The content of perception is metaphysically given. As such, perception

is unjudgeable. Just as it makes no sense to evaluate a natural occurrence like
rainfall, it makes no sense to evaluate the content of perception. Rainfall is
neither “valid” nor “invalid” — it just is. In exactly the same way, hearing the
rainfall is neither valid nor invalid — it just is.
Questions of validity or invalidity arise only where there is volitional
control of the cognitive process, culminating in a conceptual judgment
— as when you think to yourself: “the pitter-patter I’m now hearing is rain.”
That thought may be true or false, valid or invalid, correct or mistaken.
But none of these things applies to the hearing, as such. The hearing is the
physically necessitated result of the action of sound waves on one’s ears and
what one’s brain, as a physical organ, does with that input.
You control your thinking, your judgments, your reasoning, your inter-
pretation of sensory experiences, but the experiences themselves are
produced automatically, independent of your volition, which means that they
are neither valid nor invalid, but “metaphysically given” facts.
Again quoting Rand: “[man’s] organs of perception are physical and have
no volition, no power to invent or to distort . . . the evidence they give him is
an absolute.” [AS, 1041]
There is indeed a polemical value to saying “Perception is valid,” and
such a statement is unobjectionable, if one means “Perception is of reality.”
But the deeper point is that perception is, if I may put it this way, beyond
valid: as metaphysically given, perceptual data are the standard for judging
what is valid or invalid.
The technical way of putting the conclusion is that perception is “inerrant.”
The content of perception cannot be erroneous or mistaken. More simply,
the point to be affirmed is: “I see things, I hear things. I touch things and feel
them. I am aware, in various forms, of things.”
You cannot mis-see, mis-hear, mis-taste, etc. There is no such thing as
“mis-perceiving.” The very term is a contradiction: to perceive something is to
be aware of it. And there is no such thing as awareness of what doesn’t exist.
We come back to perception as a corollary of the axiom of consciousness:
perceivers perceive.

Perception vs. Conceptual Identification

There is, however, a familiar objection to the idea that perception is inerrant:
the Argument from Illusion. This argument claims that optical illusions,
and illusions involving other senses, show that the senses can be in error.
 H OW W E K N OW • 2 : P erception 74

The millennia-old example is that of the straight object that appears as if

it were bent when semi-submerged in water:

The stick looks bent, but it is actually straight. Isn’t this, then, a case of
mistaken perception? No, for we must distinguish between the act of seeing
and the use of concepts to describe what is seen. There is nothing erroneous
about the stick’s appearance; one’s eyes and brain are functioning as their
nature demands. The perceptual data are not wrong or mistaken — but they
can be misleading : a naïve observer is likely to conclude: “This stick is bent.”
If he does, it is that conceptual judgment, not the seeing, that is mistaken.
The water’s refraction of light makes the stick look bent (i.e., resemble
actually bent sticks), and one expects it to still look that way out of water:
 P erception as I nerrant 75

But seeing is seeing, not predicting. The sheer sight of the stick is not
a prediction as to how it will look out of water, or in other conditions of
perception; perception does not transcend time in that way, reaching into
the future. (There are indeed perceptual-level associations and expectations
about the immediate future formed pre-conceptually, as when a kitten learns
to associate a match flame with a painful burn and subsequently to avoid it.
This is perceptual association, not conceptual judgment.)
Expectations based on perceptual association sometimes fail to be fulfilled.
For instance, a pet cat may associate hearing a certain sound in the kitchen
with being fed but not be fed. Such frustrated expectations are not errors
in cognition. Only states of awareness, not things like forming and using
associations or expectations, count as acts of perception.31
In the bent-stick “illusion,” what we see is the way a straight stick looks
when semi-submerged in water. The image on the left is, after all, a photo­
graph. The camera does not lie, and neither do the eyes.
Vision gives us an awareness of things in a certain form — how things
look. Hearing gives us how things sound. Touch gives us how things feel.
There is not and could not be a perception of things the way they do not look,
sound, feel, etc. Things look, sound, and feel the way they must, given the
sensory inputs and the brain’s automatic processing of them. (Here, the water
refracts the light waves, as it must.)
There are countless other illusions, involving shape, color, lightness and
darkness, motion — you name it. But they all have the same form:

Our senses tell us so and so.

But so and so is not the case.

Our senses have erred.

The error is in the first premise. Our senses do not talk to us. The senses
do not form propositions. They do not make judgments. Perception is only
perception, not perception plus a proposition. Your sight of the stick does
not even include the simple proposition: “That is a stick.”
It is crucial to be absolutely clear on what is perception and what is more
advanced than perception.

31 Cf. Gregory Salmieri’s distinction between perception and “post-perceptual processing.”

[Salmieri, 2006]
 H OW W E K N OW • 2 : P erception 76

“Perception” includes: seeing, hearing, touch, smelling, tasting, and aware-

ness of things going on in our bodies (proprioception). “Perception” does not
include: association, expectation, prediction, classification, inference, propo-
sitions, intellect, reason, interpretation, judgment, thought.
There is a linguistic signal in English for the difference between the
perceptual and the conceptual: the locution “seeing that” always indicates
a judgment, never just perception. You see a tree. But to see that it is a tree
is “seeing” in only a metaphorical sense. To see that something is a tree is to
go beyond the perception to subsume what is perceived under the concept
“tree.” Likewise, we are not dealing just with perception if we see that the tree
is big, that the tree is old, that its leaves are green. Perception includes only
seeing the big, old, green-leafed tree.
Earlier, I discussed the error of reducing perception to “sensations.”
The Argument from Illusion makes the inverse error: reading the conceptual
level of awareness into perception. Where the doctrine of sensationalism dis-
integrates the perceptual level, implying that man does not directly perceive
entities, the other side of the same false coin intellectualizes the perceptual
level, claiming that perception involves the use of concepts, inference — even
theories. But perception is neither less than it is nor more than it is.
Perception is neither sensations nor ideas. Perception is an integrated, but
pre-conceptual, awareness of entities — a level of awareness not possible to
an earthworm but one enjoyed by such non-theorizers as frogs and fish.32
If we eliminate from the Argument from Illusion the confusion of percep-
tion with conceptual judgment, we have:

Under these conditions, this looks a certain way.

Under other conditions, this looks a different way.

Therefore . . . ?

No anti-senses conclusion follows.

In the stick case in particular, the actual situation is this:

32 Even Gibson succumbs to error here, holding that perception includes awareness of what
he calls the “affordances” of things — i.e., their potential benefit or harm for the agent.
Though some of things’ simple causal powers are indeed given perceptually, the learning
of many of them is either by association (which is not direct perception) or, for man,
by reasoning. Gibson goes so far as to hold that we perceive that a mailbox “affords”
mailing a letter. In other respects as well, Gibson seems to be unclear on the perceptual-
conceptual distinction.
 P erception as I nerrant 77

When semi-submerged in water, this stick looks similar in

shape to things I have classified as “bent.”
But when taken out of water, this stick looks different in shape
from those things.

Therefore . . . ?

The senses have made no “error” in responding as they do, and there is
no philosophic problem concerning the fact that some things, under certain
conditions, look like other things. One man from a certain angle resembles
another. A fox seen from a distance resembles a dog. The fact that one thing
can resemble another does not threaten the inerrancy of the senses.
There is nothing in the perception — the sheer seeing — capable of being
either correct or mistaken. Any misclassification prompted by the perceptual
appearance is an error not of perception but of conceptual classification.
A thing’s appearance depends upon both its own nature and the conditions
of perception. The green hills, viewed from a distance, look blue — i.e., they
look the way blueberries look from close up. But no error occurs — until and
unless one makes a conceptual judgment: “Those hills are blue.”
The same analysis applies to the Müller-Lyer illusion, shown below.

The two lines look the way lines unequal in length look, but they are of
equal length. But no error occurs — until and unless one makes a conceptual
judgment: “Those lines are of unequal length.”
The looks of things do not interpret themselves, or tell us how to classify
them, or how the things perceived would look in other conditions — under
different lighting, from close up, or with the two lines moved close together.
This analysis resolves all cases of illusion. They all confuse the perceptual
with the conceptual. They all attempt to make philosophical hay out of
a perceptual similarity that tempts one to make an erroneous judgment.
Animals are not perplexed by perceptual “illusions” because they do not have
concepts and therefore do not make classifications. A given “illusion” may
cause an animal to act in ways inappropriate to the actual facts — as when
a cat sees itself in the mirror and reacts as if another cat were present — but
the animal is not “making a mistake.” To call it a “mistake” is to project
 H OW W E K N OW • 2 : P erception 78

our human perspective onto the animal. The actions of the intellect in inter-
preting, describing, or classifying perceptual information can be mistaken;
the perceiving cannot be. Perception is awareness, and there is no such thing
as “mis-awareness.”
Often confused with illusions is a different type of case, that of color blind-
ness. “Who is right,” it is asked, “the man who sees red and green or the
color-blind man who sees two patches of the same color?”
The answer is: neither is right and neither is wrong. The color-blind
person lacks the power to make certain discriminations. It is not correct to
say that he sees red as being green; rather, where we see a color-difference,
he does not. Due to a physiological defect, he cannot detect certain differ-
ences that a normal person can detect. The issue, then, is not that one man
sees red where the other sees green; rather, one man is sensitive to the red-
green difference and the other is not. There is no more a contradiction here
than between a sighted man and one completely blind.
The color-blind man gets less information — but not “misinformation”:
whatever information he gets is an unevaluable, inerrant absolute. One can
make comparisons regarding the quantity of information gained, but
not about the quality of that information. Any information acquired
is information.
Different species, having different sensory physiology, pick up different
kinds of information. Dogs are able to hear frequencies too high for us
to hear; bees respond to ultraviolet light that we cannot see; bats perceive
quasi-visually by means of echo-location; migratory birds are sensitive
to the polarization of light. But whatever an organism is aware of, it is aware of.
Perception is perception.

Form vs. Object

The dependency of perception upon its causal means has deeper implica-
tions: the content of perception is shaped by the nature of the organism’s
sensory physiology. Skeptics have used this fact to argue that perception
is subjective — holding that we perceive not the world as it is, but only
the world as it is relative to us — not the real world, but only a world of
appearances.
It is in that kind of woozy language that this attack on perception is made.
But recasting it in a clear form will not only lay bare its error but also high-
light an important truth. Here is the better statement of this argument:
 P erception as I nerrant 79

The content of perception depends on our nature.

Objects in the external world do not depend on our nature.

The content of perception is not objects in the external world

(but something in us).

The above may be called The Argument from the Relativity of Perception.
The error in the argument is an insidious equivocation on “the content of
perception.” That phrase equivocates between the object perceived and the
form in which we perceive it.
Perception, like all conscious activities, has two aspects: the form of awareness
and the object of awareness. Distinguishing form and object is essential
to understanding consciousness on both the perceptual and conceptual
levels.33 The object is what you perceive — e.g., a table. The form is how you
perceive the object — e.g., as large, brown, rectangular, smooth, and so on.
We perceive the object in the form of its sensory qualities.
If we make the form-object distinction, The Argument from the Relativity
of Perception becomes quite different:

The form in which we perceive depends on our nature.

Objects in the external world do not depend on our nature.

The form in which we perceive is not objects in the external

world.

Thus, if we use the proper first premise, we reach a harmless (and fairly
meaningless) conclusion.
To accomplish its skeptical purpose, the argument would have to have as
its first premise: “The objects we perceive depend on our nature.”
But that is manifestly false. The table we perceive exists and is what it is
independently of us. The skeptic’s argument needs to equivocate — to switch
from a premise about form to a conclusion about object.
The form of perception is indeed “observer-relative.” But that does
not mean the form in which we perceive reality is subjective or distorted.
The nature of one’s perceptual systems is what it is and acts as it has to act,

33 The form-object distinction traces back to Aquinas, who has a brief, almost in-passing,
identification of the same distinction as the id quod vs. the id quo — or the “that which”
vs. the “that by means of which” [Summa Theologica, I, Q. 85, 2]. This distinction is empha-
sized in both Thomistic philosophy and Objectivism.
 H OW W E K N OW • 2 : P erception 80

given the conditions. The resulting form of perception also is as it has to

be, given the nature of the object, the perceiver’s nature, and the current
conditions of perception.
In perception, both form and object are metaphysically given facts.
Both are unjudgeable. We cannot evaluate what we perceive — the object —
as true or false, valid or invalid. If what you perceive is a tree, you cannot
judge the tree as a valid or invalid object. It is simply the object, the tree.
Similarly, you perceive the tree somehow — i.e., in a certain form, as
looking like this, with this color, shape, size, texture, etc. That specific
form of perceiving the tree results from the automatic, physiologically
determined processes of your nervous system. You do not make it up,
and you cannot change how you perceive the tree by ordering your-
self to perceive it differently. Nor can you perceive the tree in the form
that, say, a bat does — by the tree’s echoes of high-frequency sound waves.
That simply isn’t the human means of perception.
It is nonsensical to maintain that one ought to be able to perceive the
tree in a different form — as, say, vibrating atoms. There is no more basis for
that “ought” than for the idea that rain ought to fall upwards. Attacking the
senses for providing information in the form that they do amounts to what
Ayn Rand called the fallacy of “re-writing reality”: one is implicitly main-
taining, “If I had written the script for reality, I would have had rain falling
upwards and trees being perceived as collections of atoms.” But reality is what
it is — and one’s physical means of perception is part of reality; the form of
perception resulting from those means is metaphysically given.
There is no “right” or “wrong” way for things to look or sound or feel.
There is only the way that they do look, sound, and feel — as they must, given
their nature, the conditions, and the nature of our perceptual apparatus.
Things have an appearance — but the appearance is the form of perception,
not the object of perception. The visual appearance of a tree is how the tree
looks — how (form), the tree (object), looks.
The form of perception is partly dependent on our human physiology.
But, the object of perception is not. The form is the form of awareness of
objects. It is how we perceive, not what we perceive. What we perceive is the
object — that is what is meant in this context by “object.”
I want to emphasize that the form-object distinction is not some minor
point trotted out to answer skeptics. The distinction between the what and
 P erception as I nerrant 81

the how of cognition is essential to the whole approach to consciousness I am

presenting here, the approach developed by Ayn Rand. I will give a series of
examples to concretize and clarify this crucial distinction.
Consider what happens when you change the tint control on a color
TV, so that what looked green before, now looks bluer; what looked blue
before, now looks redder, etc. You still see the same objects, but in a different
form. You have not changed the channel, which would result in your seeing
a different set of objects. Changing the tint changes only the form, while
leaving the same object and the same information content.
No matter how bizarrely you set the tint control, you cannot see colors that
are wrong — merely colors that are different from the ones you are accustomed
to seeing. It is not the case that the normal form is the right way for objects to
look and the other is the wrong way; what you call the “right” tint-setting is
simply the one that makes the televised world match the way you perceive the
natural world. That is a comparison of the form of perceiving the screen
images with the form of perceiving actual objects — a comparison of form
to form. But what you cannot do is compare form to object; you cannot
compare a banana’s seen color with its “real color.” Color is a form of
perception; the “real color” of the banana is the color it is seen to have.
Suppose that, from birth, your nervous system had been peculiarly
different, so that you always saw things in a shifted spectrum. Suppose that
grass looked blue, the sky looked red, etc. If so, then those would be the
normal colors, for you. And in learning language, you would learn to use
the word “blue,” not “red,” to denote the color of the sky, even though
your form of perceiving the sky was like my form of perceiving a fire truck.
Neither your spectrum nor mine would be “right” or “wrong,” “truer” or “less true”
— just as neither the Fahrenheit scale nor the Celsius scale is the “right” scale,
and neither English nor French is the “true” language. There is no such thing
as “the way things look to God” — no standard of how things “should” look
in contrast to how things do look, as a necessary result of one’s perceptual
physiology. There is no look apart from a looker.
One’s senses are like physical instruments, which respond to physical
stimulation in the form that physical causality dictates.
Consider an actual physical instrument: a standard mercury thermometer.
The height of mercury in a thermometer does not reproduce or resemble
or copy temperature; it registers it. The senses do not reproduce reality,
 H OW W E K N OW • 2 : P erception 82

they register reality. They register, ultimately, the information contained in

the patterns of energy that strike them.34
In a mercury thermometer, sometimes a gap appears in the column of
mercury. But this does not mean that the thermometer then gives a “false”
result. It still registers the temperature, but now the person reading the
thermometer must re-calibrate the scale in order to interpret the results
correctly. The thermometer still gives a causally necessitated result that
accurately reflects the temperature. Even if a defect in the thermometer
makes it respond non-linearly, the instrument remains an accurate and use-
ful source of information as long as there is a one-to-one correspondence
between each temperature and each position of the mercury. To use this
“defective” thermometer, one need only mark it with the appropriate non-
linear scale.
In general, as long as each state of the instrument correlates with
a single state of what it is to detect, the instrument is functional, and capable of
giving an informative reading. The same is true of the senses, with the
“calibration” (i.e., determination of conceptual meaning) being done by
the intellect. The senses, like measuring instruments, can fail to function,
but if they function at all, they cannot function erroneously. Every state of
a thermometer, a voltmeter, a Geiger counter, is causally necessitated, and so
is every state of a perceptual system.
Today, modern technology makes it easy for us to understand that infor-
mation can be displayed in various ways, and that the form in which we
display information does not alter the content of that information.
For instance, a sonogram uses sound echoes to make a visual display of
shape. The display on a radar screen presents information extracted from
reflected radio waves.
Consider the fax machine. A photosensitive element scans across the
document, registering each tiny dark or light patch as a high or low voltage.
Then these patterns of voltage are transformed into tones that are sent across
the phone wires. At the other end, the tones are turned back into voltage
patterns that cause a printer to reproduce the black and white patches, giv-
ing the same dark-and-white pattern as in the original. That is an example
of how information can be changed in form without losing any of its
information content.

34 It is proper to use the term “information” here, but not in regard to the contents of
a computer (see Chapter 1) precisely because we are discussing the relation of physical
facts to a consciousness.
 P erception as I nerrant 83

A particularly dramatic example of different forms of perceiving the

same object is provided by a prosthetic device developed to enable blind
people to acquire visual information. The device is a video camera mounted
on the patient’s forehead. The information from that camera is processed
and transformed into presses on the patient’s back by an array of blunt pins
that he wears. By the pattern of the pins’ pressure, the blind person feels,
in the form of pressure on his back, the pattern of light, including how that
pattern changes as he moves around. As a result, the blind person effectively
“sees” objects in his environment by means of feeling touches on his skin.
He picks up the same kind of information, though with less resolution,
that a sighted individual does, but in a different form — as touch rather than
as sight. [Bach-y-Rita, 1969]
Perhaps the most decisive example of form vs. object occurs when the
same facts are perceived through different sense modalities. We become
aware of an object’s shape both by sight and by touch — two different forms
of perceiving the same fact. Likewise, the texture of sandpaper is perceived
both by sight and by touch: same fact, different perceptual forms.
The form-object distinction makes short work of an old philosophical
chestnut: adapting one’s left hand to cold water, one’s right hand to hot water,
then plunging both hands into the same pail of lukewarm water, which then
feels warm to one’s left hand and cool to one’s right hand. The difference
in the state of the two hands results in two different forms of response to the
same water-temperature. There is no contradiction between the two forms
of awareness, and neither is in error.
Next, consider an example that switches sense modalities. If you push on
the lower outside corner of your bottom eyelid, the pressure should stimu-
late some of the receptors in your retina. If done properly, a circle of light
will appear in the opposite part of your visual field (by your nose and near
the top of your visual field). Here, you are mechanically stimulating retinal
receptors, which are normally stimulated by light. The result is a neural dis-
charge sent to the brain which produces the experience of a circle of light.
You have a visual form of response to the pressure of your finger. It is an
unusual form of being aware of pressure, but the content is the same as when
you feel pressure by means of receptors in your skin. The object — and the
information — is the same.
As a final example, consider your awareness of heat, in the form of
warmth, when something contacts your skin. Physically, the heat is the
vibratory motion of the molecules in the object touched. Now, imagine
 H OW W E K N OW • 2 : P erception 84

a race of humanoids on another planet who experience that heat in the form
of vibrations, not warmth. The difference in the two forms of perception
would have no philosophic significance. Neither form could be declared
to be the “right” or “wrong” way to perceive heat. Both forms of perception
would be metaphysically given, and both forms would provide information
about the object.35

Perception vs. Hallucination

What about delusions or hallucinations? For instance, there is the following
common experience. When in quiet surroundings, perhaps lying down in
bed before falling asleep, one sometimes hears for a couple of seconds a pure
high tone. The sound is caused by something internal to the ear. But it is
physically caused, not the result of the mischief of some intervening spirit,
like Descartes’ evil demon. One is hearing something, something real, only it
is a state of one’s body rather than something in the external world.
Take the extreme case. Suppose someone swallows LSD and “sees” ants
crawling on his arm. But there are no ants. Isn’t he mis-perceiving? No,
because what he is doing is not “perceiving” — not in regard to the phantom
ants. He is hallucinating. Hallucination is not a kind of perception. It is an
experience in consciousness, but the hallucinatory content does not repre-
sent awareness. The LSD-ingester is not “seeing ants” — because there are no
ants there to be seen.
It must be noted that the hallucination, though not a perception, is
nonetheless causal and deterministic. What the LSD-ingester is experiencing
is the chemical action of the LSD on his nervous system. If he concludes
from the experience that ants are present, it is the conclusion that is wrong.
The experi­ence — even though not a cognitive one, not a case of perception
(insofar as the ants part of it is concerned) — is neither erroneous nor cor-
rect; it is just an experience.
The experience, though not perception, still has a form. (It has no object
but it does have a content.) The form is the colors and shapes (the colors
and shapes of ants); the cause of that content is the action of the drug on the
nervous system. One might think that the experience represents a perception
of the LSD, or of its action on the nervous system. But the experience fails to
qualify as a perception of the LSD because the experience does not include
any discrimination of the drug or its action from other things; one is not
35 This example was given by Ayn Rand in the second lecture of the series “Basic Principles
of Objectivism.”
 A F inal D efinition of P erception 85

aware of a spatial array of co-present entities within which the LSD is located.
Contrast experiencing LSD in the form of ants on one’s arm with actually
perceiving LSD — looking at it in a test tube or, using some sort of brain scan
to follow its path through another person’s brain. The delusory experience
has a form and a content but does not locate a discriminated existent among
others in the world. Thus, the experience is not a perception — neither of the
LSD nor of (non-existent) ants.
There can be cases in which one experiences a hallucination due to other
bodily disturbances, such as extreme hunger, thirst, or prolonged sleep depri-
vation. Such hallucinations are not acts of perception and are not produced
by external objects stimulating the senses. Even when the person having
such hallucinations cannot distinguish them from perception, that fact casts
no doubt on the cognitive value of perception. The remedy called for is not
epistemological — not some process of “checking” — but medical. If he mistakes
the content of his hallucinations for perceptual awareness of the world, the
error lies in that judgment.
Thus, what may be called “The Argument from Delusion” fails: hallucina-
tion and the like are not instances of perception, and neither is any erroneous
judgment one makes in response.
Dreams are not hallucinations, but obviously they are not cases of percep-
tion. There is no way to get from “With my eyes shut, I dreamed something
unreal” to “What I see with my eyes open is unreal” (an idea that steals the
concept “unreal”). Nor is there any philosophic significance to the fact that
while asleep one is not in a condition that permits distinguishing dreams
from wakeful perception: while asleep, a person is not able to make any cog-
nitive judgments. A person asleep cannot “think he’s awake and perceiving,”
because he cannot think at all. (And the question: “How do you know you
are not dreaming right now?” already presupposes that the person one is
addressing is awake and not dreaming.)
The Argument from Delusion, like the Arguments from Illusion and from
the Relativity of Perception, fails to undermine the inerrancy of perception.
It had to fail; otherwise we would know nothing, including these arguments.

A Final Definition of Perception

On the basis of all the preceding discussion, I offer the following, more
advanced, definition of “perception”:
 H OW W E K N OW • 2 : P erception 86

“Perception” is the direct awareness of reality, in the form of spatially

arrayed entities, that results from the automatic neural processing of actively
acquired sensory inputs.
(It is taken as understood that the awareness is ongoing, not momentary or
episodic, and that perception is “metaphysically given,” and hence inerrant.)

Three General Conceptions of Consciousness

1. Objectivism: consciousness as identification
As I said at the start of this chapter, the biological approach to understand-
ing consciousness recognizes that its function is to guide action, and that its
essential means of fulfilling that function is by discovering the identity of
things in the world — in Rand’s statement: “consciousness is identification.”
[AS, 1015] Identification is an active process, with definite stages or steps.
But it takes a certain sophistication to appreciate the active nature of
consciousness and much more sophistication to make the distinction
between the form of perception and the object of perception. The unsophis-
ticated attitude toward perception tends to be that known as “Naïve Realism.”

2. Naïve Realism: consciousness as reproduction

Naïve Realism is a form of realism in that it implicitly accepts the primacy
of existence: Naïve Realism recognizes that consciousness is awareness of
what exists. But the naïve part of Naïve Realism is that it construes percep-
tion on the model of mirror reflection. Just as a mirror passively reproduces
an image of an object, so consciousness, on this conception, creates internal
replicas of objects in the world.36 Naïve Realism holds that consciousness is
not identification but reproduction.
This view is captured in a phrase often used by the Ancient Greeks:
“Like knows like.” The meaning is: to know is to become like the thing known.
Seeing a tree, for the Naïve Realist, consists of forming a copy of the tree
inside one’s mind.

36 I use the mirror analogy despite the fact that flat mirrors do not actually produce an
image, just reflect light. A parabolic mirror properly set up, however, can produce such
an image of an object.
 T hree G eneral C onceptions of C onsciousness 87

(When made concretely real, as in this diagram, Naïve Realism may well
appear to be childishly silly. Nonetheless, it is the implicit premise shaping
most people’s initial approach to understanding perception, and eradicating
all traces of Naïve Realism from one’s thinking often takes some vigilance.)
The internal replica of the tree is held to be mental, not physical:
what is reproduced inside one’s mind is said to be the form of the tree
without its matter. Aristotle seems to take this view when he analogizes
seeing to the pressing of a signet ring into wax.37 The wax is like one’s
consciousness and the signet ring is like the external object perceived.
The idea is that in perception, as with impressing the signet ring, the form
(here shape) is duplicated — in the one case it is reproduced in consciousness
and in the other reproduced in wax.
What drives the idea that “like knows like”? I believe it is the desire to get
underneath consciousness, to reduce it to something else. Likeness is a famil-
iar relationship, so it is taken as the model.38 But, as I argued in the first
chapter, consciousness is irreducible. The attempt to get “beneath” awareness,
the attempt to say what awareness consists of, is fundamentally mistaken.
Awareness is a primary relation of a subject to an object and is not to be
construed on the model of any other relation.

37 De Anima, II, 12, 424a17–24. Cf. ibid., II, 5 and III, 2, 425b27–426a26.
38 In the modern era, the same motivation applies, I believe, to the “new realists” who seek
to reduce consciousness to “the intentional relation of identity.” See, e.g., Parker, 1953.
 H OW W E K N OW • 2 : P erception 88

The Naïve Realist view has important implications for how one regards
the identity and activity of consciousness. If consciousness is reproduction,
then accuracy in reproducing the external object requires that conscious-
ness be static and featureless, rather than active and possessed of its own
identity. For a pond’s surface to clearly reflect things, it must have no motion,
no waves or ripples to alter or destroy the image. Naïve Realists assume that
consciousness must likewise be still and “pure” if it is to accurately reproduce
objects. But in fact consciousness is active, and this activity is essential to its
achieving awareness.
Naïve Realism has insuperable problems. The most obvious problem
is that it conflicts with the fact that perception does involve physiological
processing, and that this processing does determine the form of the resulting
perception. Differing means of perception result in different forms of
perception of the same object, a fact exploited by skeptics in their
­Argument from the Relativity of Perception. E.g., fatal to Naïve Realism
is the example of the two pails of water: if the same water is warm to one hand
and cool to the other, how can one believe that perception is reproducing
the state of the object?
But the reaction against Naïve Realism was directed at the wrong target
— the Realism instead of the naïveté. The fact of sensory processing was taken
to mean that we do not perceive the external world but only the product of the
processing. The processing was taken to create an internal object — an image
in consciousness — with that mental image, not reality, being the direct
object of our awareness.

3. Representationalism: consciousness as self-consciousness

Naïve Realism held that consciousness is reproduction (the perceiver taking
on the form of the external object). The subsequent, utterly disastrous view
held that consciousness is representation. To perceive, according to Represen-
tationalism, is to be aware of an internal object — an image, an impression,
an appearance — which may represent the external object.
According to Representationalism, what we know directly are ­appearances,
images in our minds, not the external world. We get the report of the
senses, as interpreted by the processing, and we are aware of that report,
not things — or not “things as they are in themselves,” in Kant’s phrase.
Representationalism unwittingly accepts the Naïve Realist assump­tion
that consciousness, to be valid, would have to be a causeless revelation. Then,
 T hree G eneral C onceptions of C onsciousness 89

in response to the recognition that perception is causal, these theorists

move “true consciousness” inside the mind. What results is a kind of interior
Naïve Realism: one directly (and causelessly) perceives only contents inside
of one’s mind. Perception is then construed as the observation — in effect, by
a “little man in the head” — of these internal images. Reality recedes behind
a “veil of appearances,” never itself to be perceived, but figuring only as the
hidden cause of the images.

Representationalism is rarely advocated openly and explicitly. It lives in

the shadows, as an unidentified assumption warping theories of perception
and subverting our understanding of consciousness as such. When the repre-
sentationalist model is made explicit, as Gibson does in the following passage,
its insoluble problems and hopeless circularity become unmistakable.

[It is] tempting to believe that the image on the retina falls on
a kind of screen and is itself intended to be looked at, that is,
a picture. It leads to one of the most seductive fallacies in the
history of psychology — that the retinal image is something to be
seen. I call this the “little man in the brain” theory of the retinal
image, which conceives the eye as a camera at the end of a nerve
cable that transmits the image to the brain. Then there has to
 H OW W E K N OW • 2 : P erception 90

be a little man, a homunculus, seated in the brain who looks at

this physiological image. The little man would have to have an
eye to see it with, of course, a little eye with a little retinal image
connected to a little brain, and so we have explained nothing
by this theory. We are in fact worse off than before, since we are
confronted with the paradox of an infinite series of little men,
each within the other and each looking at the brain of the next
bigger man. [Gibson 1986, 60]

Representationalism’s unstated credo is: consciousness is self-­consciousness:

we perceive our perceptions, observe our observations, and thus are aware
only of what our minds create in us, not the external world.
But in perception, there is no internal image (which would then itself have
to be perceived, leading to Gibson’s infinite regress). The way things appear
is not what we perceive but the form in which we perceive them. We perceive
objects in the world, as they appear — i.e., in the form necessitated by our
means of awareness. The appearance of the tree is the appearance of the tree,
the manner in which the tree is given in perception.
Moreover, lacking the distinction between the metaphysically given and
the man-made, Representationalists treat neural processing as if it were
volitional and conceptual, as if perception were an interpretation of the world
made by an internal journalist. Still worse, the objectivity of the internal
journalist is not ascertainable, since we cannot get out from behind our eyes
to see, without eyes, how things “really are.”
Representationalism begins with Descartes and culminates in Kant.
Descartes held that we could be deceived about everything — except our
having internal states. Descartes retained the primacy of existence to the
extent of taking the correspondence of these states to the external world as
the standard of validity. Kant went whole hog on the primacy of conscious-
ness, rejecting the standard of correspondence to reality. Reality, he held,
is unknowable (which it would be, if Representationalism were true),
so we cannot talk about such a thing as correspondence to reality, and we
cannot take that correspondence as any kind of standard. Instead, Kant held
that all standards have to be consciousness-based — i.e., subjective (which he
calls a new form of objectivity.)39

39 “Hitherto it has been assumed that all knowledge must conform to objects. But all our
attempts . . . have, on this assumption, ended in failure. We must therefore make trial
whether we may not have more success in the tasks of metaphysics, if we suppose that
objects must conform to our knowledge.” [Critique, B, xvi]
 T hree G eneral C onceptions of C onsciousness 91

Among psychologists studying perception today, Representationalism is the

near-universal model. Seeking to avoid using terms like “consciousness” and
“awareness,” psychologists substitute the term “representation.” They assume
that “veridical” perception consists of the resemblance of the internal repre-
sentations to external objects.
This assumption, though utterly without foundation, can be hard to
dislodge. The assumption appears in a welter of absurd questions, such as:
“Since the retinal image is upside down, how does the brain turn it right side
up?” “Since the retinal image of a distant object is small, how does the brain
compensate for that to interpret it as representing a larger object at a dis-
tance?” “Since the retinal image is moving around due to both tiny and gross
movements of the eye, how does the brain stabilize the image?” — and so on.
All of these questions assume that the job of the sensory system is to
produce an image in the mind, an image that looks like the object in reality.
Since the object is right side up, large, and at rest, an accurate internal image
of it must be as well.
But the actual job of the sensory system is not to produce an internal
image that “looks like” the object. The perceptual system’s job is to detect and
make us aware of things, in a specific form. Note that the level of mercury
in a thermometer does not “resemble” temperature — whatever that would
even mean.
Representationalists respond to the skeptics’ attacks on the senses
by beating a retreat back inside consciousness. “Okay,” the Representational-
ist concedes, “I can doubt that my experiences accurately copy things out
there; I can even doubt the existence of the external world; but at least I know
that I’m having experiences, and I hope that they at least roughly correspond
to external things.” But as the history of thought following Descartes shows,
this retreat provides no refuge from the all-consuming skepticism to which
Representationalism inevitably leads.40
Naïve Realism holds that perception is internal effigy-making (“like knows
like”). Representationalism’s “advance” is to add effigy-viewing to effigy-
making. Both Naïve Realism and Representationalism hold, arbitrarily,
that valid awareness can involve no processing. The Representationalist tries
to uphold the validity of consciousness by positing an unprocessed awareness
that takes place between the “little man in the head” and the internal image.

40 Descartes attempted to escape from this skepticism by arguing that God’s goodness
guarantees the general validity of our perceptions, an argument riddled with arbitrary
assumptions. In principle, there is no way to “prove” existence from consciousness:
the concept of “proof ” presupposes that one has grasped that existence exists.
 H OW W E K N OW • 2 : P erception 92

But, in fact, processing is inherent in the nature of consciousness.

Consciousness is not mirroring but identification — identification achieved
by specific means. The false assumption of both Naïve Realism and Represen-
tationalism is: if consciousness has an identity, if it is something that works
by a specific causal means, it cannot be conscious of the identity of things
in the world. The Law of Identity can apply to the world or to consciousness,
but not both, they assume. If the world is something definite, then to
know it, the mind must be a nothing. If the mind is something, then the
world is unidentifiable, which means that consciousness is not conscious.
Accordingly, epistemology collapses into the impossible alternative of
mysticism or skepticism. Rand neatly summarizes all this in her work on
epistemology:

The hallmark of a mystic is the savagely stubborn refusal to accept

the fact that consciousness, like any other existent, possesses
identity, that it is a faculty of a specific nature, functioning
through specific means. . . . [T]he last stand of the believers in the
miraculous consists of their frantic attempts to regard identity as
the disqualifying element of consciousness.
The implicit, but unadmitted premise of the neo-mystics
of modern philosophy, is the notion that only an ineffable
consciousness can acquire a valid knowledge of reality, that
“true knowledge” has to be causeless, i.e., acquired without any
means of cognition. . . .
All knowledge is processed knowledge — whether on the
sensory, perceptual or conceptual level. An “unprocessed”
knowledge would be a knowledge acquired without means of
cognition. [ITOE, 79–80, 81]

The sensory qualities — e.g., redness, loudness, and sweetness — are actually
perceptual forms. Taking these qualities as our examples will bring the basic
issue into sharp focus. Naïve Realism holds that red­ness and sweetness
exist in the apple alone, apart from any relation it has to consciousness.
Represen­tationalism holds that redness, sweetness, and all of the sensory
qualities exist in the mind of the perceiver alone, not in the object. On
a proper understanding, the sensory qualities are not locatable in this manner
— they are neither in the object alone nor in the perceiver’s mind alone but
are the product of the interaction of the perceiver with the object.41

41 See the section headed “Sensory Qualities as Real,” in OPAR, ch. 2.

 A ppearance and R ealit y 93

The apple has certain inherent properties such that when it interacts with
our specific sensory system, the result is an awareness of the apple in a par-
ticular (red) form. The point is more easily grasped in regard to sweetness,
which resides neither in the mind alone nor in the apple alone, neither “in
here” nor “out there” but rather as the product of the interaction between the
perceiver and the perceived. To clarify this point, Rand gave the analogy of
a collision between a car and a truck.42 The collision is in neither the car
alone nor the truck alone, but in both — i.e., constitutes their interaction.
We perceive every object in a certain form, resulting from the interaction of
its identity with the identity of our sensory system.

Appearance and Reality

The understanding of perception given in the last two sections points the way
to gaining a proper understanding of the traditional distinction between
appearance and reality. We use this distinction to contrast what a thing
appears to be versus what it really is. When does this contrast exist? When
there are potentially misleading similarities: “The man appeared to be Jim,
but it was really Bob.” Or, “The stick appears to be bent but in reality it is
straight.” But the “really” or “in reality” here does not refer to something
beyond or independent of perception. Rather, “in reality” refers to what is
disclosed in other acts of perception. It is further perceptual information that
shows us the need to correct the judgment, by concluding: “That is not Jim,
but Bob” and “The stick is not bent, but straight.”
Nor does the new judgment throw out as “mistaken” the original percep-
tual data: one does not say “My eyes were deceiving me when I looked at Jim.”
Jim does have an appearance similar to Bob’s; a photograph of the two of
them would also display that similarity. The semi-submerged stick does look
similar to the way bent objects look out of water. What gets revised in such
cases is the conceptual judgment of what is perceived.
None of the perceptual data itself can be rejected or treated as invalid.
The later judgment includes the initial perceptual appearance, rather than
contradicting it or wiping it out. To be Jim includes the fact that he looks like
Bob. To be a straight stick includes the fact that it will look like a bent stick
when semi-submerged in water. To be yellow includes looking black under
a blue light. How something appears under a given condition of perception
(and of the perceiver) is a matter of ineluctable cause and effect (as studied

42 Rand in personal conversation with Leonard Peikoff, circa 1972.

 H OW W E K N OW • 2 : P erception 94

by the science of psychophysics). The judgment “This is X” integrates the

thing’s perceived qualities here and now with its perceived qualities across
all other conditions of perception.
Accordingly, there are two perfectly proper ways that we use the concept
of “appearance.” The first expresses uncertainty, as when one says, “This
appears to be gold, but I’m not sure.” (Or, when put into the past tense,
it expresses a revision of one’s judgment: “It appeared to be gold, but it turned
out to be iron pyrite.”)
The other correct use of “appearance” is one that is more relevant to the
issue of perception. This use occurs when one distinguishes a thing’s sensory
qualities — how it looks, feels, tastes — from its constituent physical prop-
erties. For instance, a surface will appear smooth to the touch when it has
only very small irregularities. In this case, we are not saying that the surface
isn’t “really” smooth; that idea makes no sense. Rather, we are saying that
a certain physical state of affairs produces a certain sensory quality.
The wrong view of the distinction between appearance and reality occurs
when philosophers claim that perception as such gives us “only” appearance,
not reality. This is a disastrous error. Do I know only the appearance of the
pen in my hand, not how the pen really is? One cannot make that kind of
distinction between appearance and reality.
The claim that perception does not provide awareness of reality (or of
reality “as it really is”) represents a massive stolen concept. If we perceived
only appearances, never reality, we could not have the concept of “reality”
(nor the concept of “appearance”).
Every appearance is the appearance of reality.
Reality appears to us in a certain form, and no form of perception can
be treated as either invalid or privileged.
The form-object distinction puts an end to all these confusions and soph-
istries. It enables one to reject the hopeless attempt to compare appearance to
reality — i.e., to compare form to object. One can compare forms of percep-
tion to each other, as in comparing how a house looks from here to how it
looks from over there. Or, one can compare objects of perception to each
other, as in comparing one house to another. What makes no sense, however,
is trying to compare a form of perception to the object of perception, as if
we could wonder whether the house’s appearance looks like the house, or
whether sugar is really as sweet as it tastes.
Of course, it makes sense to wonder whether the house is as big as it
looks (“Would that house measure as large as I would estimate it to measure,
judging by how it looks from here?”). As noted earlier, that can be a valid
 A ppearance and R ealit y 95

question to raise, because it concerns the relation of perceptual data to con-

ceptual judgment, not the relation of the form of perception to the object of
perception — i.e., not the nonsensical comparison of “things as they appear”
to “things as they really are.”
The root of the notion that we can compare form and object is the Repre-
sentationalist model of the little man in the head perceiving images, as in the
drawing on page 89. That model sets us up to ask whether or not the internal
tree looks like the external one. But the model is utterly wrong and the ques-
tion is nonsensical; it amounts to asking “Does the look of the tree look like
the tree?” But looks don’t have looks. Appearances don’t have appearances.
A look or an appearance is a form of perception, and we don’t have a form of
perceiving our forms of perceiving.
A form of perception is how a thing looks, and that is not something that
can either match or mismatch what we are looking at. Analogously, the
manner in which a thermometer responds to a temperature (by its mercury-
level) is not something that can either match or mismatch a temperature.
The relation here is not similarity-difference but cause-effect. The senses do
not copy the object, they respond to it. Like physical instruments, the senses
respond to the operative causes in the only way they can; how they respond is
metaphysically given, and thus not subject to judgment.43
Perception is an active process of detecting and grasping the identity
of objects, an automatic, inerrant process of awareness. Questions about
perception’s “fidelity” in “corresponding” to the object stem from the false
assumption that perception copies something. As Gibson states:

The information for the perception of an object is not its image.

The information in light to specify something does not have
to resemble it, or copy it, or be a simulacrum or even an exact
projection. Nothing is copied in the light to the eye of an observer,
not the shape of a thing, not the surface of it, not its substance,
not its color, and certainly not its motion. But all these things are
specified in the light. [Gibson 1986, 304–305]

The Naïve Realist says: “Rubies look red because they are red.” The Repre­
sentationalist says: “Rubies look red but maybe they are really some other color.”

43 To gain value from a thermometer requires calibrating it and establishing its scale; like-
wise to gain the cognitive value of the senses requires a similar process of interpreting
and integrating the material they provide. But in neither case is there such a thing as
“checking for validity.” Thus, the slogan here could be: Calibration sí, validation no.
 H OW W E K N OW 96

The proper view of perception holds: “Rubies looks red because the human
visual system produces that form of response to rubies under standard light-
ing and viewing conditions.”
I gave drawings that illustrate the Naïve Realist and Representationalist
models of perception. But I cannot do the equivalent for the proper theory of
perception, because the proper theory relies on the form-object distinction
and there is no way to separate form and object in a drawing. I cannot draw
a tree apart from how it looks (form), and I cannot draw the form in which
the tree is seen except by drawing the tree.
I could, of course, draw a man who is facing a tree with his eyes open.
But I cannot draw his form of seeing the tree. The form in which he sees the
tree is not a second object; it is not an image inside his head. But to draw is
to draw objects, objects in a form. Perspective is an example of a perceptual
form; I cannot make a drawing of perspective; I can only draw things in per-
spective. The way things appear is not an object and thus cannot be depicted
apart from objects.
Perceptual awareness is of some object, by some means, and in some form.
Perceptual awareness is not image-making, image-viewing, hypothesizing,
or judging: it is perceptual awareness. And perceptual awareness is not mis-
perceiving, dreaming or hallucinating: it is perceptual awareness.
 CONCEPT-
3
FORMATION

L ike perception, reason has a biolo gical function.

Man needs knowledge, knowledge that extends beyond the perceptual.
He needs to know how to obtain food and shelter, forge tools, and satisfy all
the needs of his survival, material and mental. Reason enables man to gain
the conceptual knowledge his survival requires.

For man, the basic means of survival is reason. Man cannot

survive, as animals do, by the guidance of mere percepts.
A sensation of hunger will tell him that he needs food (if he has
learned to identify it as “hunger”), but it will not tell him how
to obtain his food and it will not tell him what food is good for
him or poisonous. He cannot provide for his simplest physical
needs without a process of thought. He needs a process of thought
to discover how to plant and grow his food or how to make
weapons for hunting. His percepts might lead him to a cave,
if one is available — but to build the simplest shelter, he needs
 H OW W E K N OW • 3 : C oncept- formation 98

a process of thought. No percepts and no “instincts” will tell

him how to light a fire, how to weave cloth, how to forge tools,
how to make a wheel, how to make an airplane, how to perform
an appendectomy, how to produce an electric light bulb or an
electronic tube or a cyclotron or a box of matches. Yet his life
depends on such knowledge . . . [VOS, 22–23]

But the biological function of reason has not been widely understood
or appreciated. The long Platonic tradition, still gripping the intellectual
world, holds that reason is concerned with “higher,” “spiritual” matters,
not the allegedly “materialistic” issues of life on this earth.
Platonism pits the perceptual and conceptual levels against each other,
opposing spirit to matter, mind to body, intellectuality to worldly concerns,
theory to practice. For Platonists, man’s consciousness is split between mind
and body — between a faculty directed toward a “World of Forms” (heaven, in
effect) and a faculty dealing with this earth. Man is caught in an internal war
between his “higher” and “lower” nature; Christianity took over and intensified
this Platonic dichotomy, damning outright all things earthly.
The mind-body dichotomy is not only indefensible philosophically,
in the light of modern biology, it represents an archaic embarrassment.
We know that man evolved from pre-conceptual primates, and that our
present intellectual capacity developed gradually, as the brain evolved.
Man’s conceptual faculty arises from the nature of his brain, and the human
brain is an elaboration of the primate brain. The conceptual faculty, reason,
is an enhancement of perceptual consciousness, not an alien element
wrenching man’s soul away from perceptual concretes.
After the work of Darwin, Mendel, Fisher, Watson and Crick, we know
with full certainty that man’s conceptual faculty evolved due to natural
selection — which means: man’s conceptual faculty has survival value.
The preeminent evolutionary biologist Theodosius Dobzhansky puts
it eloquently:

The adaptive value of forethought or foresight is too evident

to need demonstration. It has raised man to the status of the lord
of creation. [Dobzhansky, 1967, 468]

The root of the mind-body dichotomy is a wrong answer to the crucial

question of epistemology, a question that remains even after rejecting any
 C O N C E P T- F O R M AT I O N 99

“higher reality.” That question is: by what means are abstractions connected
to perceptual concretes?
There are no directly perceivable referents for higher abstractions.
For instance, when we say “Honesty is a virtue,” the words “honesty” and
“virtue” do not designate perceptual objects — they stand for abstract
relationships. One cannot commune with Plato’s Form of Virtue in some
world of pure abstractions to learn or validate this moral principle. What then
is the actual relation of abstractions, concepts, to the concretes we perceive
by our senses? And, more basically, by what specific means does one, in late
infancy, advance from percepts to concepts? How are concepts formed from
perceptual concretes? And how do we go from those concepts to
more abstract ones, then to conceptual judgments, then to principles?
In other words: “How do we know?” — when the knowing is conceptual?
The validity of concepts and reason is not in doubt. If the conceptual level
were not valid, you could not be reading this book; you could not know
anything beyond what you could, like an animal, see, smell, hear, taste,
and touch. In an age of computers, cell phones, and space exploration,
we are not waiting to find out whether or not reason works. Rather, the issue
is discovering how reason works. And the purpose of discovering that is to
be able to define standards of cognition, standards to use in steering and
evaluating our mental processes.
The operation of our perceptual equipment is automatic and infallible;
the exercise of the faculty of reason is not automatic but volitional,
and therefore can be misused, leading to error. We consciously control how
we think, making constant choices regarding what to think about, how to
proceed, what counts as evidence, and what constitutes sufficient evidence
to be certain of a conclusion.

Man is neither infallible nor omniscient; if he were, a discipline

such as epistemology — the theory of knowledge — would not
be necessary nor possible: his knowledge would be a­ utomatic,
unquestionable and total. But such is not man’s nature.
Man is a being of volitional consciousness: beyond the level of
percepts — a level inadequate to the cognitive requirements of
his survival — man has to acquire knowledge by his own effort,
which he may exercise or not, and by a process of reason, which
he may apply correctly or not. Nature gives him no auto-
matic guarantee of his mental efficacy; he is capable of error,
of evasion, of psychological distortion. He needs a method
 H OW W E K N OW • 3 : C oncept- formation 100

of cognition, which he himself has to discover: he must discover

how to use his rational faculty, how to validate his conclusions,
how to distinguish truth from falsehood, how to set the criteria
of what he may accept as knowledge. [ITOE, 78–79]

A man’s survival, well-being, and happiness depend on his knowing

what to do (and how to do it). He needs to know that fire can be tamed
and that striking together two rocks of a certain kind can make sparks that
will start a fire. He needs to know that seeds can be planted, that a keystone
will hold an arch together, that 2 + 2 = 4, that a certain stock is (or is not)
a good investment, that the ratio of HDL to LDL cholesterol is important
for his vascular health, that a certain person is “Mr. (or Miss) Right,”
that political candidate X is most likely to preserve his freedom.
Each item of knowledge he needs represents a thought, a proposition.
(“A keystone will hold an arch together.”) Propositions are composed
of concepts (e.g., “keystone,” “hold,” “arch”). The mere fact that a proposition
pops into a man’s mind (“XYZ Co. stock is a good investment”) — even when
it “feels right” — is not the least assurance that it is true.
The truth of a proposition depends on the validity and precision of the
concepts of which it is composed. Will candidate X support freedom?
That depends on what the concept “freedom” means. Is abortion murder?
That, in turn, depends on what “murder,” “human,” and “life” mean. Is it
immoral to be selfish? That depends on what “moral” and “selfish” mean.
Are there objective standards for determining the proper meaning
of the terms we use? Such standards would depend upon knowing what the
cognitive purpose of concepts is, what role they play in cognition.
Concepts are the tools of thought; if the tools are useless, malformed,
or otherwise defective, the thought cannot achieve its goal: knowledge
of reality. The validity of one’s thinking depends upon the validity of the
concepts one uses.
To gain proper guidance in conceptualization, we must first know
what concepts are. And to know that, we must understand how concepts
are formed.
 T hree T heories of C oncept- F ormation 101

Three Theories of Concept-Formation

What exactly does one do, starting at about age one, to advance from
sensory perception to the possession and subsequent use of concepts?
The child certainly can direct his senses to look at this or that object; he can
selectively focus his attention on a given aspect of a thing — its size or its
shape. What happens next?
Concepts are abstract, but the world we perceive is concrete. How do
we come to have abstract ideas? What is the relationship of abstractions to
perceptual concretes?
A concept is held by means of a word — e.g., “man,” “furniture,” “justice.”
The simplest words are proper names. The word “Tom” names Tom. Likewise,
the earliest theorists assumed, when we say “Tom is a man,” we must be
using “man” as a name for some one, perceivable existent: man, or manness.
Manness is the “one in the many,” as the Ancient Greeks put it; manness is the one
“universal” found in Tom, then in Dick, then in Harry. We can call each
of them a “man,” they held, because “man” refers to this one “universal”:
manness. The unstated assumption was: for a concept to be valid, it must
be like a percept: an awareness of a perceivable existent, out there in reality,
independent of consciousness.

The Realist Theory

This assumption is the defining characteristic of the theory of concepts
known (somewhat misleadingly) as “Conceptual Realism,” which is usually
shortened to “Realism.” According to Realism, a concept is a term that desig-
nates a metaphysical universal: a special kind of non-specific element present
in all the members of a class, an element that is grasped directly by some sort
of non-sensory “intuition” or “insight.”
Realism began with Plato. Plato went far beyond the mere acceptance of
metaphysical universals; he concocted another, “higher,” reality for universals
to inhabit — his “World of Forms.” In that unperceivable, transcendent realm,
there is one perfect, unchanging Form corresponding to each of our concepts.
There is the “Form of Man,” the “Form of Triangle,” the “Form of Justice,”
etc. Each Form is supposed to be a non-material entity that is the perfect
exemplar of the corresponding concept. The differing, changing, imperfect
concretes in this world are rendered intelligible only to the extent of their
imperfect “reflection” of the pure Forms. Tom, Dick, and Harry, for example,
 H OW W E K N OW • 3 : C oncept- formation 102

count as men only insofar as they, however imperfectly, reflect the eternal
Form of Man. (Plato regards Forms as more real than the concrete particulars
of the physical world, calling the World of Forms “the really real reality.”)
The Sophist Antisthenes is reported to have said, “I have seen many men,
but never have I seen Man.” Plato responded: Antisthenes is not looking with
the eyes of his intellect. The Form of Man, Plato maintained, is available
to man through intellectual vision — a kind of (ultimately) mystical insight
into the World of Forms wherein Man resides. The direct object of concep-
tual awareness, Plato held, is a Form in another dimension.
Aristotle, though he was Plato’s student, rejected this whole metaphysics.
Aristotle recognized that there is only one reality, the world of concrete enti-
ties, the world that we perceive. But in his theory of concepts, Aristotle did
not fully break free from Platonic assumptions. (Here, I give the traditional
interpretation of Aristotle, but gratefully acknowledge recent scholarship
indicating that Aristotle’s actual view of concepts is quite close to the one
I defend in this chapter. [Lennox, 2000, chs. 1, 2, 7, 8] )
Aristotelians (if not Aristotle himself) hold a theory known as “Moderate
Realism.” Moderate Realists count as Realists because they hold that abstrac-
tions refer to metaphysical universals; the theory is “moderate” in holding
that these universals exist as aspects of perceptual concretes, not as separate
entities dwelling in another world.
In effect, Moderate Realism shatters the Platonic Form and puts a fragment
of it inside each concrete. Rejecting Plato’s separation of Forms and particulars,
the Moderate Realists hold that concretes in this world are a compound of
particular and universal elements, of matter and form (with a small “f ”).
Tom, Dick, and Harry each have their own “manness” (which is held to be
natural, not supernatural). “Manness” is what is taken to be the same in each
of them, which is why they can all be called “men.” Men differ in their matter:
I have my body, which is somewhat different from yours. Considered qua
man — i.e., in terms of their manness — all men are identical. But because
that manness is found in different “stuff,” we are many and differing: one man
is short, another tall; one has pale skin, another dark skin. The differences are
differences in matter.
Centuries later, John Locke took the further step of jettisoning the form-
matter apparatus, treating universals as attributes of things. For Locke, there
is no such thing as either “manness” or “essence.” Rather, men all have the
rational faculty and other attributes in common. In these attributes, men
“agree” — i.e., are the same — however much their other attributes may differ.
 T hree T heories of C oncept- F ormation 103

In all its variants, Realism holds that concepts reflect an awareness

of a metaphysical universal — i.e., a non-specific existent found in the
external world. Plato’s version of Realism holds that by turning away
from the physical senses, the “mind’s eye” can apprehend the universal
existing in a separate dimension. The Moderate Realists hold that we mentally
separate the universal from the concrete by narrowing our focus to ignore
what is specific and varying, in order to zero in on the universal element
lying within that concrete. Here, for instance, is a Thomist text’s statement of
the process, with the author’s own emphasis:

We confine our attention to certain elements of the [single

concrete] under consideration, shutting out all the other ele-
ments, and stripping them of all particularizing determinations.
Abstraction consists precisely in this function and in nothing
else. . . . [Abstraction is] a special spiritual power within me, which
“shines upon the sense data, and makes them capable and ready
to produce a knowledge in which reality is deprived of all its
concrete and individual features.” [De Wulf, 13, 25]

Locke’s version of Moderate Realism tries to avoid positing non-specific

universals, but implies them nonetheless:

. . . the same colour being observed to-day in chalk or snow,

which the mind yesterday received from milk, it considers that
appearance alone, makes it a representative of all of that kind;
and having given it the name whiteness, it by that sound signifies
the same quality wheresoever to be imagined or met with; and
thus universals, whether ideas or terms, are made. [Locke, II, XI, 9]

Here, “whiteness” is the universal. It has to be a non-specific attribute,

since it is found in the slightly different shades of white characterizing
chalk, snow, and milk. With regard to the concept “man,” Locke implies that
man’s rational faculty is a non-specific attribute, even though men clearly
have differences in their intelligence, their degree of rationality, the things
they think about, the language in which they think, etc.
On any Moderate Realist theory, we grasp the non-specific attribute
by abstraction, which is conceived as a subtractive process, as a process of
disregarding differences. Locke, for instance, writes that “the mind, to make
 H OW W E K N OW • 3 : C oncept- formation 104

general ideas comprehending several particulars, leaves out . . . those qualities

that distinguish them . . .” [Locke, III, VI, 32 (my emphasis)]
In summary, the three tenets of Moderate Realism are:

1. Non-specific properties exist.

2. Abstraction is subtraction.
3. Concepts are an awareness of the non-specific property.

Moderate Realism takes from Plato its basic assumption: concepts name
a pre-existent universal, which is a “one” in each of the “many.” Moderate
Realists diverge from Plato only in regard to where that universal is to be
found: not in a world of Forms but locked inside the things of this world.
For instance, “blueness” is held to be something lying beneath the various
shades of blue; we can grasp blueness in isolation by mentally stripping away
what varies in the different shades.

The Failure of Realism

The Realist theory fails at the outset, at its first and fundamental premise:
the assumption that non-specific properties exist. In fact, nothing outside
the mind is general, “blurry,” non-specific. There are no universals “out
there” in the external world. Locke’s example of “whiteness” disguises this
problem, since chalk, snow, and milk differ little in color. But white and
black are special cases; if we consider any other color, such as blue, the shade
differences are dramatic. Where is the one universal “blueness” in a blue-
berry, the sky, and a robin’s egg? The difficulty becomes more obvious as we
get more abstract — e.g., when we form the more abstract concept “color.”
All the colors that we perceive are specific, such as this shade of red or
that shade of blue. There is no “coloredness” to be perceived or “intuited.”
Likewise, “shape” is always found as this shape or that different shape, not as
some non-specific “shape-as-such.”
The same is true for concepts of entities. When we focus perceptual
attention on an entity, what we see is always something specific: this tree,
this table, this dog — which differ from that tree, table, or dog, respectively.
How does one reach the concept “tree,” “table,” and “dog,” each of which refers
to an unlimited number of differing individual entities?
To be sure, the differing concretes are in some sense “the same when
viewed abstractly.” All dogs are dogs, however much they differ in particular
 T hree T heories of C oncept- F ormation 105

respects. All shades of blue are still shades of blue. But what does “the same
when viewed abstractly” mean? That is a serious and profound question that
cannot be bypassed by a hand-waving reference to “having something in
common” or “finding the same whiteness” in chalk and milk.
The “problem of universals” is actually the question of the basis of
concepts. The issue can be stated as follows. The concretes to which a given
concept refers are neither identical to each other nor possessed of any non-
specific properties. What then warrants our treating them as the same,
as being interchangeable units, when viewed abstractly?
As Rand puts the question:

To exemplify the issue as it is usually presented: When we refer

to three persons as “men,” what do we designate by that term?
The three persons are three individuals who differ in every
particular respect and may not possess a single identical
characteristic (not even their fingerprints). If you list all their
particular characteristics, you will not find one representing
“manness.” Where is the “manness” in men? What, in reality,
corresponds to the concept “man” in our mind? [ITOE, 2]

The Realist notion of a “universal” — an existent that is non-specific —

represents the improper separation of two axiomatic concepts: “existence”
and “identity.” To be is to be something, to have a specific identity. To be noth-
ing in particular is to be nothing at all — i.e., not to be. In Rand’s statement:
“Existence is Identity.” [AS, 1016] But the Realist theory posits the existence
of non-specific, generalized, “blurry” universals. (The only alternative for
a Realist is to maintain, in defiance of plain fact, that the concretes referred
to by a concept have to be identical in some respect, as if all shades of blue
were identical, all lengths were identical, etc.)
The Realists’ separation of existence and identity reaches its clearest
expression in Locke. His concept of a “substratum” in which a thing’s quali-
ties supposedly inhere is a “something I know not what” — i.e., an existent
without any identity (since the identity pertains to the qualities not to the
substratum). But a thing is its attributes.
It is an error to take attributes to inhere in something else, like pins in
a pincushion. Each attribute is an aspect of the whole, separated mentally
but not capable of physical separation from the whole. Plato, too, separates
the concepts of existence and identity, as in his concept of formless “space,”
 H OW W E K N OW • 3 : C oncept- formation 106

in the Timaeus. Unfortunately, the Aristotelians are guilty of the same error,
which surfaces in their position that “prime matter” is identity-less, with the
identity of any concrete belonging to it only in virtue of its form.
The Moderate Realists’ notion of abstraction as a subtractive process
follows from their metaphysics. Abstraction for them is simply selective
perceptual attention to the universal element in concretes — a stripping
away mentally of whatever differs among them. But since every feature may
differ, the abstraction-as-subtraction process leaves us with . . . nothing.
The concept of “number” is the purest illustration of Realism’s conundrum:
6 and 17 differ in quantity; if we abstract away the differing quantity of each,
we are left with nothing.
Because Realism fails in its analysis of the basis of concepts, it collapses
into subjectivism: Realism provides no objective way to resolve disputes.
Is a fetus a human being? On Realist premises, there is no way to settle
this question because we don’t know how to determine if “manness” is
present in the fetus. We don’t know which differences to subtract away
and which to leave in. Should we include or strip away such differences
as: whether the entity is inside the mother’s body or outside it, whether it
breathes or doesn’t, whether it eats food or absorbs nutriments from the
mother’s bloodstream? Some people “intuit” one “essence,” others “intuit” the
opposite one.
Realists are deprived of the means of settling such disputes because their
theory comes down to: either you “intuit” the universal or you don’t — which
turns concepts into subjective “constructs.”
Consider now a second way the Realist theory lands in subjectivism:
of the many features characterizing concretes, which ones should we
use to form our concepts? The Realist theory can have no answer to this
crucial question. Realism implies that there is a nature-given set of criteria to
use in forming concepts, as if it were self-evident how we should divide up
and classify the things we perceive. There is a certain plausibility to this idea
on the lower levels of abstraction, for such concepts as “table” and “triangle.”
But more abstract concepts, like “momentum,” “freedom,” and “murder” are
more obviously made by us. What are the proper characteristics for us to use
in forming such abstractions? No guidance can be given on Realist prem-
ises. Realists can only say “Just look with your mind’s eye” — advice that is
manifestly hopeless.
Realist theories reify abstractions: they take the generality that exists only
in man’s mind and make it into a “universal” existing in the external world,
 T hree T heories of C oncept- F ormation 107

either in a separate dimension (Plato) or as non-specific characterist­ics in the

concretes of this world (Aristotle and Locke). Concepts are indeed general:
they do apply universally to a range of concretes. But in reading that mental
generality back into things, Realists are endorsing that epitome of unrealism:
the primacy of consciousness. As Rand observes, the Realists are “assuming
that reality must conform to the content of consciousness, not the other way
around — on the premise that the presence of any notion in man’s mind
proves the existence of a corresponding referent in reality.” [ITOE, 53]
Realists hold that reality contains “universals,” and that these can be
grasped passively by the intellect, as if they were perceptually given.
This, as we shall see, totally misconstrues what concepts are and do.

The Nominalist Pseudo-Theory

Realism’s problems spawned the Nominalist reaction. Accepting the Realist
claim that concepts would require pre-packaged universals, Nominalists bite
the bullet and conclude that concepts have no objective basis, that concepts
are nothing but words. Ludwig Wittgenstein, for example, writes: “To say that
we use the word ‘blue’ to mean ‘what all these shades of color have in com-
mon’ by itself says nothing more than that we use the word ‘blue’ in all these
cases.” [Wittgenstein 1958, II, 3]44
This is Extreme Nominalism, a truly lunatic theory. It flatly contradicts
every aspect of our actual usage of concepts. For instance, Extreme Nominal-
ism implies that it is an inexplicable miracle when two people independently
apply the word “blue” to the next blueberry they see. Since the blueberry and
the other things previously called “blue” have nothing in common, accord-
ing to Wittgenstein, the two people should be equally prepared, on the
Nominalist theory, to say “yellow” or “pink” — or, for that matter, “railroad.”
Just as there is Moderate Realism, so there is Moderate Nominalism.
The Moderate Nominalists hold that the referents of a given concept, though
they share no “universal,” are still similar, and that their similarity makes
it “convenient” to refer to them by the same word. Moderate Nominalism,

44 Some would argue that Wittgenstein here is not committing himself to this position,
merely proposing it for consideration, or to point out the vacuity of the phrase “have
in common.” That point of historical interpretation is not important for my purposes:
whether or not Wittgenstein finally endorses this extreme position, it expresses one
possible response to the failure of Realism.
 H OW W E K N OW • 3 : C oncept- formation 108

however, treats similarity as an unanalyzable, primary fact. Things just resemble

each other, Nominalists say, and that’s the end of the matter.45
The failure to analyze what similarity (or “resemblance”) consists of makes
it impossible to defend the objective validity of concepts and lands in rank
subjectivism. Nominalists, of any variety, are unable to answer the two
crucial questions: 1) How similar? and 2) Similar how? That is, 1) how simi-
lar must concretes be to warrant inclusion in the same class? and 2) which
similarities should we use in forming our conceptual classifications?
The Nominalists’ inability to answer these questions does not trouble them.
How similar do the conceptualized concretes have to be? No one can say,
they tell us, and then they trot out examples of “borderline cases” to argue
that the boundaries of a concept’s application are set arbitrarily. How white
does a thing have to be to be called “white”? Does cream have “whiteness”?
Is a stool a kind of chair without a back, or does it have its own “essence”?
In the continuous development of the human fetus, when does “rationality”
or “manness” appear, and why then? The Nominalists complain that “nature
doesn’t tell us” the answers to these questions, and conclude that the bound-
aries of concepts are set arbitrarily, by blind tradition or by “stipulation.”
When we raise the question: “Similar, how?” Nominalism’s utter collapse
into subjectivism becomes dramatically evident. The arch-example here is
Wittgenstein’s theory of “family resemblances.” Take a hypothetical family:
the Kowalskis. Bill and Bob Kowalski have similar noses. Bob and Jane
Kowalski have noses that look rather different but have very similar chins.
Jane and Lewis Kowalski have different noses and chins, but have similar
eyes. And so on. No one characteristic runs through all members of the
Kowalski family. The same is true of all concepts, Wittgenstein claims:
there is no one respect in which the things subsumed are even similar.
Accordingly, Wittgenstein could have no objection to the following example.
A pumpkin is similar to a basketball (in shape); a basketball is similar to
a hockey puck (in purpose); a hockey puck is similar to a hockey stick
(in the sport involved); a hockey stick is similar to a broom (in how one moves it).
We can go on indefinitely. This conceptual stew is what one ends up with on
the Nominalist theory.

45 Consider this passage from The Stanford Encyclopedia of Philosophy entry on Nominalism:
“A final version of Nominalism is Resemblance Nominalism. According to this theory . . .
what makes square things square is that they resemble one another, and so what makes
something square is that it resembles the square things. Resemblance is fundamental
and primitive . . . .”
 T hree T heories of C oncept- F ormation 109

As Ayn Rand remarks, “Wittgenstein’s theory that a concept refers to

a conglomeration of things vaguely tied together by a ‘family resemblance’
is a perfect description of the state of a mind out of focus.” [ITOE, 78]
The result is that, for the Nominalist, how concepts are to be formed and
used becomes a matter of feeling. This subjectivism has a devastating impact
upon higher-level abstractions. One man feels that aborting a fetus and
killing an innocent adult “resemble” each other, so he calls both “murder”;
another man feels differently and maintains that abortion is a right. One man
feels that a free lunch and free speech are similar, another man disagrees.
One man feels that Al Capone and John D. Rockefeller were similar enough
to be classified together, as being “rapacious” and “predatory.” But are they?
How similar and in what way? Nominalism not only has no answer, it regards
all such questions as in principle unanswerable.
Thus, the Nominalist theory deprives man of objective guidance in the
crucial aspect of his life: how to form and use the concepts on which his
control over the course of his life depends.
The Realist theory lands in subjectivism by implication, but N ­ ominalists
defiantly embrace subjectivism as a matter of principle. This is ironic, because
Nominalism entails the rejection of all principles. For Nominalism, no state-
ment of the form “All As are B” can be known to be true. After all, each A is
unique, and so is each B. Under Nominalism, no generalization (induction)
is valid. If there is no way to treat the concretes subsumed under a concept as
interchangeable units, there can be no reasoning from one case to any other;
we must face each new concrete in a cognitive vacuum. A scientist might
spend a decade studying the characteristics, properties, and capacities of
John, but then when he turns to consider Paul, he must begin all over again.
None of his knowledge of John is necessarily applicable to Paul because, to
paraphrase Wittgenstein, to say that John and Paul are both men is to say
nothing more than that we apply the word “man” to each. Or, for Moderate
Nominalists, though John and Paul resemble each other, we have no reason
to think this resemblance is sufficient to warrant applying to Paul anything
learned from studying John — just as we have no warrant to apply to a broom
things learned from studying pumpkins.
Stepping back to get an overview, we can state the problem of concepts
as follows. The concretes to which a given concept refers are similar but not
identical. The blue of the sky is similar but not identical to the blue of a blue-
berry. A beagle and a collie are similar in some way, though not identical,
and this similarity is what enables us to classify them as dogs. The question
then is: What is similarity? How does the fact that observed concretes are
 H OW W E K N OW • 3 : C oncept- formation 110

similar warrant our treating them as the same, as interchangeable units,

qua referents of a given concept?

The Objectivist Theory

Ayn Rand’s theory of similarity grounds her Objectivist theory of concepts.
She defines similarity as: “the relationship between two or more existents
which possess the same characteristic(s), but in different measure or degree.”
[ITOE, 13]
Things that are similar differ quantitatively. The blue of a blueberry
is not identical to the blue of the sky, but the two differ quantitatively,
in measurable ways. A given blueberry is a darker blue than the sky; the hue
of the blueberry has more red in it, shading towards purple; the sky’s blue
is brighter than the blueberry’s. These are differences in degree along the
three measurable axes: hue, saturation, and brightness. Modern computers
usually provide a color-setting dialog box that uses numbers from 0 to 255
to specify the setting of each of these three parameters. Any of the colors that
we can see can be specified by a trio of these three numbers. (On my moni-
tor, blueberry blue is approximately 139, 142, 74; sky blue is 139, 200, 160.)
Color differences are a matter of measurements.
Consider now the case of similar entities. The similarity of a particular
beagle to a particular collie is more complex, but still one of measurable
quantity. The beagle is smaller and stouter; the beagle’s hair is shorter and
straighter than the collie’s; the collie’s nose is longer and more tapering;
the beagle’s nose is shorter and blunter; the collie’s bark is lower-pitched;
etc. There is no non-specific “dogginess” lodged in the beagle and the col-
lie. They differ in every respect, but the differences are in how much of each
characteristic — size, straightness of hair, ratio of length to width of nose
— they possess. The differences are differences in the measurements
of commensurable characteristics.
In contrast, a very young child beginning to form concepts would not
perceive a pig and a collie to be similar. Why not, if similarity is an issue of
quantitative differences? After all, the pig’s differences from the collie are also
measurable — the pig is fatter, pinker, with a measurably different shape, etc.
The answer to this question lies in a cognitive process neglected by
traditional theorists: differentiation. Similarity is inherently perceived against
a background of difference. As I have stressed, consciousness is a difference-
detector. When a naïve, pre-conceptual child attends to two items, it is their
 T hree T heories of C oncept- F ormation 111

differences, not their similarities, that will be prominent. Although a beagle

and a collie are similar, putting them side by side serves to focus attention
on their differences (for a pre-conceptual child). But sensitivity to differ-
ence can be turned to advantage here. When the child observes a beagle,
a collie, and a pig, the huge differences between the pig and the dogs
leap to the foreground of awareness, making the two dogs appear similar.
The pig appears to be different in kind from the two dogs, while the dogs
appear to differ from each other only in degree — i.e., similar in contrast
to the pig. Even though a conceptually advanced adult observer could say
that all three animals have commensurable characteristics (color, shape,
weight), for the beginning conceptualizer, the difference that dominates his
attention is the difference between the two dogs and the pig.
In the same manner, two shades of blue put side by side will simply be
perceived as different, but when a contrasting color, such as green, is added
to the comparison, the two shades of blue appear similar, appear to go
together, as opposed to the particular green that they are being contrasted to.
Hue is the commensurable characteristic possessed by all three colors, so
hue serves as the background that pushes into the foreground the marked
difference between the blues and the green.
The grasp of similarity requires a minimum of three concretes having
a commensurable characteristic(s): two whose measurements differ slightly
and one that differs greatly in measurement from both.
The arch example is location. Is the Empire State Building near to or far
from the Chrysler Building? That depends — it depends on what we are
comparing them to. If the comparison is to a location across the street from
the Empire State Building, the Chrysler Building is far. If the comparison
is to the Sears Tower in Chicago, the two Manhattan buildings are near to
each other. And if the comparison is to the location of a mountain on Mars,
all three buildings are near to each other.
This can be represented graphically:

A B C

Considered by themselves, A and B are in different places. Considered in

contrast to C, A and B, though not in the identical place, are seen as falling in
the sameAgeneral region
B
— i.e., at the left end of the line. C

A D B C

A B C S
 H OW W E K N OW • 3 : C oncept- formation 112

“Near” and “far” are the stand-ins for “similar” and “different.” Similarity is
measurement-proximity.
“Proximity” is a relative term, depending on a contrast with something
that is more distant, which we can call “the foil.” Similarity is thus contextual,
a matter of relative proximity of measurements in contrast to the relatively
distant measurement of a foil. In such a set-up, the bigger difference swamps
the smaller difference, making the smaller difference appear as similarity.
What is experienced as similarity is, at root, lesser difference.
For this mechanism of contrast to work, all three items, the two
similars and the foil, must be compared along the same axis. They must share
a commensurable characteristic, as in the diagram above, where hori­zontal
position is the commensurable characteristic. If the position of A and
B were to be considered in the same frame of awareness with, say, the smell
of a peach, no similarity would be established, as there is no commensu-
rable characteristic uniting those three items: no single unit of measurement
can be applied to them. They are not different but disparate. Differentiation
requires a commensurable characteristic to serve as the axis along which
the things can be compared.
Rand refers to this commensurable characteristic as the “Conceptual
Common Denominator” (CCD):

A commensurable characteristic (such as shape in the case of

tables, or hue in the case of colors) is an essential element
in the process of concept-formation. I shall designate it as
the “Conceptual Common Denominator” and define it as
“The characteristic(s) reducible to a unit of measurement,
by means of which man differentiates two or more existents from
other existents possessing it.” [ITOE, 15]

It may be helpful to connect Rand’s terminology with the traditional terms

“genus” and “differentia.” A genus is the wider category to which a given
concept’s referents belong. E.g., the genus of “triangle” is “polygon.” What
is the relation of “genus” to Rand’s “CCD”? The CCD is not a category but
a measurable characteristic, one possessed by all the things in the genus,
as “having a certain number of sides” is a characteristic of all polygons.
The differentia is the referents’ distinguishing characteristic, the one that
isolates them from all the other things within the genus. For triangle, the
 T hree T heories of C oncept- F ormation 113

differentia is “three-sided.” So, we form “triangle” by differentiating shapes

according to the number of their sides (CCD), triangles having three.
Though the CCD plays a crucial role in conceptualization, neither the
CCD nor its function is explicit in the mind of the beginning conceptual-
izer. The CCD’s role is precisely to serve as the unnoticed background — not
to call attention to itself but to something else: the difference between the
similars and the foil, whose measurements lie noticeably farther away on
the same CCD. When a child differentiates two blues from green, the CCD
is hue; but the child’s attention is drawn not to hue as such (the sameness
among all three items) but to blue vs. green (the difference in hue).
Thus, the CCD plays an indispensable role, but (at the beginning levels)
an unnoticed one. The CCD serves to draw attention to the difference
between the similars and foil, a difference which is so striking that it appears
as a difference in kind (“These are blue, that is green”). Analogously, I can
draw a figure on a blackboard with white chalk because the black slate con-
trasts with the white, yet what draws your attention is the white figure, not
the black slate which, though essential, serves only as background.
Later in a child’s development, he will grasp “color,” but he can do so
only by contrasting “color” with something further afield, say shape, within
a wider (implicit) category: visual attribute. On either level, one uses the
CCD, but only as background.
Now we are in a position to find the “one in the many.” Or, rather, to see
that the one in the many, the “universal,” is man-made, not given in nature,
yet not created subjectively. A concept classifies together concretes whose
measurements fall within the same category of measurements within the CCD.
To be perceived as similar, and be grouped together, concretes need
not have identical measurements in any respect, but they each must have
measurements that fall within the same range or category. “Blueness” does
not exist per se in the sky, the robin’s egg, or the blueberry. But, contra
Wittgenstein, “blueness” does name an observed fact about each: the fact
that each shade’s specific measurement falls within a certain range of
measurements within the CCD hue. Crucially, this range is graspable only by
differentiation from the very different measurements of one or more foils
whose own hue-measurements fall outside that range.
Although a child makes color comparisons by simply perceiving the
degree-differences (bluer, less blue), today we know the exact measure-
 H OW W E K N OW • 3 : C oncept- formation 114

ments responsible for hue differences: the wavelength of the reflected light.46
The blue range is: approximately 450–485 nanometers. The neighboring
green range is approximately 500–565 nanometers. These numbers repre­
sent real lengths: the actual, physical distance between successive peaks in
the light waves. A particular hue of, say, 462 nanometers in wavelength falls
squarely within the blue range, even though there is no non-specific “blue-
ness” lodged within it. And that specific hue, will be perceived as similar
to one of, say, 475 nanometers, when contrasted with one solidly within the
green range — say 535 nanometers. 462 and 475 seem close together when
contrasted with 535.
Consider another, quite different, example. Suppose a teacher gives
a multiple-choice exam to his class, with the usual result of a wide variation
in scores. The teacher has established the familiar grading ranges: A: 90–100,
B: 80–89, C: 70–79, and so on. If Frank scores 86 and gets a “B,” that does not
mean the teacher has “abstracted the B-ness” out of Frank’s score. There is no
such B-ness to be reached by discarding specific scores, as if selective atten-
tion would reveal some residue of B-ness. Rather, the teacher recognizes that
the completely specific score of 86 falls within the range of 80 to 89. Patri-
cia’s score of 83 is categorized as “essentially” similar to Frank’s, even though
it is not identical to Frank’s. For both students, the grade of “B” is reached by
omitting the specific measurements of their scores and retaining only the fact
that their scores fall somewhere within the B-range.
The same mechanism of classification lies behind all concept-formation.
All concepts are formed by implicitly or explicitly omitting the intra-range
measurements of similar concretes. The similarity results from the closeness
of their measurements within a CCD — close when compared to a foil whose
measurements lie markedly farther away along that same CCD.
The items classified together by a concept are the same in the follow-
ing respect: their measurements all fall within the same range or category.
E.g., the different shades of blue all fall within the same range on the
spectrum; the “blueness” is that range. Or, in terms of test grades, the B-range
of 80–89 is one thing, not many, and falling in that range is what is the same
about all the tests graded B.
In general, the so-called “universal” characteristic that makes a concrete
qualify as a member of class X is not some non-specific X-ness inside the
concrete, per Realism, but is the category of measurements, the category
that includes each concrete’s specific degree of X. Rand’s important summary

46 For simplicity, I am taking only monochromatic light here, to avoid the complexity
of how different wavelengths combine (cf. the trichromatic theory of color vision).
 T hree T heories of C oncept- F ormation 115

statement is: “The distinguishing characteristic(s) of a concept represents

a specified category of measurements within the ‘Conceptual Common
Denominator’ involved.” [ITOE, 15]
Each concrete possesses its own specific degree of the distinguishing
characteristic, which may be different in quantity from the degree to which
other similar existents possess it. There is no non-specific “blueness” in
things, only specific shades of blue. In the case of concepts of entities, such as
tables, the concretes share several characteristics that jointly distinguish them
from other kinds of entities, but each of these distinguishing characteristics
always exists in a specific amount. All tables, for example, have a horizontal
surface and supports. But the number of supports may vary from having
one pedestal to having many legs, and the geometrical measurements of the
shape of the tabletop may vary rather widely.
In all cases, however, the similar concretes possess varying degrees of the
distinguishing characteristic, and those degrees fall within the “specified
category of measurements” — which is why they appear as similar.
The mental process of grasping the distinguishing characteristic is called
by Rand “measurement-omission.” Measurement-omission is the core
of the Objectivist theory of concept-formation. Having identified that similar
concretes possess the same characteristic, but vary in their measurements,
Rand was able to identify the nature of abstraction. The process of abstraction
consists in interrelating the concretes in a certain way: one grasps the range
or category of measurements that embraces all their varying measurements.
Concepts, then, are formed by omitting measurements. But omitting
measurements is not a process of deletion or excision, as if we could
mentally strip away the specific measurement from the characteristic.
Rand is very clear about the difference between “measurement-omission”
and the narrowed, eliminative focus of the abstraction-as-subtraction view:

Bear firmly in mind that the term “measurements omitted” does

not mean, in this context, that measurements are regarded as non-
existent; it means that measurements exist, but are not specified.
That measurements must exist is an essential part of the process.
The principle is: the relevant measurements must exist in some
quantity, but may exist in any quantity. [ITOE, 12]

Measurement-omission does not consist in ignoring the specific and

varying measurements of concretes (à la Locke). When we omit the mea-
surements, we are not ignoring anything; we are grasping something more:
 H OW W E K N OW • 3 : C oncept- formation 116

the relationship among the measurements, the fact that the similar concretes
are “near” to each other, as contrasted with the “far away” measurements
of the foil.
When Rand says “That measurements must exist is an essential part of
the process,” the process she is referring to is the process of measurement-
omission. One uses the measurements to interrelate the similar concretes.
One needs to focus one’s attention on the specific measurements, in order
to see “where they are” — i.e., to grasp the category of measurements within
which the similar concretes fall. E.g., the child has to focus on the specific
shapes of particular tables (and of the foil, e.g., a chair) in order to perceive
the tables as being similar, and to establish the range of shape-variation.
The process of measurement-omission is one of grasping a segment of the
CCD — the A segment B representing the range embracing the similarC concretes.
Thus, measurement-omission is measurement-integration — the establishment
of a range or category of measurements.47 This is illustrated in the diagram
below, where part of the dotted line becomes a solid line.48

A B C

Although the diagram uses thin, sharp demarcation lines, in actual

concept-formation
A
A D BB such hyper-precision is rare; usually one C establishes
C
only a rough region, allowing for borderline cases, such as greenish-blue.
(Borderline cases pose a problem for the Realist theory of a “universal,” which
must be either present or absent, but not for the Objectivist idea of an axis
of continuously varying measurements.49
WhenAoneA later
B encounters
B C another similar concrete,SD, itC is seen to fall
within the same range:

A D B C

Thus, D is classified with A and B.

47 The termA “measurement-integration”

B C replaces the term I formerly Sused: “measurement-
inclusion” [Binswanger, 1989], which I now use in regard to propositions (see Chapter 5).
48 I owe to Gregory Salmieri the idea of integrating the points into a line. [Salmieri, 2006]
49 For more on borderline cases, see ITOE 72–74 and infra 222 and 238–39.
 U nit 117

Within the established range, the particular measurements differ only

quantitatively; the range accommodates those differences, because they are
much smaller than the pronounced difference each has with the foil, C.
Thus, to grasp that things “have the same distinguishing characte­ristic” is
to grasp that their measurements, though varying, all fall somewhere within
the same range along the CCD that they share with the foil.

Unit
Concepts are formed by treating existents as units. A “unit,” in Rand’s
definition, is “an existent regarded as a separate member of a group of two
or more similar members.” [ITOE, 6]
Since similarity is a quantitative relationship, the term “unit” from
mathematics is meant literally here. “Unit” means “one,” and to regard an
existent as one is to view it in relation to a group — e.g., as a (one) book.
The group is formed by mentally isolating things that are similar in some
way, even if only in location (“objects on my desk”). Each of the similar
existents is a “one” that can be used as a standard for counting or for measur-
ing degrees.50
When two things differ only in quantity, they can be readily integrated,
because a larger one is reducible to the same units as a smaller one, plus more
of the same units. For instance, a length of 3 inches is reducible to inch + inch
+ inch, and then a length of 4 inches is reducible to that plus another inch
— i.e., another one of the units that compose the 3-inch length. Thus, “the
more” is “the less” plus more of the less. (A child, before knowing of units
and numbers, can easily become aware of this fact by seeing, for example,
the change in length of a rubber band as it is stretched and released or by seeing
the changes in shape he imposes on his modeling clay: rounder, less round.)
Understanding concept-formation in terms of establishing quantitative
relationships is the opposite of the Realist idea that we form concepts by
staring at one concrete in the attempt to achieve an “insight” into some
non-specific feature of it. On the Objectivist theory, concepts are formed

50 Counting measures quantity. Counting consists in pairing items, one by one, with the
members of a given group whose specific quantity is taken as a standard — e.g., pairing
five cows with each of the five fingers on one hand. Later, numerals are introduced,
to provide a uniform, systematic set of standards. Each number stands for the quantity
of the group formed by the sequence of numerals terminating in that numeral. “3” stands
for the quantity of the sequence: 1–2–3, a group consisting of three numerals.
 H OW W E K N OW • 3 : C oncept- formation 118

by contrast and comparison among concretes — i.e., by the conceptualizer

adopting a wider focus, not a narrower one. It is not a subtractive process,
but an additive (integrative) one.
There is no “universal” present in a concrete, so looking harder or more
selectively at one concrete will not find it. To grasp the common character-
istic that isolates the similars from the foil requires shifting one’s attention
back and forth between a minimum of three concretes — the two similars
and the foil — then omitting the measurements of the similar concretes by
seeing their range or category of measurements within the (implicit) CCD.
Measurement-omission is done on the premise that “the relevant measure­
ments must exist in some quantity, but may exist in any quantity” within that
range or category of measurements.
In speaking of “measurements” I am referring to the subconscious
mechanics of the concept-forming process, not to any consciously performed,
explicit, process of measuring. A child beginning to conceptualize things
is, of course, incapable of explicit measurement. On the conscious level,
he is aware only of similarities and differences. But the objective basis
of those similarities and differences is the quantitative variation of a commen-
surable characteristic. Where this commensurable characteristic is absent
— i.e., where the items are disparate — no differentiation and no awareness
of similarity can occur. Rand’s example is: “No concept could be formed,
for instance, by attempting to distinguish long objects from green
objects.” [ITOE, 13] To make a differentiation, one needs an A and a non-A
— e.g., blue and non-blue, or long and non-long. The disparate must be
firmly distinguished from the different. (This distinction will be important
when we discuss concepts of characteristics in Chapter 4.)
To omit specific measurements is not to discard them, ignore them or
put them aside. It is to see them as some specific amount of a characteris-
tic which varies in quantity. Length “with measurements omitted” does not
mean a length which is neither three inches nor twenty inches nor any other
specific amount; it means a length which is either three inches or twenty
inches or some other specific amount.
The Realist theory of concepts says just the opposite. Locke, for example,
says that “triangle” is a concept of what is neither equilateral nor scalene,
“but all and none of these at once.” [Locke, IV, VII, 9] The Objectivist theory
holds that the concept of a triangle includes those that are either equilat-
eral, or isosceles, or scalene. The sides of a triangle must have some specific
relation to each other but may have any.
 M easurement 119

Hegel took the “neither/nor” view to its ultimate extreme. Quoting from
Hegel’s premier interpreter, W. T. Stace:

Being . . . is the highest possible abstraction. All character,

all determinations of any kind, have been abstracted from.
Hence being has no character and is utterly empty. Because being
is thus utterly empty, it is therefore equivalent to nothing. . . .
Being, then, is nothing. . . . [B]eing has no determinations what-
soever. Therefore, there can be no difference between being and
nothing. The thought of being and the thought of nothing are the
same, and pass into each other. [Stace, II, First Division, I, A-B]

This equation of existence and non-existence is a reductio ad absurdum of

the abstraction-as-subtraction idea. A­ bstraction as measurement-omission,
operating on the “some but any” principle, regards “being” or “existence” as
the widest and fullest concept, embracing all differences. To exist is to have
some but any determinate identity, to be either an entity or an attribute or an
action, etc. Abstraction as measurement-omission means that the cognitive
content of concepts is rich — and grows progressively richer as we form pro-
gressively higher abstractions (the subject of the next chapter).

Measurement
Since measurement is the core of the Objectivist theory, we need to look
further at what measurement is. Rand writes:

Measurement is the identification of a relationship — a quantitative

relationship established by means of a standard that serves as
a unit. Entities (and their actions) are measured by their attributes
(length, weight, velocity, etc.) and the standard of measure-
ment is a concretely specified unit representing the appropriate
attribute. Thus, one measures length in inches, feet and miles
— weight in pounds — velocity by means of a given distance
traversed in a given time, etc. [ITOE, 7]

It is clear how simple, physical characteristics such as length, weight, and

color are measurable. Complex shapes, such as the shape of a leaf, are also
 H OW W E K N OW • 3 : C oncept- formation 120

measurable, by separately measuring length-width ratios, part by part. In

discussing why tables appear similar in shape, Rand notes:

. . . a given shape represents a certain category or set of geo-

metrical measurements. Shape is an attribute; differences of
shape — whether cubes, spheres, cones or any complex combina-
tions — are a matter of differing measurements; any shape can
be reduced to or expressed by a set of figures in terms of linear
measurement. . . . [A] vast part of higher mathematics, from
geometry on up, is devoted to the task of discovering methods
by which various shapes can be measured — complex methods
which consist of reducing the problem to the terms of a simple,
primitive method, the only one available to man in this field:
linear measurement. (Integral calculus, used to measure the area
of circles, is just one example.) [ITOE, 14]

At a given stage of scientific knowledge, some similarities may not be

reducible to exact, numerical measurement. Before the invention of the
thermometer, differences in hotness could not be reduced to numbers.
Even today, differences of hardness are often measured on an ordinal scale,
according to which substance can scratch which (as diamonds can scratch
steel, steel can scratch iron, iron can scratch lead) — all of which establish
rank order but not the numerical degree of hardness. To establish the rank
order of things in an ordinal measurement, one changes the unit for each
new comparison, always using the same characteristic for each of them
(e.g., hotness or hardness).51
Establishing numerical measurements, whether ordinal or cardinal,
is the most precise form of identifying quantitative relationships, but the
grasp of quantitative relationships is what goes on anytime one sees that
this is bigger or brighter than that. Approximate measurement is still mea-
surement in the sense relevant to concept-formation. One is still aware of
a commensurable characteristic, as opposed to the impossible attempt

51 A good example is the old classroom exercise of lining up students in order of height,
which was done ordinally, not by using a yardstick. The teacher would begin by having
two students stand next to each other, judge their heights, and have the taller student
stand in front of the shorter. Then, each remaining student would be positioned in the
line between the ones who were just shorter and just taller than he. This procedure meant
that, each student in the line, one at a time, served as the unit of height against which the
new student’s height was measured. Thus, the unit varied, but the characteristic did not.
 M easurement 121

to quantitatively compare, say, a rock and Saturday, which lack a commen-

surable characteristic.
An interesting case of measurement is that of measuring materials qua
materials, such as wood, copper, water. Obviously, one can measure the
attributes of the objects formed out of various materials, but in what sense
is the difference between copper and lead a difference in measurement?
On the sensory level, one uses the difference in perceptible qualities — the colors
differ, the densities differ, the hardness differs, etc. These are differences in
degree and are measured only approximately.
Modern chemistry, however, goes to a deeper level: copper and lead differ in
“atomic number.” Atomic number is a measurement. It refers to the number
of protons in the nucleus of the atom: copper has 29 protons, lead has 82.
And in the case of molecules, the differences are differences in measurement:
methane has one carbon and four hydrogens, ethane has two carbons and
6 hydrogens, etc.
Or, taking the case of food materials, bread and cake differ in the amount
of sugar. Differences in the preparation process are also specified quantita-
tively, as in a recipe: “Add ¼ pound butter to ½ cup of flour and one cup
of milk, stir slowly for two minutes, over a low heat . . .” Thus, even foods can
be specified in terms of measurements.
Throughout my discussion, I have often used the phrase “range
or category” of measurements. The reason for using two terms is that not
all CCDs vary continuously. Some CCDs come already digitized, in effect.
Triangles, for example, have exactly three sides; their number of sides
is not a range from 2.4 to 3.3. The CCD for “triangle” is: number of sides
of a polygon, but the distinguishing characteristic is not a range, but
a category: three. Or, length is extension in exactly one dimension, area in
exactly two, volume in exactly three. Length is “some but any” extension
in exactly one dimension. The CCD of “length” is: number of dimensions
of extension; the distinguishing characteristic is the category (not the
range): one.
Whether the distinguishing characteristic represents a range or a single
value within the CCD, if there are varying measurements within that range
or category (as there normally are), then they are omitted on the “some but
any” premise. For instance, length is extension in exactly one dimension,
but within that category, all the different lengths are included by a process of
measurement-omission.
 H OW W E K N OW • 3 : C oncept- formation 122

If a child considers a match, a pencil and a stick, he observes that

length is the attribute they have in common, but their specific
lengths differ. The difference is one of measurement. In order to
form the concept “length,” the child’s mind retains the attri-
bute and omits its particular measurements. Or, more precisely,
if the process were identified in words, it would consist of the
following: “Length must exist in some quantity, but may exist
in any quantity. I shall identify as ‘length’ that attribute of any
existent possessing it which can be quantitatively related to a unit
of length, without specifying the quantity.” [ITOE, 11]

In a few cases, even the distinguishing characteristic does not vary in

degree. Take the concept of “inch” for example. Inches do not vary in length.
The CCD for “inch” is length, but within that CCD, the distinguishing
characteristic is exactly this much of it. Or, if one defines the inch as one-
twelfth of a foot of length, it is still exactly that ratio, not a range. A standard
of measurement, like the inch, has to be an exactly determinate amount of
the characteristic in question.
To form the concept of a standard of measurement, such as the inch, one
must omit the measurements of other (non-distinguishing) characteristics.52
In regard to measuring length, there used to be a concrete physical object
that served as the standard meter: a platinum bar in Paris. Using that bar
as the standard of length required omitting the measurements of all its
other characteristics — for instance, its specific width, thickness, hardness,
material, location, etc. The meter-bar was given the particular width, thick-
ness, etc. that best suited it to be used as a standard of length.
The first concepts a child forms are concepts of entities, since enti-
ties are what perceptual awareness discriminates for him. As a summary
of her theory, here is Rand’s description of the process of forming the
concept “table.”

The child’s mind isolates two or more tables from other objects,
by focusing on their distinctive characteristic: their shape.
He observes that their shapes vary, but have one characteristic

52 The omission of non-distinguishing characteristics, however, is simpler: at the first stages

of forming a concept, one can simply disregard these differences, as when a child ignores
a table’s color. Sooner or later, however, the child will attend to these characteristics,
notice that they too vary in degree, and treat these differences as omitted measurements:
he realizes that a table must have some color but may have any.
 I ntegration 123

in common: a flat, level surface and support(s). He forms the

concept “table” by retaining that characteristic and omitting all
particular measurements, not only the measurements of the shape,
but of all the other characteristics of tables (many of which he is
not aware of at the time).
An adult definition of “table” would be: “A man-made object
consisting of a flat, level surface and support(s), intended to
support other, smaller objects.” Observe what is specified and
what is omitted in this definition: the distinctive characteristic
of the shape is specified and retained; the particular geometrical
measurements of the shape (whether the surface is square, round,
oblong or triangular, etc., the number and shape of supports, etc.)
are omitted; the measurements of size or weight are omitted; the
fact that it is a material object is specified, but the material of
which it is made is omitted, thus omitting the measurements that
differentiate one material from another; etc. Observe, however,
that the utilitarian requirements of the table set certain limits on
the omitted measurements, in the form of “no larger than and
no smaller than” required by its purpose. This rules out a ten-foot
tall or a two-inch tall table (though the latter may be sub-classified
as a toy or a miniature table) and it rules out unsuitable materials,
such as non-solids. [ITOE, 11–12]

Integration
I have discussed so far only the first stage of the concept-forming process:
the isolation of two or more similar items from a foil. At this stage, the
conceptualizer has mentally grouped things according to their similarities
and differences, but has no firm, enduring mental product — i.e., no concept.
To put it loosely, he has sorted but not yet conceptualized. To form a concept,
he must perform the second basic process of consciousness: integration.53
Having a concept requires integrating the similars into a new mental entity.
Integration is a process of mentally interconnecting a group of separate
items to form a permanent, usable whole. In grasping the range or category
of measurements that unites the similar items, the conceptualizer is
connecting them together in his mind, but what he has at this point

53 Measurement-omission is itself a kind of integration, as discussed above, but now

I’m talking about the enduring, open-ended integration that follows.
 H OW W E K N OW • 3 : C oncept- formation 124

is a transient state of awareness, not a permanent product that he can store,

carry forward, and use in the future. To achieve that, an act of full integration
is required. Such integration is achieved by the use of a word.
The word, as a sensuous symbol, has a crucial function in conceptual
cognition. The word completes the integration, binding the units into a single,
new mental entity. Although Nominalism is wrong in reducing everything
to words, we cannot overreact by throwing out the word with the bath water.
The use of a word is indispensable (though a visual image of a representative
concrete may play the same role for the first few concepts formed).
Contra Nominalists like Wittgenstein, the word is not what the units
have in common; what they have in common is the distinguishing
characteristic(s) — i.e., the range or category embracing their individual,
varying measurements. Thus, the word is not “the one in the many”;
rather, it is the means of retaining one’s grasp of the one in the many
— i.e., the means of retaining the awareness that these things, when
contrasted to those markedly different things, can be viewed as distinguish-
able members of a group. The word is not the object of cognition, but its form,
not the what but the how.
Rand illustrated her theory of concepts by drawing an analogy between
a concept and a file folder.54 This analogy is extremely valuable, and I will make
repeated use of it. Picture a sloppy office clerk accumulating a lot of paperwork
strewn across his desk. Soon there are hundreds of separate pieces of paper
pertaining to all sorts of different subjects. To bring order to his workspace,
his first step would be to put related pieces of paper together into separate piles.
Suppose that he makes one pile for employees’ expense accounts, another for
bills that have been paid, another for correspondence with suppliers, another
for résumés from job-applicants, and so on.
This rudimentary system brings certain advantages, but it will soon turn
into a mess, with overlapping piles scattered all over the desk, the chair
seats, the floor. Any passing breeze wreaks havoc with the piles. And after
he has formed a few dozen piles, it will become increasingly difficult to
remember where he placed a given pile or, when looking at a given pile,
to remember what it is about and why he formed it. Clearly, what he needs
to do is to introduce file folders to keep the papers in each pile together,

54 Rand drew this analogy in 1966, long before computers were common, so the analogy
is to physical file folders, not computer “file folders” (which would make less forceful
and less exact analogs to concepts). It is the old-fashioned, manila file folders,
with sheets of paper tucked inside them, that I ask the reader to bring to mind wherever
I use this analogy.
 I ntegration 125

so each can be handled as a single entity, and to begin systematizing them

by labeling the folders as to their contents, placing them in file-drawers, etc.
We can now cash out the analogy: the piles are analogous to the percep­
tual groupings that one establishes on the basis of observed similarities and
differences. The file folders are analogous to words in their conceptual function.
Words transform the piles into files. Assigning a word thus completes the
concept-formation process by linking the units together into a single, new
unit, thus forming a many-to-one network of connections.55
Speaking extemporaneously about the role of the word in this respect,
Rand noted:

. . . the word comes at the end of a process of conceptualization,

not at the beginning. One’s mind first has to grasp the isolation
and the integration which represents the formation of a con-
cept; but to complete that process — and particularly to retain it,
and later to automatize it — a man needs a verbal symbol.
But as far as the process of concept-formation is concerned,
the word is the result of the process.

Rand was then asked if that meant that the concept is formed prior to
introducing the word, with the word being used as a means of retaining
the concept. She replied:

That is a word’s main function, but its function is not merely

that. I meant exactly what I said: to complete the process.
Let me make this a little clearer. Suppose a child is forming
the concept “table.” First, he has to isolate a table from the rest
of his perceptual concretes, then integrate it with other tables.
Now, in this process words are not present yet, because he is
merely observing, and performing a certain mental process.
It is after he has fully grasped that these particular objects (tables)
are special and different in some way from all the other objects he
perceives . . . that he has to firm up, in effect, his mental activity in
his own mind by designating that special status of these particular
objects in some sensory form [i.e., by means of a word].

55 In describing the mechanics of the process, “linking” or “connecting” are better terms
to use than “associating,” which suggests a coincidental, non-logical connection.
The concept, in contrast to the word, is the logical structure and functioning of the set-up,
which remains the same whether one is linking dogs to “dog” or to “chien.”
 H OW W E K N OW • 3 : C oncept- formation 126

It is for the purpose not only of retaining the concept but also
of making and completing the process of concept-formation that
he has to designate the tables by some kind of sensory symbol.
The main function of doing so is to enable him to retain the
concept and be able to use it subsequently. But even apart from
the future, in the process of forming that concept, in order for
it not to remain a momentary impression or observation which
then vanishes — in order to make it a concept-forming process
— he has to identify what he has just observed in some one,
concrete, specific, sensory form. . . .
Now, it has to be a specific unit; and it cannot be specific,
it cannot be concrete, unless it is sensuous. Because reality
is concrete, and we perceive by means of our senses. Suppose
we attempted to have a concept which was symbolized by a cer-
tain feeling. Let’s say that I have a feeling of combined pleasure
and disgust at the concept “table” — suppose I tried to hold that
concept by means of such a feeling. Needless to say, that would
not be a concept. It would not last beyond the mood of the
moment. And I would not have performed the most important
part of the process — namely, the substitution of one handleable,
perceivable, firm, objective unit for the enormity which I want to
subsume under this concept. [ITOE, 164–165, 166–167]

Mentally linking similar existents to a word (“I’ll call these things tables”)
makes the grouping permanently accessible as a store of knowledge about
its units. Words’ information-handling power makes possible an exponential
growth in knowledge, the growth observable in a child’s cognitive develop-
ment from infancy to adulthood, as well as in the historical progression of
mankind’s knowledge.
It is a commonplace to observe that words — whose full implementa-
tion is language — permit the communication of knowledge and thus
the transmission of knowledge across generations. But Rand makes a deeper
point: beyond a certain level, language is essential to the acquisition of know-
ledge in a single, private mind.

Concepts and, therefore, language are primarily a tool of

cognition — not of communication, as is usually assumed.
Communication is merely the consequence, not the cause nor the
 I ntegration 127

primary purpose of concept-formation — a crucial consequence,

of invaluable importance to men, but still only a consequence.
Cognition precedes communication; the necessary precondition
of communication is that one have something to communicate.
. . . The primary purpose of concepts and of language is to provide
man with a system of cognitive classification and organization,
which enables him to acquire knowledge on an unlimited scale;
this means: to keep order in man’s mind and enable him to think.
[ITOE, 69]

Consider man’s predicament if he lacked words. Assume that he could

observe similarities and could form and remember perceptual groups.
Without words, his connections would be associational, rather than logical.
He would be caught in the Wittgensteinian chaos of forming not concepts
but a hodge-podge of associational clusters, as in my earlier example of
pumpkin — basketball — hockey puck — hockey stick — broom.
If we lacked words, one consequence would be the inability to form
generalizations. In order to move from observations of two or three concretes
to a generalization about all concretes of a kind, there must be a kind — rather
than a random series of associationally linked images. No generalization
could be formed from the pumpkin-to-broom cluster of associations.
Assigning a word completes the process of integration, producing a new
mental entity — a concept.
Summarizing and condensing all the preceding is Ayn Rand’s definition of
“concept”: “A concept is a mental integration of two or more units possessing
the same distinguishing characteristic(s), with their particular measurements
omitted.” [ITOE, 13]
By analyzing how concepts are formed, Rand has solved the age-old
“problem of universals.” She has shown what concepts refer to in r­eality.
Concepts do not refer to some Platonic “Form” in another dimension,
nor to an Aristotelian “essence” in things, nor to a Lockean non-specific
attribute, nor to Nominalist vague “resemblances.” Concepts refer to
existents that have, within a range or category, “some but any” degree of the
same characteristic(s). As Rand puts it:

Now we can answer the question: To what precisely do we refer

when we designate three persons as “men”? We refer to the fact
that they are living beings who possess the same characteristic
distinguishing them from all other living species: a rational
 H OW W E K N OW • 3 : C oncept- formation 128

faculty — though the specific measurements of their disting-

uishing characteristic qua men, as well as of all their other
characteristics qua living beings, are different. (As living beings of
a certain kind, they possess innumerable characteristics in common:
the same shape, the same range of size, the same facial features,
the same vital organs, the same fingerprints, etc., and all these
characteristics differ only in their measurements.) [ITOE, 17]

The Realist error is not merely that it assumes the existence in external
reality of pre-packaged universals, but also that it models concepts on
percep­tion, as if a concept were a high-powered percept (of the “universal”).
Although a concept derives from perceptual observation, to form a con-
cept is to interrelate concretes rather than engage in some kind of “insight”
into one of them. As the genus of Rand’s definition indicates, the concept is
an integration, an act that creates “a single, new mental entity which is used
thereafter as a single unit of thought.” [ITOE, 10]
When used in a proposition, to make a conceptual identification, concepts
do constitute awareness of the units to which they refer. But that awareness
is something quite different from perceptual awareness. First, perception is
automatic; conceptualizing is an active, conscious, and volitional process.
Second, perception is a source of new, primary data, but concept-formation
is not. Conception is not perception.
(Making logical connections among concepts does enable one to discover
new factual relationships, as when a child learns that 2 plus 2 is 4. But such
connec­tions merely make explicit facts that were already implicit in the
con­cepts — in this case, implicit in “4,” “2,” and “plus.” The child is not
connecting abstractions floating in the sky; he is tracing out the implications
of the observed facts he has conceptualized. Perception is still the only origi-
native source of the data used.)
The conceptual faculty, like consciousness in general, has a biological
function. To understand concepts fully, we need to consider not only how
they are formed and operate, but also the survival function they serve.
In the most general sense, it is clear that concepts serve the end of expanding
one’s range of knowledge. And the survival value of knowledge is manifest.
But the crucial question remains: what exactly is the function of concepts in
cognition? What does integrating similar concretes into a single new mental
entity do to expand the range of our knowledge?
 U nit- E conom y 129

To answer these questions, I turn to a major new topic: the inherent limit
on how much data the mind can deal with at once, and what that implies.

Unit-Economy
First some background. Everything that exists is finite. Whether it is an entity
or an aspect of an entity, each existent is exactly what it is, and no more
than that. This applies to the faculty of consciousness as well. The faculty has
a specific identity, and that identity permits it to do certain specific things,
but not more than that.
For instance, on the perceptual level, a given organism’s range of hearing
is limited to a certain range of sound intensities and frequencies. A dog can
hear sounds too high-pitched for a man to hear. The audible range is set by
the physical nature of the organism’s auditory system. The same applies to all
sense modalities; each has a specific means of operation, which determines
what it can respond to, with what resolution, how rapidly, etc.
Man’s conceptual faculty has its own specific, delimited identity. Accord-
ingly, every aspect of conceptual processing has operational limits, or “specs”
in computer-language. For instance, one cannot learn with infinite rapidity;
it takes time to take in, absorb, digest, and fully integrate a body of new
material.56
Another important operating limit pertains to intensity. The physical
analogy here is to pressure; pressure is force per unit area. The same force
applied to a smaller area produces a greater pressure, which is why a sharp
knife edge will cut something that a blunter knife will not: a given force
produces a more intense pressure when the area of contact is smaller.
Likewise, the mental intensity one can bring to bear is limited, and it can be
spread over a greater or lesser number of units. Thus, the power of concen-
tration: the same intensity is applied to a smaller number of units, increasing
the “mental pressure” applied to each unit.

56 In effect, there is some neural equivalent of the speed rating for a computer’s CPU.
The CPU rating, in cycles per second (hertz), refers to its basic “clock cycle,” the unit
of time required for its simplest operations. A computer has other rate-specifications
— e.g., the various bit-per-second ratings for memory chips, bus speed, and disk access.
There are doubtless analogs to this in limitations on the speed of brain functions,
resulting in the span of time it takes to memorize something, to recall something, etc.
There are other neural analogies to computer architecture, such as the number of CPU
registers and the “word” length (e.g., 32-bit vs. 64-bit architecture).
 H OW W E K N OW • 3 : C oncept- formation 130

A man cannot concentrate on everything presented to him at once.

Giving more attention to one thing is achieved by giving less attention
to other things. This is apparent even in vision: one’s visual field is much
wider than what is in clear, focal vision at any given moment. That limit
applies not just to vision, but to every aspect of consciousness, including the
conceptual faculty. In particular, there is a limit on the number of distinct
units one can hold in focal awareness at any given moment.
This simple fact is the basis of the conceptual level’s power and biological
value. Everything that concepts do for man, everything that raises man above
the animal levels, is traceable to the fact that concepts permit economizing
on units. “Conceptualization is a method of expanding man’s consciousness
by reducing the number of its content’s units.” [ITOE, 64]
Rand introduces the principle of unit-economy by comparing man’s
cognitive capacity to that of animals. Since the following passage goes to the
essence of the Objectivist theory of concepts, I quote it at length.

The story of the following experiment was told in a university

classroom by a professor of psychology. I cannot vouch for the
validity of the specific numerical conclusions drawn from it,
since I could not check it first-hand. But I shall cite it here, because
it is the most illuminating way to illustrate a certain fundamental
aspect of consciousness — of any consciousness, animal or human.
The experiment was conducted to ascertain the extent of the
ability of birds to deal with numbers. A hidden observer watched
the behavior of a flock of crows gathered in a clearing of the
woods. When a man came into the clearing and went on into the
woods, the crows hid in the tree tops and would not come out
until he returned and left the way he had come. When three men
went into the woods and only two returned, the crows would not
come out: they waited until the third one had left. But when five
men went into the woods and only four returned, the crows came
out of hiding. Apparently, their power of discrimination did not
extend beyond three units — and their perceptual-mathematical
ability consisted of a sequence such as: one-two-three-many.57

57 This story’s source is an 18th-century work, The Intelligence and Perfectibility of Animals,
by C. G. Leroy. His almost identical account concludes: “In fine, it was found necessary to
send five or six men to the watch-house to put [the crow] out in her calculation. The crow,
thinking that this number of men had passed by, lost no time in returning.” [Leroy, 126]
 U nit- E conom y 131

Whether this particular experiment is accurate or not, the truth

of the principle it illustrates can be ascertained introspectively:
if we omit all conceptual knowledge, including the ability to
count in terms of numbers, and attempt to see how many units
(or existents of a given kind) we can discriminate, remember and
deal with by purely perceptual means (e.g., visually or auditorially,
but without counting), we will discover that the range of man’s
perceptual ability may be greater, but not much greater, than that
of the crow: we may grasp and hold five or six units at most.
This fact is the best demonstration of the cognitive role
of concepts.
Since consciousness is a specific faculty, it has a specific nature
or identity and, therefore, its range is limited: it cannot perceive
everything at once; since awareness, on all its levels, requires an
active process, it cannot do everything at once. Whether the units
with which one deals are percepts or concepts, the range of what
man can hold in the focus of his conscious awareness at any given
moment, is limited. The essence, therefore, of man’s incomparable
cognitive power is the ability to reduce a vast amount of informa-
tion to a minimal number of units — which is the task performed
by his conceptual faculty. And the principle of unit-economy
is one of that faculty’s essential guiding principles. [ITOE, 62–63]

There are two distinguishable points here:

1. There is a limit on the number of units man can simultaneously hold
in focal awareness.
2. Concepts are man’s means of overcoming that limit (by condensing
many units into one).
The first point has become established in the literature of cognitive
psychology, stemming from George A. Miller’s classic article, “The Magical
Number of Seven, Plus or Minus Two: Some Limits on Our Capacity for
Processing Information.” [Miller, 1956]
In informal conversation, Rand often referred to this capacity-limit as
“the crow epistemo­logy,” or just “the crow.” That usage has become common
in the Objectivist literature, and I will use it as a shorthand expression from
this point on. (The value of the shorthand is itself an instance of “the crow
epistemology.”) A few more examples of “the crow epistemology” should nail
down the point.
 H OW W E K N OW • 3 : C oncept- formation 132

A quick glance suffices to distinguish two peas from three peas, but not
42 peas from 43. Counting the peas enables their quantity to be easily held
as one unit (“42”) and that quantity is easily distinguished from any other
enumerated quantity.58 An average person would begin to run into difficulty
in trying to distinguish six peas from seven by sheer perception, and almost
certainly could not reliably distinguish ten peas from eleven (without count-
ing or forming subgroups).
Bear in mind that a unit is something held as a separately discriminated
item. To hold in mind something as a unit requires that, while focusing on it,
one retains awareness of its separate, distinct identity, as well as its similarity
to the other members of the group. Merely viewing several dozen dandelions
spread across a lawn does not count as holding in focal awareness dozens
of units — i.e., holding in one frame of awareness their individual features
in addition to their similarity. But doing that for three dandelions poses
no problem.
The same need for unit-reduction applies, on a higher level, to the numerals.
It is easy to distinguish the numerals “10” and “100,” and hold in mind their
distinct identities, but the case is quite different for “1000000000000000000000”
and “10000000000000000000000.” Commas are introduced into large num-
bers to form groups of threes, thus reducing the number of perceptual units
by a factor of three.
For the same reason we introduce the new number-words “thousand,”
“million,” “billion,” and so on. Thus, we can think “four billion” rather than
four thousand thousand thousand.
Science, which routinely works with very large numbers, introduces unit-
economical notation that counts the decimal places (which are powers of 10).
E.g., the mass of the Earth is: 5,980,000,000,000,000,000,000,000 kilograms.
In the unit-economizing form of notation, that becomes: 5.98 × 1024 kilo-
grams. If we then have to divide 5.98 × 1024 by 2.0 × 1021 it becomes quite
easy: divide 5.98 by 2 and subtract the smaller exponent, giving 2.99 × 103.
Though quantities and numerals provide the simplest illustration of
the need for unit-economy, the principle applies universally to cognition.
In any field and on any issue, economizing on units magnifies the power
of one’s consciousness by concentrating a given intensity of thought
on a small number of units. This increases the per-unit mental intensity,
giving one mental leverage.
Concepts achieve an immensely greater degree of condensation than
does counting or the use of commas to break up long numerals. The number

58 For more on counting as unit-reduction, see ITOE, 63–64.

 U nit- E conom y 133

of units that a concept retains by means of a single word is not fixed and
finite, but unlimited. A concept, in Rand’s term, is “open-ended”: it applies
to all the existents of a given kind — past, present, and future. “Man,” for
instance, does not apply to merely those men whom one has seen or considered;
it applies to every man who has ever existed or will ever exist.
It is the substitution of one word for an open-ended number of similar
units that gives concepts their cognitive power. If concepts were frozen
after their formation, rather than being open-ended, they would be reduced
to proper names for the small group of units from which they were formed.
The open-ended nature of concepts permits inductive generalizations to
be formed, so that we can apply the knowledge gained from study of some
concretes to every concrete of that kind that we subsequently encounter.
Crucially, a concept is open-ended not only in regard to its application but also
in regard to its content — i.e., the information about the units that it stores.
All facts learned about the units qua units are to be added to one’s conceptual
file. When one learns that all men have kidneys, the concept “man” stores
this new information.59 Such updating is required if the conceptual file is
to function as a file, rather than being frozen in place, like a set of papers
stapled together.
How is the open-ending of a concept accomplished? What are the under-
lying mechanics? In his ongoing experience, the child comes to realize that
he has discovered a recurring type of thing, and that leads him to form an
intention (wordlessly, at this early stage) to apply the concept to all future
instances. E.g., in forming the concept “table,” the child repeatedly encounters
new tables, motivating him to, in effect, set a “standing order” to identify
every object of this type as “a table.” The same applies to his discovery of new
features of tables. (Simple identifications of this type become automatized;
the open-ending of more advanced concepts requires an active, deliberate
commitment to expanding one’s knowledge and keeping one’s mental files
up to date.)
The issue of open-endedness brings to the surface a basic difference between
the Objectivist and the Realist theories of concepts. For Realism, a concept
refers to only the “universal” feature supposedly found inside the concretes.60

59 In fact, the explicit, conscious identification, “Men have kidneys” is what accomplishes the
filing: focusing on (and valuing knowing) this fact tends to form a corresponding link in
one’s brain between (to oversimplify) the neural correlates of “man” and “kidneys.” (See
Chapter 5 on conceptual identifications.)
60 This closed-ended, frozen view of concepts is the error that generates the “analytic-synthetic”
dichotomy (see Chapter 5).
 H OW W E K N OW • 3 : C oncept- formation 134

For Objectivism, the concept refers to the units — the things out there in reality,
with all their characteristics, known and yet to be discovered.
For instance, the concept “man” refers to men. Men have the characteristics
that they have, whether or not these characteristics are yet known. Human
beings had capillaries before William Harvey discovered the existence
of capillaries. But when Socrates used the concept “man,” he meant those
entities — entities which have capillaries, even though Socrates was ignorant
of that characteristic of the entities his concept designated. Since his concept
of “man” denoted men, and since men have capillaries, the meaning of his
concept included the capillaries.
Since our concepts mean and refer to existents, not just our present
knowledge of those existents, in a sense we mean more than we know.
By observing some of the characteristics, we are able to classify them.
But we are classifying them — the existents in toto — not just our present
knowledge of them.
Objectivism rejects the assumption that a concept equals its definition.
The definition’s specific function is to be a unit-economical way of summarizing
a concept’s meaning, but the meaning of a concept includes much more than
just the characteristics that we use to conceptualize things. The distinguishing
or defining characteristics are how we classify, not what we classify.
What we classify are the existents, with all their characteristics. The concept
“pea” refers to the peas, not to only those characteristics that, in a given context
of knowledge, best distinguish peas from other legumes.
Indeed, since existence is identity, the existents are their characteristics.
By the same token, a characteristic is one aspect of a thing’s total identity.
Characteristics do not exist as such, but only as characteristics of entities
(or of an entity’s attributes, actions, relationships, etc.) Characteristics can be
isolated mentally but cannot be physically separated from the entities that
possess them.61
Accordingly, the concept of an entity is the concept of the whole entity,
not just of some characteristics of the entity. The concept of an action is the
concept of the whole action, not just of some characteristics of the action.
Even the concept of an attribute is the concept of the whole attribute, not just
of some characteristics of the attribute.62

61 Sometimes one speaks of parts of entities as being “characteristics” of them, but a part of
an entity is actually itself an entity, not a characteristic. (See ITOE, 264–274.)
62 Attributes can have characteristics, in the broad sense of that term. E.g., a man’s attribute
of being suntanned has such characteristics as: its degree of darkness, its recency, and its
effects on his appearance and health.
 U nit- E conom y 135

Since a concept subsumes units as they actually exist, with all the char-
acteristics that they possess, the information stored by the concept expands
with every new discovery about the nature of the units. Rand uses the file-
folder analogy to make the distinction between the knowledge stored in
a concept, in a given person’s mind at a given time, and what that concept
refers to in reality:

Since concepts represent a system of cognitive classification,

a given concept serves (speaking metaphorically) as a file folder
in which man’s mind files his knowledge of the existents it
sub­sumes. The content of such folders varies from individual
to individual, according to the degree of his knowledge — it
ranges from the primitive, generalized information in the mind
of a child or an illiterate to the enormously detailed sum in the
mind of a scientist — but it pertains to the same referents, to the
same kind of existents, and is subsumed under the same concept.
This filing system makes possible such activities as learning,
education, re­search — the accumulation, transmission and expan-
sion of knowledge. [ITOE, 66–67]

The cognitive function of concepts pertains to what Rand calls “psycho-

epistemology” — the psychological mechanics of the cognitive process.63
In regard to concepts, the “crow” limitation is a psychological fact about how
consciousness works, a fact that explains what concepts accomplish for us
in cognition. Since man can deal with only a few units at a time, he needs
the condensation provided by concepts — including the automatization of
the word which designates the concept. Thus, the cognitive value of concepts
is psycho-epistemological: the conceptual condensation of “the many” into
a “one” overcomes the limits of the “crow.”
The full cognitive power provided by a concept’s unit-economy and
open-endedness will become apparent in the next chapter, when I discuss
“abstraction from abstractions.”
Rand provides a vivid and memorable summary of the Objectivist theory
of concepts by means of a mathematical analogy:

63 Rand defines the field of “psycho-epistemology” as: “the study of man’s cognitive
processes from the aspect of the interaction between the conscious mind and the auto-
matic functions of the subconscious.” [RM, 18]
 H OW W E K N OW • 3 : C oncept- formation 136

The basic principle of concept-formation (which states that the

omitted measurements must exist in some quantity, but may exist
in any quantity) is the equivalent of the basic principle of algebra,
which states that algebraic symbols must be given some numeri-
cal value, but may be given any value. In this sense and respect,
perceptual awareness is the arithmetic, but conceptual awareness
is the algebra of cognition.
The relationship of concepts to their constituent particulars
is the same as the relationship of algebraic symbols to numbers.
In the equation 2a = a + a, any number may be substituted
for the symbol “a” without affecting the truth of the equation.
For instance: 2 x 5 = 5 + 5, or: 2 x 5,000,000 = 5,000,000 +
5,000,000. In the same manner, by the same psycho-epistemolog-
ical method, a concept is used as an algebraic symbol that stands
for any of the arithmetical sequence of units it subsumes.
Let those who attempt to invalidate concepts by declaring
that they cannot find “manness” in men, try to invalidate algebra
by declaring that they cannot find “a-ness” in 5 or in 5,000,000.
[ITOE, 18]

The Objectivist theory holds that concepts are formed by grasping

similarity against difference; the measurement-relationships among existents
result from their having a commensurable characteristic. Contra Realism,
concepts are not formed by passively intuiting some metaphysical “universal.”
Concepts are not formed by zeroing in on one concrete, seeking an “insight”
into its “essence,” but by widening one’s focus, interrelating several concretes
through a process of contrast and comparison, then integrating them by
means of a word.
Abstraction is not subtraction but integration.
 HIGHER-LEVEL
4
CONCEPTS

C oncepts of concrete entities, such as “tree” and “dog,”

represent only the entrance to the conceptual level. At this beginning
stage of conceptual development, a child can subsume the next tree under
“tree,” and the next dog under “dog,” and children do generally delight in
pointing to things and announcing their names. But there is little they learn
from doing so: there is as yet no conceptual knowledge in their mental “file
folders”; there are only perceptual memories of past trees or dogs plus a few
associational links to other perceptual memories.
In order to gain the full power of the conceptual level, the child must be
able to add explicit, conceptual content to his conceptual files. To get beyond
“Here tree” and “There dog,” he needs to be able to think such a thought as
“Trees grow tall.” But “grow” is an action, not an entity, and “tall” is an attri-
bute. Similarly, he needs to be able to think “Dogs are animals,” but “animal”
is a higher-level concept, not one formed from direct perception, as “dog” is.
Dogs are perceptually similar; animals, which range from flies to elephants,
 H OW W E K N OW • 4 : H igher- L evel C oncepts 138

are not. How does a child form higher-level concepts, since they are too
abstract to be formed directly from perception as first-level concepts are?
Higher-level concepts are formed by the process Rand calls “abstraction
from abstractions.” It consists of turning the concept-forming process back on
its own products: the input to the process is not concretes but earlier-formed
concepts. The input used to form the concept “animal” is the prior concepts
of, say, “dog,” “flea,” and “elephant.” The process is iterative: “animal” and
“plant” will become the input for forming, years later, the still more abstract
concept “organism.”64
Abstraction from abstractions is also the process that enables the child to
form concepts of the characteristics of entities — what the entities are and
do — to identify not only “dog,” but also “tail,” “wagging,” and “bushy,” not
only “tree” but also “tall,” “growing,” “evergreen,” and “deciduous.”
It is abstraction from abstractions that allows one to acquire the full
human vocabulary, giving one the power of thought, the power that has
enabled man to become, in Dobzhansky’s phrase, “the lord of creation.”
A child who has only first-level concepts is, metaphorically speaking,
living hand-to-mouth, without any tools of production. The child’s cognitive
progress depends upon his acquiring the equivalent of baskets, nets, weap-
ons, huts — all of which expand his efficacy, better his life, and save him time.
Just as the development of basic tools, and then tools for making tools, made
possible man’s material progress, so the step-by-step formation of concepts at
higher and higher levels of abstraction makes possible his intellectual prog-
ress (which is required for his material progress).
If somehow, overnight, we lost all higher-level concepts, industrial
civilization would be wiped out, and we would be helpless — blinking uncom-
prehendingly at the machines, computers, automobiles, and factories we
could neither value nor use. It takes higher-level concepts to understand that
the oddly shaped box is not just a “box” but a computer, that the books on
the shelf can be read to learn things, that two plus two is four — in other
words, to recognize anything beyond the level of “dog,” “berry,” and “rain.”
Higher-level concepts are essential to the human level of existence.

64 The earlier concepts are not the only input to the process; a growing body of related
knowledge is also required. Abstraction from abstractions is not a mechanical, but
a cognitive and contextual process.
 T he F ormation of H igher- L evel C oncepts 139

The Formation of Higher-Level Concepts

Let’s look more closely at the nature and genesis of higher-level concepts. At
the opening of her chapter, “Abstraction from Abstractions,” Rand identifies
the two different directions the process can take:

Starting from the base of conceptual development — from the

concepts that identify perceptual concretes — the process of
cognition moves in two interacting directions: . . . toward wider
integrations and more precise differentiations. [ITOE, 19]

Wider Integrations
Wider integrations are the simpler case. To form wider concepts, one applies
the same processes of differentiation and integration that one used to form
first-level concepts, but with earlier-formed concepts taken as the units to
be integrated. A simple example is forming the concept “furniture” from the
prior concepts “table,” “bed,” “couch,” “dresser,” etc. These first-level concepts
are taken as units, which are then differentiated from architectural features,
such as walls and doors, and/or from other objects in a room, such as plates,
appliances, and rugs. Then one integrates the items of furniture according
to their possession of a distinguishing characteristic, with their differing
measurements omitted. Rand describes the process of forming the higher-
level concept “furniture,” as it starts from first-level concepts such as “table,”
“bed,” “sofa,” etc.

The distinguishing characteristic of the new concept is deter-

mined by the nature of the objects from which its constituent
units are being differentiated, i.e., by their “Conceptual Common
De­nom­i­na­tor,” which, in this case, is: large objects inside a human
habitation. The adult definition of “furniture” would be: “Movable
man-made objects intended to be used in a human habitation,
which can support the weight of the human body or support
and/or store other, smaller objects.” This differentiates “furniture”
from architectural features, such as doors or windows, from
ornamental objects, such as pictures or drapes, and from a variety
of smaller objects that may be used inside a habitation, such as
ashtrays, bric-a-brac, dishes, etc.
 H OW W E K N OW • 4 : H igher- L evel C oncepts 140

The distinguishing characteristics of “furniture” are a specified

range of functions in a specified place (both are measurable
characteristics): “furniture” must be no larger than can be placed
inside a human habitation, no smaller than can perform the
specified functions, etc. [ITOE, 22]

Thus, the process of integrating first-level concepts into wider ones

follows the same pattern as that of forming first-level concepts: the grasp of
similarities against a background of differences, the omission of measurements
on the “some but any” principle, and the integration of the earlier-formed
concepts, by means of a word, to create a new, open-ended, mental entity.
Note that the widening process treats the old concepts’ distinguishing
characteristics as measurements to be omitted. This is entirely appropriate,
given that a distinguishing characteristic is a range (or category) of measure­
ments. E.g., the distinguishing characteristic of green is the wavelength
range of 500–565 nanometers, and we omit measurements within that range
to grasp “green”; likewise, to get “hue,” we omit the hue measurement-ranges,
expanding to the entire visible spectrum of about 380–750 nanometers,
on the premise that a hue must have some wavelength within that range,
but may have any.65
Rand shows how this point applies in omitting measurements to form
“furniture”:

The distinguishing characteristics of these units are specified

categories of measurements of shape, such as “a flat, level surface
and support(s)” in the case of tables. In relation to the new
concept [“furniture”], these distinguishing characteristics are now
regarded in the same manner as the measurements of individual
table-shapes were regarded in forming the concept “table”: they
are omitted, on the principle that a piece of furniture must have
some shape, but may have any of the shapes characterizing the
various units subsumed under the new concept. [ITOE, 21–22]

In terms of the file-folder analogy, the wider concept, “furniture,”

is a larger file folder, such as a hanging folder, which holds other file folders.

65 “Color” is wider still, embracing measurement differences in saturation and luminosity.

And, as mentioned in the previous chapter, it is somewhat an oversimplification to equate
hue with wavelength, but the complexities of color vision are irrelevant to my purpose,
which is to illustrate the concept-forming process.
 T he F ormation of H igher- L evel C oncepts 141

The resulting view is vitally different from that of the Realists. The Real-
ists’ abstraction-as-subtraction model implies that wider concepts have less
cognitive content than the narrower ones from which they are formed.
For the Realists, “table” abstracts the “universal” from individual tables by
mentally subtracting away and discarding everything that differs among tables.
Then, “furniture” discards even more, retaining only that which all items
of furniture have in common, discarding even the “universal” features of tables
(and of beds, sofas, dressers, etc.). It was this abstraction-as-subtraction
premise that led Hegel to conclude that the concept of “existence” has no
content at all (see page 119).
Rand’s theory recognizes that concept-formation is integrative, which
means that the wider concept contains more cognitive content than any of
the narrower ones from which it is formed. In forming the concept “table,”
one grasps relationships among tables (as against a foil, such as a chair).
To form the concept “furniture,” one must retain that fact plus grasp some-
thing more: the wider relationships among tables, chairs, beds, etc., in con-
trast to rugs, doors, and other items from which furniture is differentiated.
Thus, the resulting wider concept, “furniture,” subsumes and includes
all the information about the first-level concepts (plus their interrelation-
ships). The hanging file folder contains the narrower file folders; a file drawer
contains the hanging folders; a file cabinet contains several file drawers,
and so on. And since the first-level concepts denote entities in reality, with
all their characteristics, the wider concept does as well; the wider concept
represents a wider perspective, as “furniture” does by virtue of integrating
tables, chairs, etc.
For Realists, reaching a more abstract level means having a narrower “insight”
into a universal embedded inside a given universal — the “furniturehood”
lurking inside of “tableness” and “bedness.” For Realists, the wider the concept,
the emptier of cognitive content.
Realism cuts itself off from reality at the first level, since it holds that
concepts like “table” do not refer to entities, but only to the supposed universal
element in them; it then compounds the error by taking “furniture” to refer
to an even more rarefied universal, leaving the concept with even less content.
The Objectivist theory retains the link to reality: wider concepts refer,
via the narrower concepts they integrate, to existents as they are in reality,
with all their characteristics.
 H OW W E K N OW • 4 : H igher- L evel C oncepts 142

Subdivisions
The second type of higher-level concept consists of subdivisions or “narrow-
ings” of existing concepts. Narrowings have virtually never been discussed
in the history of epistemology — the discussion of higher-level concepts,
scanty as it has been, has ignored narrowings to focus exclusively on widenings.
Yet narrowings — conceptual subdivisions — are by far the more common
type of higher-level concept.66
There are two ways of subdividing an earlier concept: 1) by narrowing the
earlier concept’s measurement-range, or 2) by adding a new characteristic,
a characteristic not used in forming the earlier concept.
The simpler case is subdividing by narrowing an existing range of measure-­
ments, as “scarlet” narrows “red” or “toy poodle” narrows “poodle.”
Consider the process of subdividing the concept “table” to form “coffee
table” and “dining table.” These two subdivisions are formed by narrowing
the height measure­ments (coffee tables being lower than dining tables) and
narrowing the function-measurements (to keep objects within easy reach
while seated on a couch vs. while seated higher, in dining chairs). In the
preceding chapter, I gave the example of a teacher grading students’ tests.
The introduction of plus and minus grades is a clear case of making sub­
divisions by narrowing the measurement-ranges: if the grade of B is assigned
to 80–89, B+ is typically used for 87–89, and B- for 80–82.
In forming this type of subdivision, the old distinguishing characteristic
serves as the CCD, within which a new, narrower range of measurements
serves as the distinguishing characteristic for the new concept. Within the
narrower range, the particular measurements are omitted.
The second type of narrowing is more complex and more interesting:
cross-classification. An existing concept is narrowed according to the
presence or absence of a new characteristic. Rand gives the example of
forming “desk” as a narrowing of “table.” The essential difference between
desks and other tables consists in the presence of drawers for stationery
supplies — a characteristic not specified in, and orthogonal to, the original
concept “table.”

66 Based on making random samples of words in the dictionary, I estimate that 90% of
higher-level concepts are narrowings. The predominance of narrowings over widenings
makes perfect sense: each narrowing divides one concept into several subsets, and each
widening does the reverse, combining several concepts into one superset. By analogy
if there are an average of 20 counties per U.S. state, then there are 1,000 counties (subdi-
visions) and only one nation (wider category).
 C onceptual H ierarch y 143

The concept “man” can be subdivided in numerous ways by adding dif-

ferent characteristics to the distinguishing characteristic: “boy” (adding age
and gender), “Negro” (adding race), “Republican” (adding political affiliation),
“economist” (adding profession). To give an idea of the range of possibilities that
cross-classifications open up, here are some further examples — each using
a different basis of cross-classification: “cousin,” “student,” “criminal,” “Baptist,”
“environmentalist,” “theorist,” “hero.”
Cross-classifications range across a rich and interesting variety of cases, and
they serve many different cognitive purposes. As with all conceptualization,
the basic process remains the same: differentiation, measurement-omission,
and integration.
Subdivision is a fairly easy process to perform, because the integration
is already established, and one need only make finer differentiations within
it. Differentiation is an easier task than integration.

Conceptual Hierarchy
The process of abstraction from abstractions gives rise to a phenomenon of
immense importance for epistemology: hierarchy.
In its most general usage, a “hierarchy” is an ordered relationship among
items: each item is located in a series according to its dependency upon
the item below it.67 Examples are military rank (private, sergeant, captain,
etc.) and levels of courts (trial, appellate, supreme). The hierarchy termi-
nates in (or begins with) a primary, or set of primaries. This primary is the
fundamental item of the series, the one on which all the others depend.
For instance, the military hierarchy terminates in the Commander-in-Chief,
or each of the floors of a building depends on the one below it, terminating
in the ground floor.
The hierarchy of concepts results from the iterative nature of abstraction
from abstractions. Higher-level concepts depend upon the earlier ones that
they integrate or subdivide. “Organism” (a widening) depends on “plant”
and “animal,” and these concepts, in turn, depend on earlier concepts
formed from perception, such as “tree” and “bush,” and “dog” and “pig.”
“Celebrity” (a narrowing) depends on “fame” and “man.”

67 There is an unfortunate ambiguity in speaking of “above” and “below” here. I take the
perspective that a hierarchy moves “up” from its base or foundation, though one could
alternatively say that the hierarchy flows down from the top (e.g., the Commander-in-
Chief is usually thought of as being at the “top” of the military chain of command).
 H OW W E K N OW • 4 : H igher- L evel C oncepts 144

This dependency is absolute: without lower-level concepts to link them

back to perceptual reality, the higher-level concepts lose their meaning,
becoming empty sounds.
The hierarchy of concepts concerns the necessary order of their formation.
By virtue of the identity of man’s consciousness, concepts have to be formed
in a certain order. The concept of “stockholder” cannot be formed before the
concept of “stock,” which cannot be formed before the concept “corporation,”
which cannot be formed before the concept “business.”68
It is the condensation afforded by the first-level concepts that permits
the formation of wider concepts, because the wider concepts would not be
graspable without that condensation. Take the formation of the wider concept
“furniture.” Tables, beds, and dressers have no perceptual features in common.
A table does not look like a bed, and neither of them looks like a dresser. Nor
do these objects look similar when contrasted to such non-furniture items
as a kitchen cabinet or a refrigerator. The level on which items of furniture
are in fact similar — which involves their movability plus their function — is
too wide a level to be grasped directly from perception. In other words, the
perceptual variety of items of furniture is too great for the “crow.”
Likewise, consider the concept “organism” as denoting any living being,
whether plant or animal. A child can see the similarity among dogs, or among
horses, or among trees. But now picture the child, before he has any concepts,
looking at a grassy field on which there are dogs, horses, trees, and rocks.
Given the limits of the “crow,” no child can grasp by just looking at the
scene that the living organisms (dogs, grass, horses, trees) are similar
as opposed to the rocks. In order to reach the required scale of aware-
ness, the child must first condense the dogs into “dog,” the horses into
“horse,” then “dog” and “horse” into “animal,” and he must separately
condense “tree” and “grass” into “plant.” Only after these intermediate con-
cepts have been formed and automatized can he then consider the two units
“animal” and “plant” in opposition to “rock.” (Even this is a simplification:
one would have to form many more concepts — e.g., “growth,” “repro-
duction,” and “action.” In fact, one would have to make propositional
identifications — e.g., “The animal and the plant act on their own, but the
rock remains where it is unless pushed or pulled by an external force.”)
To fully appreciate the hierarchy of concepts, one must firmly disting­
uish between grasping concepts and merely uttering words. Just seeing

68 The overall order allows, in some cases, for options regarding the particular details of its
implementation. See ITOE, 204–217.
 C onceptual H ierarch y 145

a man wearing a uniform is not sufficient to form, from this perception,

the concept “soldier.” Other people who wear uniforms — nurses, police-
men, prep school students — are not soldiers, and the concept “soldier” does
not mean simply “people wearing this kind of clothing.” (In contrast, the
concept “man” can be, and is, formed directly from perception.) A given child
may use the word “soldier” just on the basis of the uniform, by mimicking
the language of adults, but this is imitation, not knowledge — not a grasp
of the concept “soldier.”
Even for adults, such non-conceptual imitation is all too possible,
especially in the case of higher abstractions like “justice,” “freedom,” “romance,”
and “inflation.” Many people use such words while having only a woozy,
shifting idea of their meaning and their actual connection to the facts of
reality. Rand describes the right and wrong choices one confronts in moving
beyond first-level concepts:

There are many different ways in which children proceed to learn

new words thereafter. Some (a very small minority) proceed
straight on, by the same method as before, i.e., by treating words
as concepts, by requiring a clear, first-hand understanding (within
the context of their knowledge) of the exact meaning of every
word they learn, never allowing a break in the chain linking their
concepts to the facts of reality. Some proceed by the road of
approximations, where the fog deepens with every step, where
the use of words is guided by the feeling: “I kinda know what
I mean.” Some switch from cognition to imitation, substituting
memorizing for understanding, and adopt something as
close to a parrot’s psycho-epistemology as a human brain can
come — learning, not concepts nor words, but strings of sounds
whose referents are not the facts of reality, but the facial expres-
sions and emotional vibrations of their elders. And some
(the overwhelming majority) adopt a precarious mixture of differ-
ent degrees of all three methods.
But the question of how particular men happen to learn
concepts and the question of what concepts are, are two
different issues. In considering the nature of concepts and the
process of abstracting from abstractions, we must assume a mind
capable of performing (or of retracing and checking) that process.
[ITOE, 20–21]
 H OW W E K N OW • 4 : H igher- L evel C oncepts 146

The hierarchy of concepts is the order required to grasp the concept, not
merely to parrot the word. And “grasping” means “taking firm, secure hold of,”
not merely “having a faint, brushing encounter with.” To grasp “soldier,” one
must first grasp “army” or “military,” which in turn requires grasping “war,”
“nation,” and so on, until one gets back to first-level concepts — i.e., those
that can be formed from perception, without needing to use any previously
formed concepts in the process.
Thus, the hier­archy of concepts is a logical hierarchy, not a chronological
one. The specifics of the chronological order often vary with circumstances.
For instance, an American child may not learn the concept “dingo” until
many years after he has learned the concept “animal,” but that does not mean
that knowing “animal” is a prerequisite of forming “dingo,” as shown by the
fact that an Australian child might well form “dingo” before “animal.” One
can form “dingo” directly from perception, which means that “dingo” is
a first-level concept, regardless of whether one forms it at age 1 or age 10.
“Animal,” in contrast, is inherently second-level. To grasp it, one needs to have
conceptualized some types of animals, but which particular animals a child
uses — whether dingos and kangaroos, or cats and pigs — depends merely on
the order in which he happens to observe various animals.
(In the case of two widely-separated concepts, there may be no particular
hierarchical order between them. For instance, nothing dictates the order in
which one learns the concepts “arithmetic,” and “divorce.” But a prerequisite
of forming the concept “arithmetic” is having the concept “number,” and one
cannot grasp “divorce” before having “marriage.”)
It is easy to be insensitive to the hierarchical order and to squash concepts
down to one level, because after a concept has been learned and automatized,
it becomes so much “second nature” that it feels as if one always knew the
concept, or as if it could be formed from perception — like “dog.” “Retro-
active self-evidency,” one might call this error. Retroactive self-evidency is
particularly pervasive in the computer field: using a new application passes
quickly from the befuddling to the seemingly self-evident — one feels about
such know-how, “it’s just obvious.” But in almost all cases, it is obvious only
after it has been identified.69
The hierarchical view of concepts is not that of the typical man on the
street — nor of most philosophers in their armchairs. Far more common
is the assumption that concepts are not grasped consciously but absorbed
passively from the social setting in which the use of language develops. But

69 Even natural selection was deemed “obvious” — once ­Darwin had discovered it. Biologist
G. G. Simpson wrote: “Darwinian natural selection was based on a few concepts all obvi-
ously true once they had been pointed out. After Darwin had pointed them out, honest
biologists agreed they had been extremely stupid not to see them before.” [Simpson, 51]
 C onceptual H ierarch y 147

concepts represent knowledge. To recognize that certain existents are similar

to others is to know something about the world. In the case of the simple,
first-level concepts with which a child begins, each concept is formed by
recognizing a similarity against a background of differences, and that is the
discovery of a fact. To see the similarity of dogs to each other, as against cats,
is to learn a fact of biology, not to adopt a social convention, like learning to
say “Gesundheit” upon hearing someone sneeze.
(For an infant or toddler, the knowledge condensed in a first-level concept
is not yet explicit, propositional knowledge . Upon first entering the concep-
tual level, one is not yet able to form a sentence and make an assertion. But
the child’s later ability to form propositions is based upon his having previ-
ously stored knowledge in the concepts he uses to do so. For propositions to
be cognitive, concepts must be cognitive.)
To illustrate the fact that a concept represents knowledge, not a social
practice in the use of words, consider the case of words whose meaning one
has forgotten. If one can no longer recall from high-school biology what
“monocotyledon” means, one is unable to make meaningful statements about
monocotyledons. What is gone is not the word, not a social practice, but
knowledge: an understanding of which things in reality are similar in which
respect. If one recalls that the term has something to do with plants, one can
indeed express that in a proposition. But “something to do with plants” does
not constitute a grasp of the concept, even though one had grasped it years
ago. The knowledge can be regained, of course — in this case, by looking it up,
in a source written by someone possessing the knowledge.
Neither the concept “monocotyledon” nor even the concept “dog” repre-
sents a gift granted by society. Grasping any concept is an achievement, one
that requires effort, attention, and doing mental work. The product of that
work, a concept, is a grasp of facts of reality — i.e., knowledge.
We need to bear in mind Rand’s observation that “no matter how many
men mouth a concept as a meaningless sound, some man had to originate it
at some time.” [ITOE, 21] To originate a new concept requires making fresh
observations, and, for higher-level concepts, that origination also requires
a sustained effort to relate things, in a quest to find common denominators.
A fairly simple example of the difference between origination and learn-
ing from others is provided by the concept “velocity.” High school physics
students today have little difficulty in integrating the concepts of “speed” and
“direction” to form the concept “velocity,” which denotes speed in a given
direction. But consider the fact that the origination of this concept, which is
the key to understanding the laws of motion, required the genius of a man
 H OW W E K N OW • 4 : H igher- L evel C oncepts 148

like Galileo. The simple concept “speed,” even precisely measurable speed,
does not suffice to explain motion. For example, the basis of understanding
planetary orbits lies in realizing that, though a planet’s speed remains roughly
constant, its velocity is continuously changing as it follows its curved path.
By using the Galilean concept of “velocity,” Newton was able to explain
how gravity is the force responsible for the changing velocity of the planets,
and that the planets are constantly accelerating towards the sun, even though
their distance from the sun remains almost constant. “Acceleration” is a still
more abstract concept, defined not as a change in speed but as a change
in velocity. Only this conceptual hierarchy makes it possible to understand
such seemingly paradoxical facts as that a ball thrown upwards, though
losing speed as it rises, is accelerating downward from the moment it leaves
one’s hand, in accordance with Newton’s Second Law of motion (F = ma).
Newton’s three laws of motion, as simple as they are, cannot be grasped
until one has grasped, then related, the concepts of direction, speed,
velocity, and acceleration. These concepts do not lie flat, in effect, on a plane;
instead, they have a hierarchical structure: a necessary order of learning.
“Velocity” is a cross-classification and cannot be grasped before “speed”
and “direction.” Likewise for “acceleration,” which is a change in velocity.
If one forgets the actual meaning of “velocity,” and attempts to use it as
a synonym for “speed,” one’s concept of “acceleration” will be sabotaged.
A higher-level concept depends for its meaning on the prior concepts
used in its formation. Deprived of that base, a concept becomes a mean-
ingless sound. That is exactly the fate of “stolen concepts” — e.g., of “illegal”
in “The U.S. Constitution is illegal.”
The hierarchy of concepts is not metaphysical but epistemological.
To be sure, a concept’s place in the hierarchy is partly determined by meta-
physically given facts — e.g., by the scope of differences that must be embraced
— but what makes the scope of differences matter is “the crow epistemology.”
Certain differences are too big, or too subtle, to be taken in perceptually
and register as perceptual similarities, even when juxtaposed with other
things sharing a CCD. The differences between civilians and military men
are real, but too wide-ranging to be taken in at the purely perceptual level.
At the other end of the range, the subtle differences between two varieties of
beetle can be recognized by the prepared mind of the naturalist, but not by
a child of two who is just forming the concept “bug.”
Earlier, I gave the example of “organism,” which embraces differences too
wide-ranging, in terms of perceptual features, for a child to grasp without the
 A bstract vs . C oncrete 149

aid of the condensing power of prior concepts, such as “animal” and “plant.”
But metaphysically, apart from the cognitive requirements of man’s mind,
there are no “levels of existence.” Apart from man’s perspective, organisms
exist, animals exist, and Lassie exists, period. The hierarchy of concepts is
epistemological, not metaphysical.
The hierarchy of concepts has another source besides “the crow epistem­
ology”: the nature of man’s senses. The concepts “ultraviolet” and “molecule”
cannot be formed directly from perception — because neither can be perceived.
Thus, this hierarchy stems from facts about the identity of man’s conscious­
ness, a factor that plays a crucial role throughout epistemology.

Abstract vs. Concrete

The fact that there are conceptual subdivisions (“narrowings”) alters our
understanding of what “abstract” means. Previous philosophers, counte­
nancing only widenings, equated “more abstract” with “more general.”
But concepts also become more abstract when they make subdivisions
of an existing concept. Consider “crimson” and “red.” Crimson is a specific
shade of red, so “crimson” is less general than “red.” But a child would have
to learn “crimson” as a subdivision of “red.” “Crimson” is thus both more
abstract than “red” and less general.
This point is more obvious with narrowings that cross-classify, like
“employee,” which is a cross-classifying subdivision of “man” (a man who
works for another). “Employee” is more abstract than “man,” but less general.
Grasping “employee” requires a host of prior conceptual knowledge — which
is why a two-year-old is not prepared to grasp “employee.”
Abstractness, then, does not equal generality. Generality is a matter
of scope — the number of units encompassed by the concept. Abstract-
ness is a matter of what has to be conceptualized on the basis of what.
As Peikoff notes, “Concepts, therefore, differ from one another not only in their
referents, but also in their distance from the perceptual level.” [OPAR, 91–92]
One can move further away from the perceptual level in either direc-
tion — going “up” (widening) or “down” (narrowing). Starting from the first-
level concept, “man,” one gets more abstract both when one widens to form
“animal,” and when one narrows to form “employee.”
The hierarchy of concepts soon becomes quite complex. Not only can one
move “up” to a wider concept, or “down” to a narrower concept, one can,
 H OW W E K N OW • 4 : H igher- L evel C oncepts 150

so to speak, reverse directions — moving “up” from a concept that one

reached by moving “down,” and vice-versa. A few examples will illustrate the
rich complexity of conceptual development.
“Female” is an example of moving “up” on the basis of having previ-
ously moved “down.” “Female” is a widening. It integrates such concepts
as “woman,” “mare,” and “hen,” to reach the idea of the female of some but
any animal species. But “woman,” “mare,” “hen,” etc. were each reached by
narrowing. “Woman” is a narrowing of “human being,” “mare” of “horse,” and
“hen” of “chicken.” So, “female” is an integration (moving “up”) from a set
of narrowings (each reached by moving “down”). Adding to the complexity,
the process of reaching “female” also depends on having as an automatized
context the wider concept “animal,” and, for the full concept “female,” such
concepts as “reproduction.”
Conversely, here are some examples of moving “down” on the basis of
having previously moved “up.” From “animal,” a widening, one moves “down”
to make subdivisions, such as “predator.” Or, from “number,” a widening of
“one,” “two,” “three,” etc., one moves “down” to form “even” and “odd.”
In forming cross-classifications, an earlier concept is narrowed by the
addition of another characteristic; the added characteristic must have already
been con­cept­ualized.70 The concept of the added characteristic may itself
have been reached by either a widening or a narrowing. For instance, take
the case of narrowing “man” to reach “artist.” The concept “artist” narrows
“man” by reference to “art,” which is a concept reached by widening from
the individual arts — painting, literature, music, etc. “Collie-owner” narrows
“man” by reference to a certain subcategory within “dog” plus “ownership,”
a concept that depends upon many earlier widenings and narrowings.71
As a result, beyond the first, simple steps in the development of concepts,
the hierarchy of concepts becomes an increasingly complex affair, involving
“two interacting directions,” in Rand’s phrase quoted earlier. The hierarchy
is not at all a linear progression, not even if we allow for proceeding in two
directions — “up” and “down” — from the perceptual level. The hierarchical

70 The added characteristic may initially be grasped only perceptually, as when one notices
a similarity among personality types or musical styles; but the completion of the concept-
forming process requires a definition, and that, of course, requires an explicit, conceptual
grasp of the characteristic.
71 Strictly speaking, a concept has to be one word, but in a few of the examples in this
book I ignore this requirement — or fudge it with a hyphenated word, like “collie-owner.”
Interestingly, English is moving in the direction of eliminating hyphens and forming one
longer word, as with “healthcare” and “website.”
 A bstract vs . C oncrete 151

dependency of some concepts on others is real and unavoidable, but this

hierarchy is not a simple sequence, like the number line, which extends in
two directions (positive and negative).
Hierarchy has sometimes been analogized to structures — to skyscrapers
or pyramids. But the most accurate analogy for the hierarchy of concepts
is the suspension bridge. Some of the higher parts of a suspension bridge
serve to hold up the structures below them; other parts do the opposite,
supporting what lies on top of them. But every part of the suspension bridge
is subject to and works in relation to the force of gravity, just as every concept
in the hierarchy is subject to the necessary order of learning. Just as a given
part of a suspension bridge has support but also supports other parts,
so a concept in the hierarchy may have prior concepts that make it possible,
while also making possible grasping other concepts that rest on it.

And just as any bridge part will fall unless it is supported, ultimately,
by the ground, so any concept which does not reduce through intermedi-
ate concepts back to the perceptual level will fail to function cognitively.
Ungrounded “concepts” are mere sounds, without a cognitive link to the facts
of reality. Rand calls them “floating abstractions.” (Methods of preventing
floating abstractions will be discussed in Chapter 7.)
In the field of abstraction from abstractions, certain concepts require
special attention: concepts of characteristics, concepts of consciousness, and
axiomatic concepts.
 H OW W E K N OW • 4 : H igher- L evel C oncepts 152

Concepts of Characteristics
The first concepts a child forms are concepts of entities — e.g., “dog,” “table,”
“cookie.” These concepts are formed directly from perception, rather than
requiring prior concepts, and perception is geared toward discriminating
entities from each other. Though we also perceive the attributes and actions
of entities, perception does not discriminate attributes or actions from
entities. When we perceive a big dog barking, the dog is given as discrimi-
nated from the ground on which it is standing (and from every other entity in
the scene), but we are not given any discrimination of the dog’s size from the
dog; nor are we given any discrimination of the dog’s action of barking from
the dog. Yet, after a period of development, we are of course able to form
the concepts “big” and “barking.” By what means do we do this? How do we
form concepts not of entities but of their characteristics: their colors, shapes,
locations, what they are doing, and what they can do?72
This is an important question, because concepts of characteristics are
our means of identifying the nature of a thing, breaking down what is,
perceptually, an unanalyzed whole. It is one thing to see a red ball rolling by,
it is quite another to isolate its color, shape, or action and to name it. Concepts
of characteristics offer this analysis and identification, enabling us to gain
explicit, conceptual knowledge of what things are and do. Doing so enables
us to connect the properties of a thing to its actions — i.e., we can identify
causal factors.73 E.g., the ball’s round shape is necessary for it to roll, but its
red color is not.
By breaking down perceived entities into separate characteristics and then
identifying the action-consequences of each, man has been able to harness
the power of wind, water, fire, sunlight, and petrochemicals.
For instance, man uses concepts of characteristics to grasp that wind
pushes in a certain direction; he uses that analytical knowledge to develop
the sailboat. Similarly, he observes that flowing water pushes things, and
uses that analytical knowledge to invent the waterwheel. Even at a more
primitive stage, in learning how to tame fire, he must recognize that fires are
hot, that fire has fuel, that different fuels burn differently, that a fire can be
starting, steady, or dying. This is the kind of analytical knowledge that

72 I include “actions” under “characteristics” to avoid having to repeat “characteristics

and actions.”
73 “The law of causality is the law of identity applied to action. All actions are caused by
entities. The nature of an action is caused and determined by the nature of the entities
that act; a thing cannot act in contradiction to its nature.” [AS, 1037]
 C oncepts of C haracteristics 153

permits primitive man to progress from fearing fire to taming it, to mak-
ing wood fires, inventing torches, then candles, then ovens and foundries.
None of this would be possible if man were restricted to the pre-­analytical
concept of “fire” as being just “that kind of thing.” Without concepts of
characteristics, man would be restricted to “Here is wind” and “There is
fire,” which is in itself hardly any advance over the perceptual associations
formed by animals. Concepts of characteristics make possible man’s mastery
of nature — and of himself.
In view of the inestimable value of causal knowledge, it is imperative
to understand how characteristics are isolated and conceptualized.
Concepts of characteristics are formed by the same basic method used
to form concepts of entities; the difference here pertains to what is being
conceptualized, not directly to the means of doing so. Nevertheless, there
are some wrinkles — and some questions — that arise regarding concepts
of characteristics.
To form a concept of an entity, we contrast two or more instances of
the entity with a foil — e.g., some tables vs. a chair. Likewise, to form the
concept of an attribute, we contrast this and that instance of the attribute
with a foil — e.g., two or more shades of blue vs. a shade of green. To form
the concept of an action, we contrast this and that instance of an action
with a foil — e.g., two or more instances of a thing moving vs. being at rest.
And, as with entity-concepts, concepts of characteristics are formed by
measurement-omission, on the “some but any” principle, and are integrated
into a new mental unit by means of a word.
We form higher-level concepts of characteristics just as we do in the case
of higher-level concepts of entities.74 “Blue” is first-level, within attribute-
concepts; “color” is a widening; “indigo,” and “ultramarine” are narrowings.
Narrowing by cross-classification is exemplified by “pastel blue,” if we allow
two words to count as a concept, since “pastel blue” stands for those shades

74 A technical issue arises regarding “levels,” because the term has two senses. In one sense,
only concepts of entities are “first-level”: only entity-concepts presuppose no prior
conceptualization. Since concepts of characteristics presuppose concepts of entities,
concepts like “blue” are not “first-level” in this sense. But, in another sense, “first-level”
denotes concepts that do not integrate or subdivide any prior concepts, and in this sec-
ond sense “blue” is first-level: it conceptualizes what is directly perceivable. Accordingly,
these concepts need no validation or checking (there’s no such thing as getting “blue”
wrong). Concepts like “blue,” “round,” and “moves” are part of the incontestable base to
which more abstract concepts must be reduced and against which their validity is to be
judged. As such, they could be called “reductively first-level.”
 H OW W E K N OW • 4 : H igher- L evel C oncepts 154

that contain a considerable amount of white.75 “Pastel” itself is a narrowing

of a previous widening (“color”).
Colors are attributes, but the mechanics of concept-formation are
the same in the case of characteristics in categories other than “attribute.”
For concepts of actions, the first-level includes something surprisingly wide:
“motion.” The motion of things is given directly, in vision, hearing, and touch.
In conceptualizing “motion,” the child has, in one way, an easier task than
in conceptualizing attributes: he can observe one and the same entity when
it is moving and then when it is not, as in observing his mother walking,
then standing still. He also observes the same alternative with other entities:
his ball rolls then stops; he waves his hand then holds it still; a car, a bug,
a bird — all move and then don’t move. By omitting the measurements
of what moves — whether it is big like the car or small like the bug, whether
it is round like his ball or shaped like his hand — and by omitting the
measurements of the motion (fast, slow, toward him or away, rotating
or translating), he forms the concept “move” — which, in English, is likely
to be first captured as “go.”
The concept “motion” has many, many narrowings. One narrows “motion”
by narrowing one or more measurement-ranges that were left open in form-
ing “motion.” For example, “rising” is moving upwards, “slowing” is decreas-
ing in speed, “walking” is a certain kind of movement of an animal with legs,
“sinking” is falling in water.
And the narrowings can be further narrowed. An impressive array of
English verbs use subtle differences to narrow “walking.” Consulting several
thesauruses, I found 20 different narrowings of “walking.” Here are
a few: ambling, striding, pacing, sashaying, strolling, limping, parading.
All these different ways of walking differ in the measurements of the
component aspects of walking. (Note that, hierarchically, it would be impos-
sible to form the concept of, say, “strolling” before the concept “walking.”)
The concept “parading” is a cross-classification, narrowed not only by the
features of the motion itself, but by the purpose of parading: to display one-
self to others.
Concepts of relationships are the most complex type of concepts of char-
acteristics, since these concepts involve not just relating things (as all con-
cepts do) but also isolating the relationship itself. Spatial relationships are the
simplest case, since they are given in visual perception. Examples of concepts

75 I do think there are some concepts which, in English, are held as two or even three words.
“Black hole” and “stolen concept” are examples. German routinely combines words into
one compound word.
 C oncepts of C haracteristics 155

that are first-level within this category are: “in,” “on,” “beneath,” “near,”
“behind,” etc. The words for these, as well as for temporal relationships, are
prepositions:

Prepositions are concepts of relationships, predominantly of

spatial or temporal relationships, among existents; they are
formed by specifying the relationship and omitting the measure­
ments of the existents and of the space or time involved
— e.g., “on,” “in,” “above,” “after,” etc. [ITOE, 17]

A widening of “on,” “in,” “beside,” etc. would be “with” (in one sense).
A narrowing of “on” would be “atop” (which excludes being on the side or
bottom). Later come myriad cross-classifications, from the relatively simple
“astride” to such complex (non-spatial) relationships as “marriage.”
In regard to any concept of a characteristic, we must distinguish between
an adjective and the abstract noun, which refers to the characteristic as
such. “Small” is less abstract than “smallness.” “Small” is an adjective, a word
that describes; “smallness” is a noun, a word that names that attribute.
Likewise, “sharp,” “flexible,” and “polite” describe, but “sharpness,” “flexibility,”
and “politeness” name the attribute involved. (Verbs can also be nominalized,
which creates gerunds, as in “Walking is good exercise.”)
In English, a suffix, such as “ness,” or “ity,” is normally used to distinguish
the nominalized adjective from the adjective. But we do not use such suffixes
in the case of color-concepts: we do not say “greenness,” just “green,” when
we are making that color the subject of our thought. The shrinking of
“greenness” to “green” is exploited in the following equivocation, which helps
concretize the difference between an adjective and the corresponding noun:

Grass is green.
Green is a color.

Grass is a color.

— as if grass were an attribute. The equivocation becomes clear if we force

the green-greenness distinction:

Grass is green.
Greenness is a color.

— which is not a syllogism, and no conclusion follows.

 H OW W E K N OW • 4 : H igher- L evel C oncepts 156

Note that the equivocation cannot be slipped by us when the lan-

guage does make the noun-adjective distinction, as in the case of “long”
and “length”:

This rope is long.

Length is an attribute.

To produce a syllogism, the second premise would have to read: “Long is

an attribute” — which is neither proper English nor meaningful.
Going from a verb (e.g., “move”) to its nominalized form (“motion”)
offers certain advantages. The nominalized characteristic can be made the
object of separate study, as if it were an entity — as in studying “motion”
or “immigration” (from “immigrate”), or “happiness” — which is in one way
the same as, and in another way quite different from, “happy.”
Any characteristic that rewards specialized study gets its own nominalized
form. Nominalized forms of characteristic-concepts enable us to direct
attention to the characteristic and its sub-characteristics, rather than being
distracted by its relations to the other attributes of its entity. Consider the dif-
ference between: “Whenever one object acts on a second object, the second
object reacts equally and oppositely on the first object” and “For every action
there is an equal and opposite reaction.” The second statement is far more
unit-economical, and it directs our attention to the action itself, in abstrac-
tion from the other characteristics of the objects that act and react.
Now let’s confront the intriguing question of how we isolate a characteristic
from the entity that possesses it, which we must do before we can conceptu-
alize the characteristic (as, e.g., an adjective).
In forming concepts of entities, such as “table,” we attend to characteristics,
but we do so perceptually: we selectively focus on characteristics, e.g., shape,
in order to classify entities. The child at this early stage is not conceptualizing
“shape” as a characteristic. Rand writes:

In the process of forming concepts of entities, a child’s mind has to

focus on a distinguishing characteristic — i.e., on an attribute — in
order to isolate one group of entities from all others. He is, there-
fore, aware of attributes while forming his first concepts, but he
is aware of them perceptually, not conceptually. It is only after he
has grasped a number of concepts of entities that he can advance
to the stage of abstracting attributes from entities and forming
 C oncepts of C haracteristics 157

separate concepts of attributes. The same is true of concepts

of motion: a child is aware of motion perceptually, but cannot
conceptualize “motion” until he has formed some concepts of that
which moves, i.e., of entities.
(As far as can be ascertained, the perceptual level of
a child’s awareness is similar to the awareness of the higher
animals: the higher animals are able to perceive entities,
motions, attributes, and certain numbers of entities. But what an
animal cannot perform is the process of abstraction — of mentally
separating attributes, motions or numbers from entities. It has
been said that an animal can perceive two oranges or two pota-
toes, but cannot grasp the concept “two.”) [ITOE, 15–16]

In thinking about how we form concepts of characteristics, we must take

care not to commit the fallacy of retroactive self-evidency. Having so long ago
automatized concepts like “white” and “round,” we might easily assume that
these characteristics are given to us directly in the same way that entities are.
But perception isolates entities from other entities, not characteristics from
other characteristics. As adults, when we look at a white, round table, it is
self-evident to us that the table has the color and the shape as characteristics;
but that was not self-evident in early childhood, before we had ever performed
the process of isolating those attributes from the entities that possess them.
What is “obvious” to a modern adult is not necessarily a “given” for the begin-
ning conceptualizer. What needs to be explained is the mechanics of the
process of abstracting characteristics from the entities that possess them.
At the entrance to the conceptual level, awareness of attributes serves
only as the means of distinguishing among entities; the perceptual focus on
an attribute is not, per se, the isolation of one attribute from all the others
that constitute the entity. But that isolation is precisely what is required for
conceptualizing the attribute. How is this isolation performed?
In principle, an isolation of one existent from others can be achieved by
one of several means. If the existent is an entity, it can be physically separated
from surrounding entities, as a single piece of gravel in a pile can be picked
up and inspected individually. Parts of entities can be physically separated
from the whole entity, as when one pulls a leaf from a tree. (Parts of entities
are themselves entities.)
Even without detaching it, a part can be isolated by attending to its spatial
position — by tracing around its edges with one’s eyes and fingers.
 H OW W E K N OW • 4 : H igher- L evel C oncepts 158

But attributes (and actions and relationships) are not parts of entities,
and they cannot be isolated by separability, location, or any physical means.
While physical means can be used to isolate a leaf from the tree it is part of,
no physical means can isolate from the tree its height, its age, or its action
of swaying.
To isolate a characteristic from the entity that possesses it, one must use not
a physical but a mental means: differentiation. Consciousness is a difference-
detector, and there are two basic types of differentiation that we can use to
isolate a characteristic from the entity.
The simplest form of isolating a characteristic occurs when the character­
istic in question may be present or absent in the same entity. The same
entity — the child’s mother — can be walking or still, standing or seated,
happy or sad. The very same blossom that was white yesterday is brown
today. In such cases, the child is aware that there is one entity that has
changed, so he will not use the difference to make a concep­tual subdivi-
sion (as if there were two types of entity: walking-mother vs. still-mother,
white-blossom vs. brown-blossom). The difference of which the child
is aware is a difference in a characteristic: his attention is drawn to the one
factor that changes, against the background of those that do not. (This repre­
sents the application to concept-formation of Mill’s Method of Difference,
which he formulated as a technique for isolating causal factors.76)
Next comes the stage of integration. When the child observes a similar
change in other entities (bread in the toaster goes from white to brown),
if he has already conceptualized the entity involved (bread), he is prepared to
differentiate what changes (color) from what does not (the other attributes
of the bread). Then he can omit measurement-variations within the brown
range, while retaining that range.
How is the isolation of a characteristic accomplished in cases of charac-
teristics that are permanent rather than sometimes present and sometimes
absent? The child can isolate the characteristic by considering a wide range of
entities having only one perceptual feature in common. For instance, a child
can notice the same color, blue, in many far-flung cases: blue shirt, blue berry,
blue hills, blue lake. What is the same and what is different across all these

76 In Mill’s wordy statement: “If an instance in which the phenomenon under investigation
occurs, and an instance in which it does not occur, have every circumstance in common
save one, that one occurring only in the former, the circumstance in which alone the two
instances differ, is the effect, or the cause, or an indispensable part of the cause, of the
phenomenon.” [Mill, III, VIII, 2]
 C oncepts of C haracteristics 159

cases? The sameness is in the one attribute, blue: it is constant while all the
other attributes vary.
Here the child is making a second-order differentiation: he is distinguish-
ing between what differs and what does not.77 In a first-order differentiation,
one distinguishes A from B; in a second-order differentiation, one distin-
guishes the differing from the non-differing. In a first-order differentiation,
the characteristic is the means; in the second-order differentiation, it is the
object, the thing being conceptualized. E.g., color is one means by which
a child differentiates blueberries from other berries, but then he is forming
“blueberry” not “blue.” It is the second-order differentiation that makes the
color into the object of his concept-forming process; it permits him to isolate
the blue color from the other attributes of the concretes, such as their size,
shape, solidity, and texture: these other attributes vary widely among the
shirt, the lake, etc. while the attribute of color does not vary. The similar color
stands out: it is the one thing that does not vary. The highlighted similarity
is in color, not in the type of thing having the color.
This second-order differentiation constitutes the equivalent, for concept-
formation, of Mill’s Method of Agreement: “If two or more instances of the
phenomenon under investigation have only one circumstance in common,
the circumstance in which alone all the instances agree is the cause (or effect)
of the given phenomenon.” [Mill, III, VIII, 1]
A child at this early stage of forming concepts needs to have automatized
the concepts of the various blue things — shirts, berries, etc. — to avoid being
swamped by all the differences among them. If he has already classified
shirts as one type of thing and berries as another, and so on for the other
concretes, he is prepared to grasp that, among these familiar, conceptualized
entities, there is one thing that is the same: they are all blue.
The second-order differentiation (the Method of Agreement) prevents one
from using attribute-differences to subdivide entities, as when a child uses
blue to subdivide berries into blueberries vs. others. Subdivision requires
a Conceptual Common Denominator (CCD), but the shirt, the lake, and
the berry are disparate things, which lack any obvious CCD. Consider the
different axes along which these concretes vary: the berry is small, the lake
is big; the lake is liquid, the berry and shirt are solid; the shirt is for wearing,
the others are not, and so on. All the salient attributes — except color — vary.

77 See Chapter 2 for a discussion of how the same kind of second-order differences are
exploited by the perceptual system.
 H OW W E K N OW • 4 : H igher- L evel C oncepts 160

Thus, among the characteristics that are differing, the child’s attention is
drawn to the one that does not differ: the blue color.
There is a further reason why entity-concepts must precede concepts of
attributes, and of the other kinds of characteristics. To be a characteristic
is to be something that characterizes an entity. Attributes are attributes of
entities, actions are actions of entities, relationships are relationships among
entities, etc. To grasp “blue” is to grasp it as something that entities can be.
If a child has not yet learned that “blue” implies a blue thing, then whatever
sounds he may utter, he does not yet have the concept “blue.”
In fact, a child does not have the full concept of “blue” until he open-
ends the classification. He needs to understand, implicitly not explicitly, the
wide range of phenomena that can be blue: a physical object, like a shirt,
a living thing, like a berry or a fish, a liquid, the sky, etc. He does not have
the concept “blue” if he knows as “blue” only his shirt; he does not have “fast”
if he knows only cars as things that move fast. This is another way in which
concepts of characteristics require the prior formation of entity-concepts.
Once the child has mentally isolated the attribute of being blue, he can
omit two categories of measurements: 1) the measurements pertaining to
different shades of blue, and 2) the measurements of all the other varying
characteristics of blue things (size, shape, solidity, etc.). The omission of the
second category of measurements is what enables him to apply “blue” to each
of the wide range of things that are blue.
An interesting variant of the Method of Agreement can be used to isolate
a given characteristic, even in a primitive state of knowledge: the metaphorical
use of entity-concepts as adjectives.
Imagine how a primitive man might form the concept “tiger lily.” Suppose
he is a jungle dweller of 50,000 years ago. Assume that he possesses many
concepts of entities, such as “lion,” “tiger,” “flower” as well as subdivisions
of “flower,” such as “lily,” “rose,” etc. but has no concepts of characteristics.
One day he notices a new flower that clearly is a lily but is orange with black
stripes or spots. It reminds him of tigers. In a burst of inventiveness, he calls
it a “tiger lily.” Though “tiger lily” is still an entity-concept, his metaphor
makes “tiger” into an implicit adjective, just as if he had called it a “tigerish”
lily (“tigerish” being an adjective).
In creating the metaphor “tiger lily,” he is regarding the coloring of these
lilies as being similar to that of tigers, and he has incorporated that similarity
into the name: “tiger lily.” Because tigers and lilies are so different, he is not
tempted to unite “tiger” and “lily” into a wider entity-concept (as he does
when he integrates “tiger” and “lion” into “feline”). What he has done is to
 C oncepts of C haracteristics 161

form a subdivision of “lily,” but the means of doing so is by the metaphor that
spotlights the characteristic shared by otherwise disparate things.
If he later encounters many things that are tiger-striped, the same
kind of subdivision would be applied to other entity-concepts. Eventually,
that would lead to the development of “tigerish,” “tiger-like,” or some other
linguistic method of distinguishing nouns from adjectives.
Of course, not many things are in fact tiger-striped. So I now switch
the example to “orange.” “Orange” originally meant the fruit, not the color.
When someone early on got the idea that his friend’s hair was like the fruit,
he probably called it “orange hair” — that being simply a clever metaphor,
like “apple cheeks” or “hook nose.” Later, as all sorts of things having that
color began to be likened to the fruit, this use of “orange” died as a metaphor
and became an adjective (“orange-like,” in effect).
Note that many of the subtler shades of color are still named for entities:
turquoise, aqua, burnt umber, amber, rose, pearl. Color-concepts derive from
an awareness of the similarity between one entity with a distinctive color and
another entity, already conceptualized as a different kind of entity, that has
the same color.
Again, after many, many instances, in different contexts, of using a noun
as an implicit adjective (rose lips, rose cloud, rose pebble), someone reached
the level of differentiating “rose” used in this way from “rose” used in the
literal sense, to denote the flower, and some linguistic differentiation was
made — e.g., “rosy.”
The noun used metaphorically (e.g., “rose” in “rose lips”) could have
served the function of adjectives before adjectives (“rosy”) were devised.
(As I noted earlier, one learns the adjective, “rosy,” long before one learns its
nominalized form, “rosiness.”)
Thus, by this “tiger-lily” route, concepts of attributes can be formed without
beginning by abstracting the attribute from the entity, but, instead, by initially
combining one entity-concept with another entity-concept, as a metaphor.
The metaphorical noun-combinations focus attention on characteristics, and,
by their nature as words, provide the means of eventually denoting charac-
teristics as such: the invention of adjectives.
Here’s another example. The origin of “red” in English turns out to be
blood — “rud” in Old English. Imagine primitive men continually using
the blood metaphor: blood leaf, blood cloud, blood hair. Their repeated
use over time of the same word, “blood” (or “rud”), focuses attention on the
linguistic similarity, and eventually the same word is sensed to have two dif-
ferent functions: to refer to literal blood and to compare something else to
 H OW W E K N OW • 4 : H igher- L evel C oncepts 162

blood. This sensed difference then gets expressed linguistically: “rud-ish” or

“ruddy,” which became “red.”
One can also use the “tiger-lily” method to form concepts pertaining to
other characteristics, for instance concepts of actions. In modern English,
many action-concepts have been derived from entity-concepts — e.g.,
“to ski,” “to roof ” (a house), and most recently “to mouse” (move a cursor
with a mouse). It is likely that more ancient concepts, like “to wave” were
similarly derived from nouns (to act like a water wave).
I have outlined three ways to use differentiation to isolate a character-
istic: by differentiating the same individual entity when it has or lacks the
characteristic (the Method of Difference); by a second-order differentia-
tion of those characteristics that vary and those that do not (the Method of
Agreement); and by the tiger-lily method of combining entity-concepts
metaphorically (another form of the Method of Agreement). The use of
one or more of these means enables a beginning conceptualizer to isolate
a characteristic, which then permits him to omit its varying measurements
and form the appropriate concept.

Concepts of Consciousness
On this base, we are prepared to analyze a much more advanced — and more
powerful — set of higher-level concepts: those pertaining to consciousness.
Some examples are: “hear,” “remember,” “desire,” “think,” and “understand.”
Since consciousness, at root, is an activity, the primaries in this category are
concepts of mental actions78­ — notably: “perception,” “evaluation,” “emotion,”
“thought,” “reminiscence,” and “imagination.” [ITOE, 30]
The formation of concepts of consciousness requires introspection
and is considerably more difficult than the formation of concepts of external
existents. Mental phenomena do not have the stability and the observer-­
independence of physical objects. In investigating a physical object, say
a bracelet, one can take hold of it, methodically scrutinize every part of
it, manipulate it, turn it this way and that to observe it from different angles.
In introspecting a thought or an emotion, one cannot use such tactics;

78 This is not to imply that these processes are “purely” mental in some Cartesian sense:
though irreducible qua conscious, all conscious actions have a neural base (see Chapter 1),
and emotions have a bodily component.
 C oncepts of C onsciousness 163

only some partial, rough substitutes are available. Mental actions and states
are continuously changing; as soon as one begins to scrutinize a mental
content, it changes, and it will fade out unless continually “refreshed.”
Moreover, conscious states are “transparent to the user,” in computer jargon;
it takes a special introspective act to make the means of awareness of the
world into an object of awareness in its own right.79
In this difficult and largely uncharted territory, I will limit myself
to explaining how concepts of consciousness fit the Objectivist theory of
concept-formation.
The central question for the Objectivist theory here is how the formation
of concepts of consciousness involves measurement-omission within a CCD.
Take a concept that is first-level within this category: “seeing.” How does
a child form the concept “seeing”? The concept “seeing” abstracts the action
of consciousness from its varying contents, i.e., the objects seen. Contra the
Realist theory, one cannot subtract the things seen from the seeing of them.
As I now look out at the cityscape beyond my apartment window, I cannot,
by any act of selective attention, separate the scene from my viewing of it.
If I attempt to ignore the content — the buildings and sky that I see — I am
not left with any residue to consider. Again, abstraction is not subtraction but
differentiation: if I close my eyes, the action of seeing, the visual perception,
ceases. By deliberately shutting and re-­opening his eyes, even a toddler can
differentiate the seeing from the things seen, the action of his consciousness
from its content.80 One can also differentiate the seeing of an object from the
feeling of touching it, the hearing of the sounds it makes, etc., though these
differentiations are more subtle than the dramatic difference resulting from
closing and re-opening one’s eyes.
As in isolating a characteristic from an entity, the child isolates the action
of his consciousness from its content by differentiation, not by Realist

79 It is difficult to do much introspection “in real time,” because that requires splitting
one’s focus between consciousness and self-consciousness, between the intro­spection and
the extrospective activity being introspected. Most introspection is retrospection, but
retrospection is also not as stable, manipulable, and repeatable as extrospection.
80 An interesting question is whether an animal, when it closes and opens its eyes, is aware,
implicitly, of the difference between seeing and not seeing. I doubt that it is; probably, the
animal simply sees, then does not see. The child, unlike the animal, can voluntarily, delib-
erately shut and re-open his eyes, as an act of agency. The child’s level of control over his
body is of a different order from that of an animal’s, and, I believe, gives the child a sense
of self not possible to an animal. A child’s sense that he caused the changed experience
may be required for his isolation of the action of his consciousness from its content.
 H OW W E K N OW • 4 : H igher- L evel C oncepts 164

subtraction. As discussed above, the differentiation is, in effect, an implicit

use of Mill’s Methods.
But do acts of seeing differ quantitatively? When acts of seeing differ,
do they differ in measurement? Regarding the measurable aspects of
consciousness in general, Rand writes:

In the realm of introspection, the concretes, the units which

are integrated into a single concept, are the specific instances
of a given psychological process. The measurable attributes of
a psychological process are its object or content and its intensity.
The content is some aspect of the external world (or is derived
from some aspect of the external world) and is measurable by
the various methods of measurement applicable to the external
world. The intensity of a psychological process is the automati-
cally summed up result of many factors: of its scope, its clarity,
its cognitive and motivational context, the degree of mental
energy or effort required, etc.
There is no exact method of measuring the intensity of all
psychological processes, but — as in the case of forming concepts
of colors — conceptualization does not require the ­knowledge
of exact measurements. Degrees of intensity can be and are
measured approximately, on a comparative scale. For instance,
the intensity of the emotion of joy in response to certain facts
varies according to the importance of these facts in one’s hierarchy
of values; it varies in such cases as buying a new suit, or getting
a raise in pay, or marrying the person one loves. The intensity of
a process of thought and of the intellectual effort required varies
according to the scope of its content; it varies when one grasps
the concept “table” or the concept “justice,” when one grasps that
2 + 2 = 4 or that e = mc2. [ITOE, 31]

Mental intensity is not a single attribute but a composite, an “automatically

summed up result” of several, more specific attributes. The analogy from the
external world is size: size is the product of a thing’s length, width, and thick-
ness. Just as a body gets bigger when it increases in any of these dimensions,
so a conscious state is made more intense (“bigger” in internal, psychological
terms) when it increases in clarity, scope, emotional impact, etc.
Let’s concretize this by taking the case of seeing. The intensity of seeing
varies in regard to: clarity, acuity, time, attention, and purpose (and prob-
 C oncepts of C onsciousness 165

ably with other factors). One sees a given object less clearly when one’s eyes
are focused on something closer than that object, when one is looking at it
through fog or smoke, or when the object is seen with peripheral rather than
foveal vision. Acuity, in the sense of the amount of detail seen, varies with
distance, with the condition of one’s eyes, and with lighting conditions. Time
makes the difference between glancing, gazing, staring. More or less attention
can be paid to what one sees, and to seeing vs. listening, feeling, tasting, etc.,
in accordance with one’s purpose at the time.
Consider the wealth of conceptual subdivisions of “seeing” that have
been formed to capture sub-ranges within the above axes of measurement.
Here are some, listed in alphabetical order: descry, espy, gawk, gaze, glance,
glimpse, look, ogle, peek, peer, scan, scrutinize, stare, watch. The measurements
used in making these subdivisions are only approximate, but precision here
is irrelevant: the child need observe only a variation in degree — in the
more and the less. And these subdivisions do exhibit that kind of variation.
Scrutinizing differs from glancing in being more intense in time, attention,
and purposefulness. Ogling differs from gazing in being more emotionally
involved and attentive. Descrying differs from gawking in being more
purpose­ful, more discrim­inating, and far more acute. Scanning differs from
merely looking by being more purposeful, more systematic, and less static
(scanning involves a systematic movement of one’s gaze).
The point of discussing all these types of seeing is to show that instances
of seeing differ in their intensity along the various axes of measurement.
To conceptualize “seeing,” a child needs to become aware of the similarity
among different acts of seeing, but he does not need to know explicitly how to
measure them nor what the different elements of intensity are. Different
instances of seeing, despite their variations in intensity, are experienced
as similar when contrasted to what goes on in darkness or with eyes shut,
or when contrasted to instances of hearing.
After one has separated the seeing from the thing seen, or more broadly has
separated any action of consciousness from its content, one conceptualizes
that action in the same way that one forms any other concept:

The formation of introspective concepts follows the same

principles as the formation of extrospective concepts. A concept
pertaining to consciousness is a mental integration of two or
more instances of a psychological process possessing the same
distinguishing characteristics, with the particular contents and
the measurements of the action’s intensity omitted — on the
 H OW W E K N OW • 4 : H igher- L evel C oncepts 166

principle that these omitted measurements must exist in some

quantity, but may exist in any quantity (i.e., a given psychological
process must possess some content and some degree of intensity,
but may possess any content or degree of the appropriate cat-
egory). [ITOE, 31–32]

The widest categories — entity, action, attribute, relationship, etc. — apply

to consciousness in much the same way that they apply to external existents.
In the external world, the basic category is “entity.” While there are no
actual entities residing in consciousness, the nearest mental counterparts to
entities are products of consciousness. This category includes such things as
concepts, knowledge, syllogisms, sciences, memories, decisions — each of
which exists as a permanent, recallable unit. I do not regard particular sights
or feelings as falling into the category of products of consciousness: although
they do result from a process, they are transient states rather than stable,
retained, usable, entity-like packages.
Conscious processes and states have attributes — for instance, they are clear
or vague, pleasant or painful, general or specific — and there is the umbrella
attribute of intensity. Actions of consciousness are, as mentioned earlier, the
primaries in this field (products of consciousness are products of some mental
action). There are relationships among aspects of consciousness — e.g., clearer
than, more pleasant than, derivative of, contradictory to — and relationships
to reality — e.g., truth and reference. There are capacities of consciousness
— e.g., vision, intelligence, or any mental ability. There are states or conditions
of consciousness — e.g., confusion, alertness, dejection, concentration.
In chapter 4 of ITOE, Rand discusses some of the complex ramifications
of concepts of consciousness. She covers such higher-level concepts as “concepts
of method” (e.g., “logic”) and “composite concepts” (e.g., “marriage,” “trial”),
which integrate concepts of consciousness with existential concepts. She also
provides a discussion of grammatical concepts, such as “but” (in my view,
not necessarily hers, these concepts are instructions to consciousness, rather
than being concepts that name conscious processes).
Concepts from the value-realm, such as “love,” are subject to what Rand
calls “teleological measurement.” Teleological measurement involves using
a standard of value to rank the degree to which things serve as means to
an end, with one’s life as the ultimate end, and the time devoted to that
end as the CCD. The reader is directed to her chapter for more discus-
sion on all these topics; here, my purpose has been limited to showing
 A x iomatic C oncepts 167

that concepts of consciousness are formed by the same type of processes

— differentiation, integration, and measurement-omission — that is used to
form existential concepts.
As we consider the wide range of things that are subject to measurement,
it is perhaps necessary to observe that not every difference is a difference
in measurement. For example, the difference between consciousness and
external existents is not one of degree. It is not the case that “being” and
“knowing” are two different amounts of something. Nor does an action
differ quantitatively from an attribute. To hold otherwise would be to
endorse some ludicrous form of Pythagoreanism, in which everything
reduces to numbers. The Objectivist theory does not imply that there is
one common unit of measurement, or CCD, in terms of which all existents
have their place, as if “running” were 732 units and “big” 28.
The only common denominator among all existents is that they exist
and are what they are; but neither existence nor identity is a characteristic,
let alone a commensurable one.81

Axiomatic Concepts
In the first chapter, I discussed the fact that axiomatic concepts are prior
to axiomatic propositions. The primary axiomatic concepts are: “existence,”
“identity,” and “consciousness.” (There are a few other axiomatic concepts,
the most noteworthy being “entity” and “action” or “change.”) Axiomatic
concepts warrant special discussion here: since they are in one sense first-
level and in another sense higher-level, they constitute a special case.
The sense in which axiomatic concepts are first-level concepts is that the
facts they integrate are given in perception. Indeed, these facts are present
in every act of awareness, which is why they are axiomatic. Every act of
awareness consists of being conscious of the identity of something that exists.
From the newborn’s first sensory impression, he is aware; he is aware
of something; something that exists. There is no awareness that is of an
identity-less nothing. Even a dream, which isn’t awareness of the world,
is a state of consciousness that has content; that content exists (as images
or as a mental state) and is what it is (e.g., a dream that one is flying is some-
thing different from a dream that one is walking on the ground).

81 For a discussion of how one forms such basic concepts as “attribute” and “action,”
see ITOE, 274–276.
 H OW W E K N OW • 4 : H igher- L evel C oncepts 168

As aspects involved in every experience, the axioms are implicit from

the start of a consciousness, which apparently begins even before birth,
in the womb. And from the first sensory experience, there is nothing further
to be learned about what it is to be conscious, or what it is for there to be
something, or what it is for the something to be something. The concepts of
existence, identity, and consciousness are first-level in this sense: the facts
they designate are perceived or experienced directly.
However, axiomatic concepts are only implicit in experience. To identify
them explicitly, to conceptualize them, requires a sophisticated develop-
ment. Long before knowing the abstract terms “existence,” “identity,” and
“consciousness,” we use the simple words “is,” “this,” and “know.” But to iso-
late these basic, omnipresent facts and name them abstractly, we must have
a plethora of earlier concepts — plus extensive familiarity with the process of
conceptualization. We saw above how forming the concept of “consciousness”
requires differentiating action from content; it also requires omitting large
sets of measurements: “Whatever the size, shape, color, etc. of what I see,
I see it when my eyes are open but not when my eyes are shut.”
By contrasting seeing with not seeing, hearing with not hearing, etc., one
can generalize to reach the idea of consciousness as such. But what a man
cannot do is project what total unconsciousness would be like for him. For in
that condition, there is no mental content — and no him.
Just as a particular action of consciousness, seeing, can be contrasted to its
cessation, so a particular thing’s existence can be contrasted with its absence,
when that particular thing no longer exists. For instance, George Washington
no longer exists. Or, after a log is burned, though the ashes remain, the log is
gone; it exists no more. From such examples, it is easy to form the idea of any
particular thing’s existence or non-existence.
But the axiomatic concept of “existence” goes much further: it refers to the
totality of that which exists — just as the concept of “consciousness” refers to
the totality of one’s experience. The totality of one’s consciousness does not
cease when one shuts one’s eyes, and the totality of existence — the universe
— does not go out of existence when a log is burned or a particular man dies.
How, then, do we form the full concept of “existence,” the concept that
Rand uses in stating “Existence exists”? The full concept is reached as the
final step in a series of progressive widenings. E.g., one goes from: all the
dishes, to all the man-made objects, to all the objects — to all the everything.
This final, unrestricted use of “all” is not based on any process of differentia-
tion, but none is required. Existence has no contrary: there is no nothing;
 A x iomatic C oncepts 169

what is not, is not. The process requires no differentiation because, after one
can undertake a deliberate, self-conscious process of widening, the question
naturally arises: “What is the widest category of all?”
By a process of progressive widening, one can, without any differentiation,
arrive at the “integration of all existents” that Rand describes:

Since axiomatic concepts are not formed by differentiating one

group of existents from others, but represent an integration of
all existents, they have no Conceptual Common Denominator
with anything else. They have no contraries, no alternatives.
The con­trary of the concept “table” — a non-table — is every other
kind of existent. The contrary of the concept “man” — a non-man
— is every other kind of existent. “Existence,” “identity” and
“conscious­ness” have no contraries — only a void. [ITOE, 58]

What measurements are omitted in forming axiomatic concepts?

The measurements omitted from axiomatic concepts are all the

measurements of all the existents they subsume; what is retained,
metaphysically, is only a fundamental fact; what is retained,
epistemologically, is only one category of measurement, omitting
its particulars: time — i.e., the fundamental fact is retained inde-
pendent of any particular moment of awareness. [ITOE, 56]

For instance, one’s awareness of a table occurs at some particular time,

but the table exists across time, independent of one’s consciousness of it.
This is the primacy of existence (see Chapter 1). Consciousness perceives
what exists; it does not create its objects. (The same is true of consciousness
itself: introspection does not create the mental activities that one is intro-
specting; self-consciousness does not create consciousness.)82
In omitting what Rand calls “psychological time-measurements,” axiom-
atic concepts serve an important function:

Axiomatic concepts are the constants of man’s consciousness,

the cognitive integrators that identify and thus protect its continuity.
They identify explicitly the omission of psychological time
measurements, which is implicit in all other concepts.

82 Time and space are usually considered a pair, but phenomena of consciousness, such as
thoughts, have no spatial location.
 H OW W E K N OW • 4 : H igher- L evel C oncepts 170

It must be remembered that conceptual awareness is the only

type of awareness capable of integrating past, present and future.
Sensations are merely an awareness of the present and cannot
be retained beyond the immediate moment; percepts are
retained and, through automatic memory, provide a certain
rudimentary link to the past, but cannot project the future.
It is only conceptual awareness that can grasp and hold the total
of its experience — extrospectively, the continuity of existence;
introspectively, the continuity of consciousness — and thus
enable its possessor to project his course long-range. It is by
means of axiomatic concepts that man grasps and holds this
continuity, bringing it into his conscious awareness and knowledge.
It is axiomatic concepts that identify the precondition of
know­ledge: the distinction between existence and conscious-
ness, between reality and the awareness of reality, between the
object and the subject of cognition. Axiomatic concepts are the
foundation of objectivity. [ITOE, 56–57]

In order to form a judgment about a thing’s objective nature — to judge

what something is, not just how it looks now — one must have grasped
the difference between existence and consciousness, at least implicitly.
Every judgment asserts that its subject exists and does in fact have the identity
ascribed to it. Even “I dreamed that I was flying” asserts that such a dream
did occur and was of flying, not of something else. Even a statement about
what seems to be the case or is wrongly believed to be the case assumes the
existence of the appearance or the wrong belief. To judge is to seek to identify
what something is. Thus, all judgments presuppose and use the (implicit)
concepts “existence,” “identity,” and “consciousness.”
Chapter 5 takes up the deeper question: what is the nature of a judgment?
To answer that question, we must investigate the process by which we advance
from forming concepts to forming propositions.
 PROPOSITIONS
5
C oncepts are tools. The value of any tool lies in its use.
What concepts are used to do is to identify: to state in words what
something is.
The form in which we make conceptual identifications is the proposition.83
A proposition is the combination of two or more concepts into a single
thought, as in “The table is brown,” “Tops spin,” or “Plants need water.”
A proposition must be distinguished from the sentence used to express it.
A sentence is a series of words; the proposition is the thought behind those
words. A sentence is the concrete, sensuous form of a proposition in just the
way that a word is the concrete, sensuous form of a concept. Just as the word
“table” denotes the same concept as “Tisch” does in German, so “The table
is brown” is the same proposition as that expressed by “Der Tisch ist braun.”
The linguistic symbols differ, the thought is the same. The proposition is
the cognitive content of the sentence, as distinguished from its linguistic
form. (More technically, a proposition represents the form of a judgment,

83 Ayn Rand said that she intended to write a “Volume II” of Introduction to Objectivist
Epistemology, to explain propositions, but she died before undertaking it, and ITOE
has only two or three sentences on conceptual identification. The present chapter
is very largely my own theory.
 H OW W E K N OW • 5 : P ropositions 172

in abstraction from whether one is asserting it, denying it, considering it, etc.,
but I am putting aside this subtlety.)
A given sentence may combine two or more propositions — i.e., make
two or more identifications. “Plants need sunlight, and animals need food”
is a single sentence, but two propositions. More complex sentences may
express several propositions. But we are concerned here only with understand-
ing the basic form of the proposition: the assertion that a predicate, P, applies
in a specified way to a subject, S: “S is P.”84
The first step to making such a conceptual identification is to have some-
thing to identify — i.e., to have isolated and observed some subject-matter, S.
When one grasps some fact about that thing — a feature of it, what it is made
of, what it is doing, etc. — one can name that fact explicitly by applying an
appropriate predicate, P.
Suppose, for instance, that a child is concentrating on a green top as it spins.
The ability to form propositions comes when he can use words — i.e., concepts —
to name what he is aware of, as in: “That is a top.” “The top is spinning.”
“The top is green.” The propositions he forms identify facts in conceptual
terms — i.e., he subsumes what he sees under a concept, in an act of class-
ification. “That is a top” classifies the object he sees under the concept for that
kind of entity (“top”). “The top is spinning” classifies its action. “The top is green”
classifies one of its attributes, its color. (As will be discussed, the grammar
specifies just how the concept functions in any given case.)
The essence of the proposition is the application of a concept to observed
facts. Perceptual observation is the basic form of discovering what things
are, their identities; propositions apply concepts to subjects known directly
or indirectly by perception. By applying concepts, the proposition identifies
the kind of fact one has observed — based on its similarity to other things
known — enabling one to connect the observation to the whole network of
one’s knowledge.
Take “The top is spinning.” A pre-conceptual child may enjoy watching
the motion of the top, but he needs the concept “spin” in order to think the
thought “The top is spinning,” and thus to relate its motion to the similar
motion of other things — a spinning coin, a revolving carousel, a twirling
ice-skater. Or, suppose that during a drought the grass gets browner each
day. The color of the grass is present in the visual fields of both the men

84 Propositions are not the only kind of locutions. Imperatives (“Sit up straight!”) issue
orders; interrogatives (“Where is she?”) ask questions; conditional statements (“If it rains,
we’ll get wet.”) make contingent judgments, etc. However, the base of all these is the
“S is P” proposition.
 C lassificatory P ropositions 173

and the animals in that area, but only the men can form the proposition
“The grass is turning brown.” An animal would not be able even to take the
initial step of isolating the color of the grass as a separate fact, let alone take
the subsequent steps of comparing the present color to that of previous days.
Everything beyond viewing the current scene (which includes grass that
happens to be browner than yesterday’s grass) requires concepts and their
use in propositional thought.

Classificatory Propositions
The simplest and earliest propositions are those that classify an entity under
a first-level concept, e.g., “That is a dog.” When a toddler points at a dog
and exclaims, “Dog!” he is, in effect, making that kind of proposition: he is
subsuming the dog he sees under his previously formed concept, “dog,” just
as if he had uttered the full sentence, “That thing is a dog.”
To subsume something under a concept is to classify it — i.e., to grasp
that it qualifies as a unit of an existing concept. What qualifies an existent
as a unit is its essential similarity to the other existents already integrated by
the concept. It is this dog’s essential similarity to the animals already concep-
tualized as “dogs” that enables the child to classify it as a dog. The child says
“Dog!” because he is aware, perceptually, of this similarity.85
Underwriting the similarity is measurement-proximity — i.e., the fact that
this animal’s shape, size, diet, keenness of smell and hearing, etc. have measure­
ments falling within the “dog” range or category. The canine shape, though
instantly recognizable, is a complex set of ratios, such as of maximum height
to maximum length and thickness, head size to body size, eye separation
to head size, etc. For simplicity in diagramming the role of measurements
in classification, I take as the CCD the relatively unimportant characteristic
of overall size (volume).

Lassie’s size
CCD: Animal size
Dog sizes

85 The proposition, however, is about its subject, not about the similarity involved.
Grasping similarities is the how, not the what. A proposition about similarity would take
that similarity as its subject-term, as in: “The similarity of A to B is . . .”
 H OW W E K N OW • 5 : P ropositions 174

Thus, measurement-relationships underlie the two basic operations of

the conceptual level: concept-formation and conceptual identification.
Concept-formation operates by measurement-omission (to establish a range)
and conceptual identification operates by measurement-inclusion (inclusion
in an established range). Concept-formation creates the file folder; a classi­
ficatory proposition applies the information in the file folder to the subject.86
The child, of course, is aware only of similarity and difference, not of the
underlying measurement-relationships, which is a phenomenon identified
by the epistemologist.
A further step in the same direction comes when the child forms higher-
level concepts; these enable him to make wider classifications. Forming
the wider concept “animal,” allows him to make the wider classificatory
identification: “Dogs are animals.” Here, the subject is a class of thing, dogs,
not a single concrete, Lassie. But the process is the same: “dog” represents
a measurement-range that is, from a wider perspective, only a sub-range
within the animal range of measurements, as distinguished from that
of plants.87 In effect, he places the “dog” file folder inside the “animal” folder.
Or, more literally, he grasps the fact that the concept “dog” is a unit of the
wider concept “animal.”
Just as wider concepts, like “animal,” make possible propositions having
greater generality, so conceptual subdivisions, like “collie,” make possible
propositions having greater specificity. “Lassie is a collie” is more specific
than “Lassie is a dog.”
Many of our concepts represent “cross-classifications” (see Chapter 4),
and these are frequently used as predi­cates. E.g., “Lassie is a pet.” (“Pet” is
a cross-classification.) From this array of quite simple higher-level concepts,
one can already see the permuta­tions start to proliferate: “Collies are animals,”
“Some collies are pets,” “Pets are animals,” etc.

86 All mental content, including concepts, is stored in the nervous system in physical form.
Neuroscience has found that this storage involves changes in neurons and their synaptic
connections. (For a well-written, popularized history of some very recent discoveries
concerning how the brain stores memories, see McDermott, 2011.)
87 In forming “animal” (as a category ranging from, say, fish through mammals) two
broad CCDs are involved: “form of locomotion” and “type of consciousness,” and each
of those involve several axes of measurement. E.g., both a fish and a dog move themselves
around and are conscious, in contrast to a tree or a bush, but there are many measurable
differences in how they do so. Their different forms of locomotion and consciousness are
adapted to their different means of survival in their distinct “ecological niches.”
 D escriptive P ropositions 175

Descriptive Propositions
All propositions apply concepts to a subject, but only some propositions
classify their subject as a whole. Classification occurs only when the
predicate is a noun, in which case the predicate names a class of things.
Since concepts start from perception, and since the content of perceptual
awareness is entities, a child’s earliest propositions have concepts of entities
as their predicates. But soon thereafter, the child develops concepts of char-
acteristics — of attributes, actions, relationships — as discussed in Chapter 4.
(For ease of expression, I continue to refer to actions as “characteristics,” even
though actions are not attributes.)
Thus, there is a difference between propositions that classify and propo­sitions
that describe, a distinction first made by Aristotle. [Categories, II, 1a20–1b8]
Where a classificatory proposition classifies its subject as a whole, a descrip-
tive proposition analyzes out of the whole subject a part, an attribute, the
material of which it is made, etc. In a descriptive proposition, the predicate
can be an adjective (“Tom is tall”), a verb (“Tom ran”), or a prepositional
phrase (“Tom is in the kitchen”).88
One function of classificatory propositions is to connect concepts to
other concepts, which organizes one’s mental file folders into a network.
Having one’s concepts organized logically is of inestimable value; it means
that identifying something as P carries with it or implies that it is also
Q, R, S, T — and so on, for everything to which P is logically connected.
Some of the most valuable of these connections are to the characteristics of S.
“Man” stores all the characteristics of men: that they have a certain range of
size and shape, that they walk, talk, think, learn, have parents, form societies,
buy and sell things, create works of art. Having the concept “man” is cognitively
valuable not because it “labels” men, but because it stores knowledge of their
characteristics. Descriptive propositions are the means of identifying char-
acteristics. And bear in mind that it is the analysis expressed in descriptive
propositions that permits the identification of causal factors, the key to
man’s mastery of his environment and the progress of civilization.
How, then, do descriptive propositions work? What is the process of
propositional judgment when the predicate is not a concept of an entity,
but of an entity’s characteristics? What are the underlying mechanics when
a child thinks “Lassie barked,” or “Lassie is big,” or “Lassie is on the couch”?

88 Nouns may appear in the predicate of descriptive propositions, as shown by “Tom opened
the door” and “Tom is the winner” — neither of which classify Tom.
 H OW W E K N OW • 5 : P ropositions 176

Here, the child is not classifying Lassie under “barked,” as he did when
he classified her under “dog.” We cannot say “Lassie is a barked,” as we say
“Lassie is a dog.” Lassie is not an instance of barking (nor of bigness nor of
on-the-couchness). But in applying such terms to Lassie, the child is still
proceeding conceptually. He is not merely hearing Lassie’s bark, but identi­
fying it conceptually, by applying the concept “bark” to that particular sound
he just heard her make.
The difference between classificatory and descriptive propositions is that
classificatory propositions put the subject as a whole into a class of similar
existents; descriptive propositions analyze the subject, isolating a character-
istic of the subject, then identify that the measurements of that characteristic
fall within the measurement-range denoted by the predicate. Thus, one first
zooms in on one characteristic, isolating it from the subject’s other charac-
teristics, recalls the term for that characteristic, and then zooms back out to
regard that characteristic as being part of the subject’s overall identity.
In a descriptive proposition, there is indeed a classification made, but one
that is implicit not explicit, and the predicate classifies a characteristic, not
the subject as a whole.89 When one says, “The rope is short,” one is classifying
its length, while retaining the fact that it is the rope’s length. An adult with an
advanced vocabulary could express the same thought as: “The rope’s length
is short.” Likewise, “The frog jumped,” is the same as “The action of the frog
was to jump.” But these expanded formulations use abstract terms (“length”
and “action”) that the young child does not yet know.
In a descriptive proposition, the predicate-term classifies a characteristic,
but only as a means of describing the subject. In contrast, in a classificatory
proposition, the classification is not the means but the end, i.e., the point of
the statement.
In both types of proposition, the overall purpose is to bring to bear on
the subject the knowledge stored in the predicate. The effect of classifi­
catory propositions is to realize: “All that which is true of the Ps is true of
this S.” (E.g., all that which is true of dogs is true of this thing.) The effect
of d­ escriptive propositions is to realize: “All that which is entailed in having
characteristic P is true of this S.” (All that which is entailed in being a short
rope is true of this rope.)
The basis of the child’s use of “short” is his recognition that the rope’s length
falls into a measurement-range, the measurement-range established earlier,

89 I am indebted to Gregory Salmieri (personal discussion) for defending this analysis,

which I had considered, but prematurely rejected. See Salmieri, 2013.
 D escriptive P ropositions 177

when he formed the concept “short.” Applying the adjective “short” relates
the rope’s length to the length of other things. “The rope is short” places
its length within the “short” range within the implicit CCD: (relative)
length. (But the CCD is only implicit; at this first stage, he does not have the
concept “length,” only “short” and “long.”) In sum, descriptive propositions,
like classificatory ones, work by measurement-inclusion.
The process of forming any proposition is essentially the same as the
process of forming a concept. In concept-formation, one grasps certain
similarities, based on measurement-proximity in contrast to the measure­
ment-remoteness of a foil, and one establishes a range or category
of measurements. In proposition-formation (i.e., conceptual identification),
one also grasps a similarity, but the range of measurements has already
been established, so that one recognizes that the subject’s characteristics
are included in that established range. Thus, the diagram given above,
to show the mechanics of “Lassie is a dog,” applies to descriptive propositions
as well as to classificatory ones.
Both descriptive and classificatory judgments require awareness of char-
acteristics. Descriptive propositions focus on characteristics individually,
but classificatory propositions are based on a connected set of character-
istics, those common to the members of the class. In classifying Lassie as
a dog, we identify the kind of thing Lassie is. The point of doing so is to
be able to apply to Lassie our knowledge of the set of canine characteris-
tics she shares with other dogs. Contrary to the Realist theory of concepts,
the proposition “Lassie is a dog” does not state her “being” or her total identity.
It does not state, for instance, the fact that Lassie is a movie star, or even that she
is a collie. The proposition “Lassie is a dog” identifies Lassie’s possession of
those canine characteristics that are universal to dogs, not her possession
of those additional characteristics.90
In traditional logic, descriptive propositions are assimilated to classi­
ficatory ones by making up artificial classes in which to place the subject.
“This rope is short” is treated as if it subsumed the rope in the class
of “short things.” But there is no such class and no concept “short things.”
We should not attempt to reduce description to a kind of classification.
Though classificatory propositions are chronologically prior to descriptive

90 This marks a difference between concepts and propositions. The concept “dog” subsumes
all the characteristics of dogs, even those not known at any given stage of knowledge,
and includes individual differences as potentialities (e.g., “a dog can be a movie star”).
This point about concepts is the basis for rejecting the dichotomy between “analytic” and
“synthetic” propositions, to be discussed below.
 H OW W E K N OW • 5 : P ropositions 178

propositions, even the first classificatory propositions (“Dog!”) depend upon

perceiving the characteristics of the subject.
Thus, description is not a disguised classification. “Big” is an adjective,
not a noun; nouns name, adjectives characterize. This takes us to the subject
of grammar.

The Role of Grammar

Grammar is essential to propositions. The different parts of speech are
designed to indicate differences in metaphysical status: entity vs. attribute
vs. action, etc. Nouns and pronouns stand for entities; the other parts of
speech pertain to characteristics of entities. I say “pertain to” because the
other parts of speech — adjectives, verbs, adverbs, etc. — do not name
characteristics of entities; instead, the other parts of speech use concepts
of characteristics to permit the formation of descriptive propositions.
For instance, “soft” is an adjective. In contrast, the noun “softness” names the
attribute in order to make it into a subject (“Softness is good in a pillow”).
We say, “The pillow is soft,” not “The pillow has a soft.” Likewise, “jumped”
uses our concept of the action to describe what an entity does — vs. saying:
“The frog is (or has) a jumped.”
Grammar provides the means for specifying the particular way the
predicate applies to the subject. It is different, for instance, to say “He lied,”
and “He is a liar.” The former describes his action; the latter classifies him as
a certain kind of person. Since English grammar makes much use of word
order, there is quite a difference in meaning between “He loves only her,”
and “Only he loves her.” There is even a difference between “He loves her,”
and “She is loved by him.” The first is about him; the second is about her,
directing our attention to the effects on her of being the object of his love.
(In inflected languages, such as Latin, the “case” of the terms does the work
that in English is done by word-order.)
Grammar also provides the mechanism for combining concepts within the
subject (“The proud, thin man is speaking”) or within the predicate (“Joe is
a proud, thin man”). Characteristics can be combined cumulatively (“proud,
thin”) or, by using adverbs, to make one characteristic modify another
(“proudly thin”). “A proud, thin man” means a man who is proud and
thin, but “a proudly thin man” means a man who takes pride in being thin
— the grammar specifies which is meant.
 N egative P ropositions 179

As this brief foray into a vast field indicates, grammar is an essential part
of the mechanics of the proposition. In Rand’s words, “Grammar is a science
dealing with the formulation of the proper methods of verbal expression
and communication, i.e., the methods of organizing words (concepts) into
sentences.” [ITOE, 37]
In sum, a proposition identifies a subject by grasping that it, or a charac-
teristic of it, is a unit of the predicate. The mechanics of the process reduce
to grasping the measurement-relationships involved. The grammar keeps
straight what is being related to what and precisely how. All this can be put
into a definition: A “proposition” is a grammatically structured combination
of concepts to identify a subject by a process of measurement-inclusion.
The grammatical structure of a proposition indicates whether it classifies
or describes — i.e., whether it identifies the subject as a whole or a characteristic
of it. In either case, a proposition uses one’s awareness of a similarity,
which means that one is aware, at least implicitly, that the measurements
of a characteristic(s) of the subject fall within the category of measurements
covered by the predicate concept. E.g., we classify a given figure as a penta­
gon because it has five sides, or describe a given man as tall, because his
height exceeds six feet.

Negative Propositions
Up to this point, I have been considering only affirmative propositions.
Negative propositions are those having the form “S is not P” — e.g., “Lassie is
not a beagle” or “Lassie is not small.” Negative propositions are differentiations.
Negative classificatory propositions assert that the subject is different
from the existents subsumed by the predicate, and thus is to be excluded
from the predicate-class (Lassie is excluded from the class of beagles).
Negative descriptive propositions assert that, within the predicate’s CCD,
the subject’s measurements fall outside the range covered by the predicate-
concept (within the CCD “size of animal,” Lassie’s size-measurements
fall outside the “small” range for dogs). Negative propositions work by
­measurement-exclusion.
Thus, negative propositions do not refer to some supposed “negative
facts.” Everything that exists is something. To be non-P is to have a positive
identity, but one that is different from P. A thing that is red is non-green
— by virtue of what it is: red.
 H OW W E K N OW • 5 : P ropositions 180

(Outside of philosophy classes, negative propositions are used only when

there is a CCD permitting a differentiation, not to make pseudo-assertions
like “Saturday is not in bed,” or “Man is not a triangle.” To make a measure-
ment-exclusion, there must be a CCD; the characteristic is excluded in the
sense of being placed on the CCD, but outside a specified measurement-
range. There is no CCD uniting man and triangle. In contemporary logic
classes we are told that “non-P” includes every existent of which P cannot
be truly asserted, but in actual cognition, “non-P” means that which is
differentiable from P, not that which is disparate from P. The question
“Is or isn’t man a triangle?” is illogical.)
A special kind of negative statement is: “S does not exist.” There are two
different cases regarding this proposition. In the first case, we are talking
about what did exist but has gone out of existence: “George Washington no
longer exists.” Here, we are not, contradictorily, saying “S exists as some-
thing different from what exists.” We are saying, “George Washington died.”
The same principle applies in the case of inanimate things that go out of
existence — they go out of existence qua the kind of thing they were, but
their constituents remain. E.g., a drinking glass, when smashed, goes out
of existence qua drinking glass, but the glass of which it was made remains,
just in a new arrangement. Or, a newspaper is burned and goes out of exis-
tence qua newspaper, but the chemical constituents remain, although some
are dispersed in the air.
But the interesting case is the one in which the subject never existed:
“God does not exist,” “Unicorns do not exist,” “An integer square root
of 17 does not exist.” In these cases, the actual subject is not an existent,
because there never was and never will be one. Instead, the actual subject is
an idea — the idea of God, or of unicorns, or of an integer square root of 17.
What the proposition is differentiating is not an existent from other
existents but a valid idea from an invalid one. The proposition “S does not exist
(and never has or will)” means one of two things: “The idea of S is false”
(contradicts known facts) or “The idea of S is a fiction, one produced by
imagination not cognition.”
The corresponding reverse statement, “S exists” (in the sense of “S is real”)
means, in this usage, “The idea of S is valid: I perceive or infer something
fitting the description of S.” Normally, the point of saying “S exists” is to
counter the idea that S does not exist — i.e., is mythical or fictitious, as when
someone asserts “The Abominable Snowman exists.” “S exists” cannot mean
that S has the property of existence, since there is no such property. Nor does
 T he A naly tic - Sy nthetic D ichotom y 181

“S exists” classify S among the existents; classifying S already presupposes

that S exists to be classified. Also, “the existents” does not isolate one class
among others. There are no others. Rather, “S exists” means: “I am aware of
something that fits the description of S, so the idea of S is not a fiction.”
The overall point is that negative propositions are about the difference
between one existent (including ideas) and others. All propositions are
fundamentally positive: they state facts of which one is aware, including the
(positive) fact of differences. All propositions are about what is. There is
nothing else to talk about. Existence exists; non-existence does not.91

The Analytic-Synthetic Dichotomy

There are further ways of classifying propositions, but I will restrict myself to
warning against one unwarranted and destructive division: the false dichot-
omy of “analytic propositions” vs. “synthetic propositions.” According to the
advocates of this widely accepted dichotomy, propositions like “Dogs are
animals” are radically different from propositions like “Dogs need water.”
The first statement, philosophers have held, is an “analytic proposition,”
one known by analyzing the concept “dog.” It is true “by definition,”
the proponents of the dichotomy say. But “Dogs need water” is supposed
to be “synthetic,” because we cannot arrive at this identification just by
analyzing the concept “dog”; to learn it, we must observe the empirical facts.
“Analytic” statements are held to be “truths of logic,” which supposedly can
be known without sensory observation. “Synthetic” statements, on the other
hand, are held to be “empirical truths,” as they add to the concept facts that
are only “contingent” and, therefore, cannot be asserted until we see what
“happens” to be the case (and which, it is claimed, might “happen” to be
otherwise, tomorrow).
There are many, many errors and fallacies in this separation of logic from
observation, and in the related notion of “contingent” facts, but rather than
discussing them, I refer the reader to the comprehensive and devastating
criticism of this dichotomy in Peikoff ’s “The Analytic-Synthetic Dichotomy”
(an article that Rand edited and chose to include in her book Introduction to
Objectivist Epistemology).

91 There is not even a concept of sheer “non-existence.” On this, see ITOE, 149–150.
(To apply the lesson of this section, the first sentence of this footnote means:
“ ‘Non-existence’ is not a valid concept” — i.e., it is different from a valid concept.)
 H OW W E K N OW • 5 : P ropositions 182

The basic error behind the analytic-synthetic dichotomy is a wrong view

of what a concept is. A concept is based on the observation of similarities
and differences. Thus, both “Dogs are animals” and “Dogs need water” are
facts learned on the basis of observing things in the world — in the one case,
by observing dogs, animals, and plants, and in the other case by observing
dogs and what happens to them when deprived of water (plus what happens
to other animals when deprived of water).
The analytic-synthetic dichotomy treats a concept as if its content were
limited to only those characteristics used to form it or define it. But the cog-
nitive role of a concept is precisely to serve as an open-ended file folder —
i.e., as a device for storing (and then applying) facts learned by observation,
new facts being stored as they are learned. The dichotomy, in effect, staples
the file shut right after it is started. The result is that new conceptual identifi-
cations are treated as either stipulations (“true by definition”) or rank guesses
about “contingent” matters.
The open-ended nature of a concept is essential to its cognitive (and
action-guiding) function. The complaint “But this is new information about
the referents of the concept” reveals, in Rand’s description:

. . . the unadmitted presupposition that concepts are not

a cognitive device of man’s type of consciousness, but a reposi-
tory of closed, out-of-context omniscience — and that concepts
refer, not to the existents of the external world, but to the frozen,
arrested state of knowledge inside any given consciousness at
any given moment. [ITOE, 67]

A concept refers to certain existents, and those existents have the char-
acteristics that they have. All the units’ characteristics, all the information
in the conceptual “file,” including the characteristics used to form the
concept, are learned from perceptual observation, directly or indirectly.
No characteristic is learned or validated by “analyzing the concept,” if that
means analyzing something other than the integration of observed facts.
Observing reality is the only way to acquire knowledge, and all the so-called
“analytic truths” — from “A is A” to “Bachelors are unmarried” to “Dogs are
animals” — are based on perception. There is no other source of knowledge.
Peikoff summarizes:

. . . concepts mean existents, not arbitrarily selected portions

of existents. There is no basis whatever — neither metaphysical
 C oncepts vs . P ropositions 183

nor epistemological, neither in the nature of reality nor of

a conceptual consciousness — for a division of the characteristics
of a concept’s units into two groups, one of which is excluded
from the concept’s meaning.
Metaphysically, an entity is: all of the things which it is. Each of
its characteristics has the same metaphysical status: each consti-
tutes a part of the entity’s identity.
Epistemologically, all the characteristics of the entities sub-
sumed under a concept are discovered by the same basic method:
by observation of these entities. . . . The fact that certain charac-
teristics are, at a given time, unknown to man, does not indicate
that these characteristics are excluded from the entity — or from
the concept. [ITOE, 98–99]

Concepts vs. Propositions

To some extent, the analytic-synthetic dichotomy is due to ignoring the
differences between concepts and propositions, as in the common belief that
a concept is only a shorthand tag for its definition, which is a proposition.
But there are crucial differences between concepts and propositions, and one
cannot understand either until these differences are recognized.
To form a concept is to create a new file folder; to form a proposition
is to use existing file folders. Concepts are information-handling tools, not
(primarily) identifications. Propositions are the identifications. Concepts
condense and store information discovered. Propositions make use of
concepts to formulate explicit identifications. A concept is a cognitive device,
a means not an end. Propositions are the end in relation to which concepts
are the means. Concepts are the elements of thought; propositions are
the thoughts.
Accordingly, a proposition does not have the kind of unity that a concept
has. Propositions are not integrations — not in the sense that concepts are.
A proposition is an organized combination of concepts (or proper names)
not an integration of them: the parts of the proposition — the individual
words — remain apparent as components.
In contrast, the unity of a concept is seamless. As Rand describes concept-
formation: “The uniting involved is not a mere sum, but an integration, i.e.,
a blending of the units into a single, new mental entity . . .” [ITOE, 10] That
kind of “blending” does not occur in a proposition.
 H OW W E K N OW • 5 : P ropositions 184

A consequent difference between concepts and propositions concerns

their permanence. A concept is a permanent file folder, but the majority
of the pro­positions that one forms are about particular, transient circum-
stances — e.g., “There goes a fire truck,” or “It’s time to have lunch.” Such
propositions are used, but not stored. In contrast, any concept, e.g., “truck,”
represents an open-ended file carried forward to form a permanent part of
one’s cognitive equipment.
A concept is stored by a specific word, and one needs to store that word
and be able to recall it. But the equivalent is not usually true for proposi-
tions. One does not need to memorize specific sentences, not even for
general propositions expressing a truth that is to be carried forward.
For example, there are many facts that I know about hammers — all hammers
— but I have not memorized any sentences regarding that knowledge.92
A concept’s function is to store an ever-growing body of information
about its units; the proposition is the application of that stored information
to a new unit — i.e., an act of conceptual identification.
Concepts function as file folders that store knowledge about their units,
including all their common characteristics. “Dog” applies to all those four-
footed, tail-wagging entities out there in reality. Those entities have all the
characteristics that they have, including their need of water, whether or not
one already knew that fact. The function of the concept is to store that know-
ledge, as it is learned, and to make it permanently accessible. But the function
of a proposition is more limited: to make a specific connection. A proposition
states only what it states. E.g., the proposition “Dogs are animals” states that
dogs have the characteristics that are common to all animals; it does not
state that dogs were domesticated from wolves, even though the concept “dog”
does include that fact, because the concept stands for dogs, and dogs were
domesticated from wolves.
A proposition says what it says, nothing more. A proposition does imply
more than it says, but such implications are not necessarily recognized by the
person asserting the proposition. If someone says, “Dogs that bite people should
be kept muzzled,” his statement has all sorts of implications, but he may
not be aware of them. Grasping implications takes a separate mental act:
inference. As grammarians remind us, “implied” does not mean “inferred.”
A given proposition does not in itself contain all its implications; nor does it

92 To be sure, there are the rare cases in which it is important to remember a specific
linguistic formulation — e.g., for definitions and for statements of basic principles,
such as scientific laws. In such cases, one normally names the formulation — “Ohm’s Law,”
“The Golden Rule,” “the definition of ‘man,’” etc.
 C oncepts vs . P ropositions 185

contain all its presuppositions or all its possible integrations with (or contra-
dictions to) the rest of one’s knowledge.
A concept, in contrast, is open-ended and gets continually updated to
store newly learned facts. A typical proposition does not ever get updated.
(The idea of updating a proposition makes no sense.) E.g., a child first form-
ing the concept “man” does not know that man has a pancreas, but when he
learns that fact, it takes its place in his mental file folder. But suppose that, at
age three, he states, “Man has two hands.” When, years later, he learns about
the pancreas, that does not convert his earlier proposition into: “Man has
two hands and a pancreas.” A proposition states only what it states. The very
nature of a proposition is to make something explicit, as opposed to leaving
it implicit.
The purpose of a concept is to store whatever is learned about its units qua
units. The purpose of a proposition is to bring to full, conscious awareness
the specific, delimited fact(s) that it states.
Thus, concepts and propositions differ in open-endedness. A concept is
open-ended in two ways: extensively and intensively. Extensively, a con-
cept subsumes all the units: past, present, and future. Intensively, a concept
subsumes all the units’ characteristics: known and yet to be discovered.
The situation is different for propositions. Only general propositions, such as
“Dogs have tails,” are, like concepts, open-ended extensively (true of all dogs,
across time); most propo­sitions c­ oncern particular facts and are closed-ended.
The point is that even g­ eneral propositions are not open-ended intensively:
one does not add new i­nformation to them over time, with the growth
of knowledge. (A given propo­sition can become richer in meaning over time,
but that is a different issue: new data is not stored in them.)
The closed-endedness of propositions stating concrete facts (“It’s time for
lunch”) does not make them inferior to generalizations. Just as concepts are
for the sake of conceptual identification, so general propositions are for the
sake of their concrete application. The ultimate purpose of all propositions is
to guide one’s actions in the world, and that means guiding particular actions
dealing with particular concretes. Metaphysically, only concretes exist.93 From
the broadest perspective, the entire conceptual level — the sphere of science,
philosophy, wisdom, the intellect — is for the sake of: “I should now do this,
not that.”

93 Even in the mind, only concretes exist: abstractions are concretes as mental existents, and
it is only in regard to their use in cognition that they are abstract. See ITOE, 153–158.
 H OW W E K N OW • 5 : P ropositions 186

Though “meaning” is a term used in more than one sense, the point can
be put in the following way. The meaning of a concept is the existents it
integrates, with all their characteristics. The meaning of a proposition is the
limited fact it states: the subject’s possession of a certain characteristic(s).

The Cognitive Function of a Proposition

In a tantalizingly brief lead to the cognitive function of propositions,
Rand states:

Since concepts, in the field of cognition, perform a function

similar to that of numbers in the field of mathematics, the
function of a proposition is similar to that of an equation:
it applies conceptual abstractions to a specific problem. [ITOE, 75]

An analysis of propositions in terms of subject and characteristics suggests

how to flesh out Rand’s statement. First, her statement has to be understood
correctly. Her point, as I read it, is not that a proposition equates subject
and predicate — not, absurdly, that “Lassie is a dog” says that “Lassie = dog”
— but, as she says, that there is a similarity in function between a proposition
and an equation. The overall function of propositions and of equations is to
advance our knowledge. “Lassie is a dog” advances our knowledge in a manner
similar to the way “2 + 3 = 5” advances it.
But just how does the arithmetic equation advance our knowledge? What
is the “specific problem” that “2 + 3 = 5” solves? The problem is to identify
the overall quantity of a group, a group known to consist of a pair and a trio.
But there are an unlimited number of other equations that would also
identify the quantity of 2 + 3:

2+3=4+1
2 + 3 = 18 − (4 + 9)
2 + 3 = the cube root of 125

And so on. Every valid equation makes a connection between terms on

the left side of the equal sign and terms on the right. Propositions, likewise,
make a connection between subject and predicate. Establishing a connec-
tion between arithmetic terms or between subject and predicate means
 T he C ognitive F unction of a P roposition 187

that knowledge can be applied: knowledge of what is on the right side of

an equation, or knowledge stored by the predicate of a proposition.
Why, though, is “5” the answer to the question, “What is the quantity
of a group composed of two units and three units?” What is wrong with
giving as the sum “18 − (4 + 9)”? The answer is that “5” is the most unit-
economical way of stating that group’s quantity. Giving the sum as “5” makes
available the greatest amount of other knowledge. “5” gives direct access to
all the facts in the “5” file folder, such as that the quantity is 1 more than 4,
that it is the same as the quantity of fingers on one hand, is odd, prime, has
no integer square root, is the cube root of 125, etc. Identifying the sum of
“2 + 3” as “5” implicitly relates that quantity to the whole integrated set of
mathematical facts that we have stored, carry forward, and continue to learn
more about. The one term “5” on the right-hand side of the equation relates
the two terms on the left to a kind of “central repository” for information on
this quantity.
Propositions perform a similar function. The proposition “Lassie is a dog”
enables us to apply to Lassie all the knowledge of dogs stored in the “dog”
file, which is the “central repository” for information on this kind of animal.
We can see the same point by approaching it from another direction.
Since the source of all information is perceptual observation, the ques-
tion arises: how can applying concepts in a proposition move us forward,
cognitively? Since the Realist theory of concepts is false, since concepts are
not an independent means of access to reality, revealing an “essence” that is
hidden from perception, what does the application of concepts add to what
we already know from perception?
The answer is: relationships. Identifying that Lassie is a dog relates her
to other dogs, to other animals, to plants, and (going in the other direc-
tion) to other breeds of dogs, to kennels, leashes, dog-owners, dogfood, etc.
To identify, in a descriptive proposition, that Lassie is friendly, relates her
to other friendly dogs, unfriendly dogs, friendly and unfriendly people,
stories of friendship, psychological causes of friendship, types of friendship
(Aristotle distinguished three of them), friendships that turn into romances,
issues of loyalty to friends, etc. Some of these relationships are perceptually
available, others are graspable only through intermediate abstractions, but
are reducible back to perception. The point is that the word or concept is not
just “a label,” but a file folder stocked with a growing wealth of knowledge
about the units.
In this regard, the file-folder analogy can be misunderstood: some physical
file folders store concretes. E.g., a file folder for bank statements contains all
 H OW W E K N OW • 5 : P ropositions 188

the individual bank statements one has received. One can pull out that file
folder and open it to find any particular statement. But one’s mental file folders
(i.e., concepts) are not for storing the concept’s units. Rather, the folders are
for storing knowledge about the units.
When one makes a conceptual identification of a concrete, one is not
(normally) doing the equivalent of placing the concrete into a folder; rather,
the contents of the folder are applied to the concrete. The purpose of identi-
fying Lassie as a dog is not to have another example of “dog,” but to under-
stand Lassie. The purpose of identifying that the U.S. government is deeply
in debt is not to have another example of debt, but to understand the
economic condition of the nation. And in every case, the purpose of identify-
ing that S is P is not to add to some catalog of instances of P, but to recognize
what S is — i.e., to recognize that S is characterized by being P.
Introspection supports the point that concepts do not agglomerate con-
cretes: bringing a concept to mind does not pull up all or most of the things
one has identified as instances of it. If I say “lunch” to you, that does not
make it easy for you to recall the tens of thousands of lunches you have eaten
(not even just those that you consciously identified as lunches). And there
would be no purpose in doing that if one could. The purpose of a concept
is to store not rafts of concretes but items of knowledge about any concrete
of a given kind. This permits one to apply knowledge gained from the study
of some concretes to new instances, as they arise.
The point of identifying, for instance, that Joe is a (college) sophomore
is not to enlarge one’s set of known instances of sophomores; the point is
to understand Joe. The proposition “Joe is a sophomore” applies to Joe the
knowledge stored in the “sophomore” file. Here again, any proposition is
about its subject, not its predicate. The predicate is the means of identifying
in conceptual terms what the subject is.
A concept has cognitive content, not just referents. The concept stores
knowledge about its referents, knowledge acquired by observation or by
inference from observation. To use a different metaphor, the concept is like
a water reservoir, and the proposition is like a pipe that enables water to flow
down from that reservoir to fill a particular basin.94 The flow of information
is from the predicate to the subject.

94 Treating the pipe and basin together as the proposition makes for a more accurate,
though less neat, analogy. Overall, this metaphor for propositions is less apt than the
file-folder metaphor for concepts. But it does capture the idea that the concept’s contents
exist for their use, as the water is stored in the reservoir so that it may be piped out.
 T he C ognitive F unction of a P roposition 189

How does information get added to the conceptual file? By means of

a higher-order proposition: a proposition whose subject is the predicate
of the original proposition.
In the higher-order proposition, one identifies something about that kind
of thing — the kind denoted by the predicate of the lower-order proposition.
In the “sophomore” example, the higher-order propositions would include,
first a definition: “A college sophomore is a college student in his second year
of study,” and then such other information as: “Sophomores are typically 19
to 20 years of age,” “Sophomores are qualified to take certain courses not
offered to freshmen,” “Sophomores are said to have an exaggerated sense of
what they know,” “Sophomores who have not yet picked a major are expected
to do so in that year of study,” and so on.
These higher-level identifications, which take “sophomores” as their
subject, make it possible for that concept to store knowledge about all its
units. Then we call upon this knowledge when we reason deductively:
“Joe is a sophomore, therefore he is qualified to take certain courses that he
was not qualified to take as a freshman.”
The proposition “S is P” has its cognitive value in the fact that “P” is
not merely a symbol for all the things similar to S in a given respect,
but that “P” has itself been identified by means of higher-order propositions
of the form “P is Q.” The higher-order identifications permit us to apply that
know­ledge (with whatever qualifications are necessary) to S, by the simple
deductive process:

S is P
P is Q

S is Q

Because concepts store knowledge, once one learns that S is P, the fact
that S is Q becomes implicit in the structure of one’s conceptual filing system:
since the P folder is stored “inside” the Q folder, placing S in P also means
placing S in Q.
Accordingly, the purpose of deduction is to make the implicit explicit — to
bring an item’s implications into full, conscious awareness.
(The process of acquiring entirely new knowledge, as in learning that P
is Q, is essentially inductive. One does not deduce that sophomores are in
their second year of study; that fact is part of the knowledge used to form
 H OW W E K N OW • 5 : P ropositions 190

the concept “sophomore,” which is, in pattern, an inductive process; the pro-
cess is also inductive when one learns, e.g., that colleges expect a student to
commit to a major by the end of their sophomore year.)
Thus, concepts and propositions interact. Concepts make propositions
possible, but then higher-order propositions enable us to add content
to those concepts. Propositions make thought itself possible; indeed, propo-
sitions are the form that thinking takes. And some of that thinking leads
to the formation of new concepts, concepts that could not have been formed
without propositions using earlier knowledge.
The concept of “furniture” provides a simple example. Items of furniture
are not perceptually similar. A bed does not look like a chair. In order to grasp
“furniture,” one has to do some thinking, and that means forming such
propositions as: “The tables, beds, and chairs can be moved around, but
the built-in cabinets can’t be; we put things inside both the stove and the desk
drawers, but the stove is for heating things and the drawers are for storing things.”
The need for propositional thought is, of course, greatly multiplied in the
case of scientists forming new concepts (e.g., “gene,” “electron,” “factorial.”)
Adding to the complexity and richness of cognitive processes is the fact
that new perceptual observation is also involved in this spiraling process.
The hierarchical progression is: perception, conceptualization, propositional
thought; but that of course does not mean that one ceases performing one
stage when one begins performing the acts of a higher stage. Just as concep-
tualization does not replace but supplements perception, so propositional
thought does not replace but supplements, and extends the possibilities of,
conceptualization.
The proposition is a means of bringing to mind and applying to the
subject the knowledge stored in concepts. The primary direction of information
flow is from the predicate to the subject; the predicate illuminates the subject
by bringing stored knowledge to bear upon it.
Propositions are knowledge-appliers.
 LOGIC: THEORY
6
T he ability to form propositions gives rise to a crucially
important phenomenon: the ability to think about one’s own thinking and
thus to judge one’s own judgments. Concepts of consciousness enable one to
identify what one’s mind is doing — whether one is perceiving, thinking,
emoting, imagining, etc. Axiomatic concepts enable one to distinguish
between what is only in one’s mind and what is a fact of the external world,
to distinguish between what merely seems to be and what really is, between
one’s wishes or fears and the facts of an independently existing reality.
The evaluation one makes of a mental process or product affects sub-
sequent cognition; it affects what next comes to mind, and how what is
in one’s mind is filed in memory. Judging affects retrieval and storage.
Different results ensue from judging “That stick is bent” and “Though that
stick seems bent, in reality it is straight.” The effects of passively spinning
out a daydream are very different from those resulting from actively judging,
“Though I wish things were otherwise, they are what they are.”
In consequence, a man’s ability to evaluate his mental processes gives him
a power of self-control that liberates him from the reactive nature of the per-
ceptual level. Perceptual cognition is automatic; perception is world-generated,
hard-wired, and deterministic. Conceptual cognition, in contrast, is volitional:
self-initiated, self-directed, and controllable.
 H OW W E K N OW • 6 : L ogic : T heory 192

Thus, man can use his mind to control the operation of his mind; he
can use the ideas he has formed to steer the process by which he reaches
further ideas; he can use axio­matic concepts to judge the relation of his
mind’s actions to reality and to direct his mind on that basis.
By using concepts of consciousness and axiomatic concepts to make intro-
spective judgments, man becomes cognitively self-determining. He can decide
by conscious, explicit choice what questions to ask, what issues to consider,
what aspects to focus on, how to proceed.95
But conceptual functioning is fallible. Although perception is inerrant,
conceptual processes, by virtue of being volitionally controlled, can be
misperformed, resulting in conclusions that are false — i.e., that contradict
perceived fact. To correctly identify reality on the conceptual level, man
needs a method, with standards, to guide him. We give this method the
name “logic.”
Since the purpose of logic is to align one’s thinking with the facts,
logic requires that one start from and work with the only source of infor-
mation about the facts: perceptual observation. In the conceptual use of
that data — in concept-formation, propositional judgment, and inference
— logic defines the kind of procedures one must follow in order to keep
one’s thinking connected to perceptual reality. (Logic is not only the method
of proof but also the method required to gain knowledge.)
Man survives, progresses, and prospers by acquiring and using conceptual
knowledge. Logic is the method of acquiring knowledge and of ensuring that
it is knowledge.

The Three Laws of Logic

In dealing with perception-derived material, the guidance logic provides
flows from a single imperative: Be consistent. Because contradictions
do not exist in reality, a mental process that involves or implies a contra-
diction has departed from reality and is invalid; a conceptual product that
contradicts any fact is false. Aristotle, “the father of logic,” identified the
Law of Non-Contradiction, stating that it is the basic principle of all knowledge.
He gives this careful formulation of the Law:

95 In an interesting paper, Lee Pierson and Monroe Trout argue that all consciousness,
even the perceptual level of animals, is volitional, not deterministic. [Pierson, L. & Trout,
M., 2005]. The problem with this view is that the animal would be “choosing” blindly;
concepts are required to project the future, evaluate alternatives, and direct one’s actions.
 T he T hree L aws of L ogic 193

The same characteristic cannot both belong and not belong to the
same thing at the same time and in the same respect.96

A thing cannot be hot and not be hot at the same time. If it changes,
over time, from being hot to being cool, that does not violate the Law of
Non-Contradiction. In fact, change is embraced by the law: to be cooling
down is not to be, at the same time, not cooling down. “In the same respect”
is also essential: a thing may be hot in one part and not in another;
it may be hot in relation to ice and not hot in relation to lava. But at a given
time and in a given respect, it is what it is, and is not what it is not.
The corollary of the Law of Non-Contradiction is the Law of Excluded
Middle: a given thing must either have or not have a given characteristic
at a given time and in a given respect. It must either be A or not be A.
(Violations of Excluded Middle reduce to contradictions. A violation
of Excluded Middle would be something that isn’t A and isn’t non-A.
But, as explained in the preceding chapter, what isn’t A is still something,
just something different from A. Thus a violation of Excluded Middle
would be something that is different from A and isn’t different from A
— a contradiction.97)
Later Aristotelians recognized that both these laws stem from the axiom
of identity: “A is A.” A thing is what it is.
There are many ways of formulating these three laws. I suggest the fol-
lowing formulations are those that get to the fundamental: the metaphysical
issue of what it is to be:

The Law of Identity: To be is to have a specific identity. (A is A.)

The Law of Non-Contradiction: To have a given identity
is to have no other. (A isn’t non-A.)
The Law of Excluded Middle: Not to have a given identity
is to have some other. (What isn’t A is non-A).

96 Metaphysics, IV, 3, 1005b19; I added “characteristic” to what, in the Greek, is just

“the same.” Literally: “[For] the same simultaneously to belong and not to belong
is impossible to the same and according to the same.” τὸ γὰρ αὐτὸ ἅμα ὑπάρχειν τε καὶ
μὴ ὑπάρχειν ἀδύνατον τῷ αὐτῷ καὶ κατὰ τὸ αὐτό.
97 As mentioned in Chapter 5, I presuppose the exclusion of such absurdities as,
“The square root of two is red.” Rather than saying, “The square root of two is neither red
nor not red,” the correct statement is: “The square root of two is not the kind of thing
that can have one color or another.” “Not” and “non-” express difference not disparateness.
 H OW W E K N OW • 6 : L ogic : T heory 194

The following formulations are more economical and more memorable,

but they are consequences of what is stated in the more careful formulations:

Identity: Everything is something.

Non-Contradiction: A thing can’t be everything.
Excluded Middle: A thing can’t be nothing.

The Laws of Non-Contradiction and Excluded Middle are reformulations

of the Law of Identity made for the purpose of guiding cognition. To think
that a thing is A and non-A in the same respect, is implicitly to hold that
the thing is everything in that respect. But to be everything in a given respect
is to be nothing in particular in that respect — i.e., to lack identity.
For instance, if a ball simultaneously is and isn’t all red, then it has no
specific identity with respect to color. If the ball is simultaneously here and
not here (or neither here nor not here), then it has no identity in regard
to location. (Epistemologically, a concept represents a differentiation; to file
a thing under the concept and simultaneously under that from which the
concept is differentiated would wipe out the concept. If “table” and “non-
table” have the same referents, then “table” becomes a meaningless sound.)
Knowledge is an awareness of the identity of things; logic enjoins us to use
our conceptual faculty in ways that recognize that things are what they are,
rather than being contradictory or identity-less. Thus, Ayn Rand’s definition:
“Logic is the art of non-contradictory identification.” [AS, 1016]98
Logic applies to every level of conceptual functioning: to concept-
formation, to propositional judgment, and to inference. Inference, though
the most advanced of these three processes, is the simplest one to analyze.
Here, Aristotle’s great discovery was the syllogism, which is the “atom” of
deductive reasoning. The simplest type of syllogism has the following form:99

Lassie is a dog.
Dogs are animals.

Lassie is an animal.

98 It is an “art” in the sense of a method; the corresponding science that identifies

the required methods is epistemology. Epistemology is the theory of the method,
logic the method.
99 I reverse the traditional order of major and minor premise, which, in my opinion,
wrongly subordinates the progression of one’s thought to the issue of which term has the
wider scope. Also, I eschew “All” as both misrepresenting the terms in which one actually
thinks and, in some instances, raising needless questions about special cases.
 L ogic and the I dentit y of C onsciousness 195

The “middle term” — “dog” — establishes the conclusion because of the

law of identity. It is because a dog is a dog that the syllogism establishes
the conclusion — i.e., identifies the fact that the subject term, “Lassie,” is sub-
sumed under the predicate term, “animal.” The connection made depends
upon the terms having the same referents with the same identity. Where the
same word is used but with different meanings, the fallacy of equivocation is
committed — e.g., switching between the metaphorical and literal meanings
of “pig” in: “He is a pig, pigs have four legs, so he has four legs.”
The cognitive mechanics of the syllogism are also based on the law
of identity. To use the file-folder metaphor on this example, the premise
“Lassie is a dog” represents recognizing that the file of information on dogs
applies to Lassie. But that folder is connected to the wider “animal” folder,
as stated in the second premise. If the information about animals applies to
dogs, and the information about dogs applies to Lassie, that implies that the
information about animals applies to Lassie. The deduction makes this impli-
cation explicit. To deny that Lassie is an animal would be to use the “dog” file
inconsistently: both as a subfolder of “animal” and not as a subfolder thereof.
Aristotle identified not only the rules of syllogistic deduction and the
Law of Non-Contradiction as the standard of logic but also the fact that
sensory perception is the self-evident base and court of final appeal
for all conceptual conclusions. His achievements in logic lay dormant
for many centuries, but their recovery almost a millennium later led
to the Renaissance and the Scientific Revolution. [Randall, 1940]

Logic and the Identity of Consciousness

As fundamental and pathbreaking as Aristotle’s work was, it is incomplete.
The principles of logic he formulated pertain to the objects of cognition.
But logic must also take into account the nature of the subject of cognition:
the nature of the thinker’s mind. Identifying this fact and working out its
implications for logic is the achievement of Ayn Rand. [ITOE, ch. 8]
She stresses that man’s cognitive equipment has a specific identity, with
specific terms of operation, which the principles of logic must reflect:
“. . . the rules of cognition must be derived from the nature of existence and
the nature, the identity, of his cognitive faculty.” [ITOE, 82]
Knowledge is a mental product. Any product is made by working up
the proper materials, following a proper method. Automobiles cannot be
made out of bricks; nor can they be made by combining the right materials
 H OW W E K N OW • 6 : L ogic : T heory 196

— steel, glass, plastic — in the wrong way. Likewise, one cannot make know-
ledge out of errors, vague approximations, wishes, or daydreams. Nor can
one make knowledge out of truths improperly combined, as in combin-
ing “Men are human beings,” and “Women are human beings,” to conclude,
“Men are women.” Both the right input and the right method of production
are required. Identifying the right method of production depends upon
identifying the nature of the equipment; a given drill cannot penetrate faster
than a certain rate, a given crane cannot lift more than a certain maximum
weight. There are equivalent limits on the nature of man’s mental equipment
— notably the limit imposed by “the crow epistemology.” (See Chapter 3)
A mental procedure is logical if it adheres to the identity of the materials
supplied by reality (via perceptual observation) and accords with the identity
of man’s consciousness. Any mental procedure or product that contradicts
either reality or the requirements of cognition is illogical.
The requirements of cognition are requirements, not merely matters of
“convenience,” as 20th century philosophers liked to say. Some concep-
tualizations are intelligible, usable, cognitively productive, others are
not. The issue is wider than cognition: achieving any end has its neces-
sary means. To achieve the end, it is necessary, not merely “convenient,”
to conform one’s actions to those means. It is not “inconvenient” but
wrong to travel (unnecessarily) north when one’s goal lies to the south.
It is not “inconvenient” but wrong to attempt to travel from Boston to
London by swimming the Atlantic. Making mental choices that flout the
requirements of cognition results not in “inconvenience,” but in cognitive
paralysis, confusion, and plain error.
Man’s cognitive mechanism is what it is and cannot function in contra­
diction to its nature. The nature of man’s consciousness includes two facts
that are central to logic:

1. Perception is the base of all conceptual cognition (the primacy

of perception).100
2. Only a few distinguishable units can be held in one frame of
awareness (“the crow epistemology”).

100 Although Aristotle grasped this point, it is more explicit and more developed in the
Objectivist epistemology. The term “primacy of perception” is my own and is discussed
in Chapter 11.
 L ogic and the I dentit y of C onsciousness 197

Knowledge is based on the data given in perception. From that base one
builds new knowledge upon old, in an incremental, step-by-step process.
Knowledge is not gained by revelation from on high and it is not gained in
huge gulps. Knowledge is built up, in “crow-friendly” steps, from specific
perceptual observations. We start somewhere: 1) where we are in the
universe, and 2) with the kind of information our perceptual system is
scaled to detecting. E.g., an infant lives in a specific location, with sense
organs sensitive to a specific range of energy differentials and not others.
He can perceive the people and furniture in the room, but not the atoms that
compose them and not the ultrasonic frequencies that a dog can hear.
From his first perceptions, the child’s knowledge grows step by step. His
progress is made possible by further observation, observation of new things
and of new aspects of old things. At a certain point in his development, he
is able to form concepts. From then on, observation and conceptualization
reinforce each other, in a spiraling process, as I discuss later in this chapter.
Now we are prepared to revisit the issue of fallibility. When a child forms
his first concepts, the ones for perceived objects, though some mental effort
of directed attention is required, the process is infallible; there is, for instance,
no such thing as getting the concept “table” wrong. Even if a child forms
a concept of “table” that differs somewhat from the concept that adults name
by that word, the child is nonetheless aware of whatever similarities he
is aware of, and his condensation of the similar concretes cannot be faulted.
The same is true of the application of simple concepts when a child
forms his first propositions, such as “That is a table.” All that is required is
the minimal effort of attention to the thing before him and the recollection
of what the word “table” means. Even though he may err in application,
as when he thinks the distant hills are blue, the process is too simple to require
method­ology, nor could the child at such an early stage grasp rules to follow.
But on the middle and (especially) the higher levels of conceptual cognition,
conscious adherence to rules (logic) is required in order to identify facts
objectively.
In accordance with the identity of reality and the identity of concep-
tual consciousness (especially, the two facts stated above), there are two
over­arching facts about the nature of knowledge that we must adhere to:
knowledge is contextual and hierarchical. The two basic injunctions of
the Objectivist conception of logic are: hold context and obey hierarchy.
[OPAR, ch. 4] Though context and hierarchy are closely interconnected,
let’s look at each in turn.
 H OW W E K N OW • 6 : L ogic : T heory 198

Context
Knowledge is contextual: it consists not of isolated atoms of information
but of a network, a fabric of interwoven connections.
The word “context” suggests by its structure (“con” + “text”) the s­ implest
meaning of “context”: the surrounding text. In reading a given word or phrase,
one needs to hold in mind the sentence of which it is a part. Likewise, the
sentence is part of a paragraph, section, chapter, etc. Meaning is contextual:
without a context, one does not know the right way to interpret an isolated
term — e.g., the word “one,” as used earlier in this sentence.
Peikoff defines “context” as: “the sum of cognitive elements conditioning
an item of knowledge.” [OPAR, 123]
The contextual nature of knowledge reflects a metaphysical fact and an
epistemological one. Metaphysically, reality is an interconnected whole;
not only are there no contradictions in reality but also each existent is part
of the same, causally interconnected universe.101 Epistemologically,
consciousness works by relating things — i.e., by detecting differences and,
in the case of human consciousness, by detecting similarities (based on
differentiation from a foil).
On the perceptual level, integration occurs automatically: one’s brain
automatically combines sensory inputs to present one with an integrated,
global field of awareness embracing all sense modalities. Perceptual
awareness is automatically contextual. But on the conceptual level, in
order to go beyond obvious things that occur to one automatically,
one must actively choose to do the work of integrating new observations and
ideas into the existing structure of one’s knowledge.
Contextuality is inherent in man’s conceptual activities at every level.
The basic act of concept-formation is clearly contextual: concepts are formed
by grasping relationships of similarity against a background of difference
(see Chapters 3 and 4). Propositions are formed to relate subject and predi-
cate (see Chapter 5), which allows the knowledge embodied in the predicate
to be applied to the subject, and also to integrate into the wider context,
since the predicate itself is connected to other conceptualized information
(see the end of Chapter 5). Going still higher, principles, theories, and whole
sciences are each formed in a context and apply in that context.

101 Hume’s claim that every “event” is “loose and disconnected from every other” results from
his wrong, sensationalist view of perception. Contra Hume, in perception we are aware
of a world of entities, not of floating sensory qualities (see Chapter 2).
 C onte x t 199

The network of knowledge has its base in a literal, physical network:

the neural network of the nervous system. We must distinguish process and
product in this regard. The process of grasping a fact is a mental activity;
the storage of that grasp is accomplished physically, by some alteration of the
neural network.
I can say a little, in a general way, about the mechanics of the inter­action
between the mental process and its physical products. When one brings
together two or more items in one frame of awareness and grasps how they
relate — such as that they are similar in a certain way or that one causes the
other — that grasp, if intense enough, gets “saved to disk” — i.e., is encoded
physically by a specific type of change in the brain. The stored and recallable
grasp of a factual relationship is the stored knowledge, the knowledge qua
permanent product. (Note that even when asleep, one still possesses know­
ledge — which exists as a potential: the ability, when awake, to recall things
previously grasped; that potential is due to permanent neural encoding.)
But this ability of the brain to form lasting connections is of little cogni-
tive or practical value if done by random association. When the two items
in focal awareness are co-present only by coincidence, what the brain stores
(when such storage occurs) is a random association, not a logical connec-
tion. The difference is all-important. Only a conscious, rational process can
distinguish mere coincidence from a logical relationship. What logic demands
is not random association but integration, and to achieve integration requires
working to achieve both clarity and precision. Attaining clarity and precision
is a precondition of checking for consistency. One cannot check the unclear
or the vague for its coherence with the rest of one’s knowledge.
The random associations that occur in the normal course of life pose no
problem — provided one recognizes them as being random associations.
For instance, it is said that most people remember where they were when
they learned of some shocking event, such as the shooting of John F. ­Kennedy
or the attack on the World Trade Center. But everyone is aware that this
is random association: it is obvious that there is no logical connection
between, say, one’s having been on a picnic in Pasadena and Kennedy’s
having been shot in Dallas. In other areas, however, people can be unaware that
they have mistaken correlation for causation, details for essentials, familiarity
for universality.
The ability to rise above random associations is what separates man
from the animals. Animals function by random association of perceptual
concretes. An animal that is burned by a flame will subsequently fear and
avoid anything that is perceptually similar. A man can analyze the factors
 H OW W E K N OW • 6 : L ogic : T heory 200

that make the flame more or less dangerous, such as the flame’s intensity
and proximity; he can learn what is flammable and what is not, and use the
latter to shield himself from the flame’s heat; he can learn what extinguishes
a flame and what fuels it. Thus, he learns how to tame fire and make it serve
his needs.
The wider point is that logic directs us to work to reach a clear, precise
awareness of the specific nature of the relationships existing among things.
Logic directs us not to rest with “fire has to do with stuff that burns,” but
“fire requires heat, oxygen, and a flammable substance that serves as its fuel.”
Prometheus’ gift of fire would have had no value to a being restricted
to the gross, unanalyzed associations occurring on the perceptual level.
A conceptual being is able to break down the perceptually given into its char-
acteristics and components and to identify each conceptually. Man’s m ­ astery
of nature stems from this ability to rise to the conceptual level, to use
concepts to analyze concretes, to logically identify causal factors, and thus
turn the nature-given causal powers of entities to the service of his needs.
All items of knowledge are interrelated, and every item of knowledge
is acquired, maintained, and used according to its relationship to the
rest of one’s knowledge.102 Knowledge is a network of discriminated and
inter­related content, not scattered bits acquired in separate “revelations.”
Expanding one’s knowledge is like adding the next pieces to a jigsaw puzzle:
each new item must attach to those adjacent to it, if it is to cohere with and
add to the whole.103 Thus, the point that knowledge is contextual goes to the
essence of what knowledge is: a growing network of interconnected material.
The immediate context of an item is the knowledge directly connecting
to it — as the knowledge that Phoenix is in Arizona connects directly to the
knowledge that Arizona is a U.S. state. The wider context consists of the
things directly related to that, and then the things directly related to them,
and so on — such that the full context is the totality of one’s knowledge at
a given time.
Since every fact bears some relationship to every other fact, however remote,
one must work to integrate one’s knowledge into a non-contradictory whole.
Rand’s important summary statement is:

102 Even the first item of knowledge one acquires is contextual. Knowledge begins with
perception, and perception is not of an isolated datum but of the world, with its panoply
of entities. Perception is inherently wholistic and contextual.
103 The jigsaw metaphor was suggested by Jean Moroney.
 C onte x t 201

No concept man forms is valid unless he integrates it without

contradiction into the total sum of his knowledge. [AS, 1016.]104

Doing the work of integration is a multifold process, requiring more

than good intentions. One has to understand and follow the rules of the
“non-­contradictory identification” that constitutes logic. Take: “Tax cuts
would stimulate the economy.” Is there any contradiction in that idea?
That question cannot be answered by simply posing it. There are myriad
questions and subquestions one has to consider in order to deal logically
with that issue. The whole science of economics and of its philosophic base,
as well as a knowledge of history and of the nature of man, is involved in
reaching a logical conclusion.
In integrating something into the total sum of one’s knowledge,
there are two stages. First, one checks the coherence and consistency of
the new item with its “nearest relatives” forming its immediate context.
One needs to ask oneself, “How does this relate to the things in the next
wider category (its genus)?” “What are its immediate causes (both of the
thing and of one’s knowledge of it)?” “What are its direct applications
and/or implications?” and, “What are other examples or instances of this?”
In each case, one is both looking for positive connections and checking
for possible contradictions. (Depending on the results one reaches, this may
call for further integration.)
The second stage of integration is wider: one sets a “standing order”
in one’s mind — a settled intention — to carry forward the process of inte-
gration into the future. That is, if one has reached a conclusion (or formed
a hypothesis), one commits oneself to being alert to additional information
that one can connect to the item.
The injunction to integrate, i.e., to “hold context,” pertains to the applica-
tion of knowledge as well as to its acquisition. In applying one’s knowledge,
one must give due attention to the specifics of the case at hand. Concepts
and generali­zations cannot be treated as contextless formulas or rote rules.
The mechanical, context-blind application of knowledge constitutes a fallacy,
one identified by Aristotle and now known as “The ­Fallacy of Accident.”105

104 In this passage, Rand uses “concept” to mean any conceptual product, including proposi-
tions, theories, etc.
105 A Wikipedia article gives a good example of this fallacy: Cutting people with a knife
is a crime. Surgeons cut people with a knife. Therefore, surgeons are criminals.
 H OW W E K N OW • 6 : L ogic : T heory 202

Rand’s term “context-dropping” names the wider error: ignoring available

facts that would alter or contradict one’s conclusion.
A simple form of context-dropping is that involved in irrational behavior.
An action is irrational if it stems from evading the long-range consequences.
It is context-dropping to function on the basis of a kind of tunnel-vision that
restricts one’s range of view to the here and now. Overspending facilitated
by a credit card may be satisfying for the range of the moment, but it carries
a long-term penalty. Defrauding someone may bring you cash today, but in
the full context it exacts an unpayable price: setting yourself up as an enemy
of every honest man and instituting a state of war with reality.106 Any action
that sacrifices the long-range to the short-range represents context-dropping.
Peikoff recounts a striking example of what context-dropping means.
Here is his analysis of Neville Chamberlain’s appeasement of Hitler’s demands
regarding Czechoslovakia, at the Munich conference in 1938:

Mr. Chamberlain treated Hitler’s demand as an isolated fact to

be dealt with by an isolated response; to do this, he had to drop
an immense amount of knowledge. He did not relate Hitler’s
demand to the knowledge already gained about the nature of
Nazism; he did not ask for causes. He did not relate the demand
to his knowledge of similar demands voiced by aggressor nations
and even local bullies throughout history; he did not ask for
principles. He did not relate his own policy to mankind’s know-
ledge of the results of appeasement; despite ample indications,
he did not ask whether his capitulation, besides satisfying Hitler,
would also embolden him, increase his resources, hearten his
allies, undermine his opponents, and thus achieve the opposite
of its stated purpose. Chamberlain was not concerned with any
aspect of a complex situation beyond the single point he chose
to consider in isolation: that he would be removing Hitler’s imme-
diate frustration.
Deeper issues are involved in this example. Chamberlain
was proposing a course of action while ignoring the field that
defines the principles of proper action, ethics. He did not ask
whether his course comported with the virtues of honor, courage,
integrity — and, if not, what consequences this portended.

106 On the latter, see OPAR, 267–276.

 C onte x t 203

He dropped the fact that foreign-policy decisions, like all human

actions, fall within a wider context defined by moral philosophy
(and by several other subjects as well). The prime minister wanted
“peace at any price.” The price included the evasion of political
philosophy, history, psychology, ethics, and more. The result
was war. [OPAR, 124–125]

The more one works to integrate, the better one’s mental filing system
and the easier further integration becomes. A proper filing system deals
in essentials.107 As a consequence, the task of integration becomes progressively
more efficient. And of course one does not set about integrating at random
(“Does my conclusion about tax cuts integrate with the Pythagorean
theorem?”). Moreover, if one has formed concepts properly, one’s mental file
folders will have a certain hierarchically nested structure, and that also aids
the integrative process. One asks: “What is taxation?”, “What are the conse-
quences of taxation on the individual? On government financing?”, “What is
the source of economic progress?”, etc.
The effects of tax cuts are part of the science of economics, which is part
of the social sciences. Knowing the “tree of knowledge” vastly accelerates
the needed integration, allowing for things to be integrated wholesale.
The Pythagorean theorem, for instance, is part of geometry, which is part of
mathematics; the required integration is not that of tax cuts with the theorem
but rather the integration of the entire field of mathematics with the entire
field of the social sciences. The overall integration of two such sciences
is very abstract, but fairly straightforward, and one would have no reason to
expect to find an inconsistency between them. (E.g., economics presupposes
and uses mathematics, since economics deals with quantitative relations,
but mathematics does not presuppose or use principles of economics; mathe-
matics, though devised by man, studies all quantitative relationships, whereas
economics studies an aspect of man — his interactions in production and
trade; mathematics is a science of method — of calculation and measure­ment,
whereas economics is a theoretical science, etc.108)

107 The Objectivist theory of “essential” will be discussed in the next chapter.
108 In fields other than philosophy and mathematics, integration usually requires acquiring
new data by fresh perceptual observations. (For philosophy, the observational base is
the common knowledge we have gained as civilized adults; for mathematics, the base
is provided by simple observations about quantity and shapes. In neither field does one
need to acquire new information about concretes to get new data — although doing
so may aid one’s thinking.)
 H OW W E K N OW • 6 : L ogic : T heory 204

Integrating an idea into the full context is a demanding activity — far more
demanding than the simple act of combining two premises in a syllogism.
Proper integra­tion consists of a conscientious process of checking the idea
against (the essentials of) all of one’s knowledge, actively looking for possible
contradictions or counter-data. It also requires the commitment to doing
fresh thinking should any interesting or problematic facts arise in the future
(such thinking is normally an enjoyable, rewarding process.)
The fact that knowledge is contextual has crucial implications for the
field of epistemology: the standards of judging an idea’s validity must take
into account the contextual nature of knowledge. No cognitive standard
can require one to have more knowledge than is possible at a given stage
of cognitive development. One cannot require that, for instance, in order
to attain certainty, one must know everything that could bear upon one’s
conclusion — i.e., be omniscient. One cannot downgrade or invalidate
actual cognition because it is contextual. Standards must adjust to reality,
not begin with a fantasy (omniscience) and downgrade what is real by
reference to what is dreamed up. [Binswanger, 1981] Standards must judge
by reference to what is possible within the available context of knowledge.
If an idea is supported by observation and integrates with all the knowl­
edge available, that idea is valid. Its validity is not retroactively undone
if, as occasionally happens, the idea has to be subsequently qualified,
or even rejected, on the basis of new data. Given the context of knowl­
edge that one possessed, the idea was either reached logically or it wasn’t
— and that fact about the past situation never changes. The issue is not:
“What would one conclude if one were omniscient?” but: “what is the proper
conclusion to draw given all the facts available now?” Epistemic standards
are prospective, not retrospective.109
Contextuality thus has a dual application: 1) one must be consistent with
all the knowledge currently available, and 2) a standard cannot require
consistency with the as-yet-unknown. “Man cannot know more than
he has discovered — and he may not know less than the evidence indi-
cates, if his concepts and definitions are to be objectively valid.” [ITOE, 46]
The purpose of standards is to guide one’s present choices.

109 Even retrospectively, the issue is: “What should I have done, given what I knew (or should
have known) at the time?”
 H ierarch y 205

Hierarchy
In Chapter 4, I discussed the hierarchical nature of concepts, but hierarchy
is a much wider phenomenon: every aspect of conceptual knowledge
is hierarchical. Hierarchy is essential to logic, and the anti-hierarchical
approach is the one factor most responsible for the confused and chaotic
state of today’s intellectual world.
“Hierarchy” as a general term pertains to a number of ways in which things
exist in an order of dependency, but the specific meaning of hierarchy I will
focus on is: the hierarchy of learning — i.e., the necessary order of acquiring
knowledge. This is the kind of dependency that occurs when knowing
A is a prerequisite of learning B, as knowing arithmetic is a prerequisite of
learning algebra. In this hierarchical relationship, A grounds B.
The dependency here is causal dependency — but the causality is that
operative in cognition, not physical causality. “Hierarchy” refers to how
certain knowledge makes possible other knowledge, not to what causes what
in the external world. For instance, consider the order of grandfather and
father. Existentially, the grandfather came into being before and caused the
existence of the father; but epistemically, the order is exactly the reverse:
first one forms the concept “father,” and only then can one grasp the idea
of a parent’s father — i.e., “grandfather.”110
The hierarchy of knowledge is epistemological; it is not some structure
that exists in the external world (see Chapter 4). But the hierarchy reflects
two metaphysical facts. First, the range of things that the senses can respond
to is metaphysically given. We cannot, for example, respond to ultraviolet
light, so our knowledge of ultraviolet depends hierarchically upon inference
from what we do perceive. Second, some differences are too great to be
accommodated by our limited “crow” capacity. For instance, the differ­ences
between trees and dogs are more numerous and extensive than the differ­
ences between dogs and cats. Accordingly, to integrate “dog” and “tree” into
“organism,” we need to advance in “crow-friendly” steps. Only after we have
formed and become familiar with the concepts “animal” and “plant,” are we
able to contrast both of them to non-living things, such as rocks and rivers.

110 The hierarchy of knowledge pertains to the content of knowledge, not to its external
conditions. E.g., to learn algebra, one has to stay alive, but knowing how to obtain food,
shelter, etc. is not part of the hierarchical base of algebra. The hierarchy of know­ledge
concerns cognitive prerequisites: the earlier identifications that have to be used in grasping
later ones. To grasp “the father of a father,” one must use the knowledge of what a father
is; to grasp that 2 + 2 = 4 one must use the knowledge of what 2 and 4 are.
 H OW W E K N OW • 6 : L ogic : T heory 206

The intermediate concepts are condensations that make “room” for grasping
how dogs and trees are similar. (By analogy, ladders are necessary to reach
things that are too high; and the spacing of rungs is set by the size of man’s
step, just as man’s cognitive tools must be adapted to the size of the “crow.”)
A vivid illustration of how the “crow” creates the need for a sequence of
condensations, and thus hierarchy, is provided by numbers. To grasp the con-
cept “million,” a child first has to have formed the concepts of the numbers
one through ten, then “hundred” as ten tens, “thousand” as ten hundreds,
before he can finally reach “million” as one thousand thousands.
Hierarchy is an aspect — the structural aspect — of context. Hierarchy
is the context insofar as it exhibits epistemic dependencies. For instance,
the concept of “dog” and the concept of “cat” are clearly related parts
of one context, the context of one’s knowledge of domesticated animals.
But there is no hierarchy between “dog” and “cat”: one does not need to form
“dog” on the basis of already having formed “cat” or vice-versa. In contrast,
the concept of “kennel” is not only contextually related to the concept
of “dog,” it is hierarchically dependent on “dog.” One could not form the
concept “kennel” if one did not already have the concept “dog.”
Peikoff summarizes the relation of hierarchy and context:

A hierarchy is a type of context. The contextual view of know-

ledge states that cognition is relational. The hierarchical view
identifies a particular kind of cognitive relationship: it states not
only that every (non-axiomatic) item has a context, but also that
such context itself has an inner structure of logical dependence,
rising gradually from a base of first-level items. The principle
of context takes an overview; it looks at the sum of knowledge
already acquired and says: it is a sum. The principle of hier­
archy looks at the process by which a given item was learned and
says: the simpler steps made the more complex ones possible.
[OPAR, 132]111

The need to obey hierarchy has long been recognized in regard to infer-
ence. The well-known fallacy of “begging the question” (petitio principii)
consists of a hierarchy violation: illicitly using what was to be proved, in
the attempt to prove it — i.e., circular reasoning. The textbook example of
question-begging is: “There must be a God because the Bible says there is,

111 In general, I depend heavily on Peikoff ’s discussion of hierarchy in OPAR pp. 129–132,
and in many of his articles and lectures.
 H ierarch y 207

and I know the Bible is true because it is the word of God.” That is a hierarchy
violation: A cannot be established by B, if B has to be established by A.
More widely, the injunction to obey hierarchy means never to treat an
idea as if it could be known or used in disregard of the knowledge that is
hierarchically prior to it — i.e., on which its grasp and intelligibility depend.
In contemporary philosophy, some minimal attention is given to the issue
of an item’s immediate context, but there is total, blithe unconcern with
hierarchy. To understand the issue of hierarchy, one must be very clear on the
difference between uttering a word and grasping its meaning. But, under the
influence of nominalism, philosophers have denied there is any difference
between “socially approved linguistic behavior” and grasping the meaning
of a concept. Thus, the philosophic establishment first opposed, then became
blind to the issue of which concepts depend on which prior concepts.
For instance, a leading philosopher of the 20th century, Willard van
Orman Quine, proclaimed:

Any statement can be held true come what may, if we make drastic
enough adjustments elsewhere in the system. Even a statement
very close to the periphery can be held true in the face of recalci-
trant experience by pleading hallucination or by amending certain
statements of the kind called logical laws. . . . Revision even of
the logical law of the excluded middle has been proposed . . . .
[Quine 1953b, 43]

Quine’s is not a minority position.112 But his flippant dismissal of hierarchy

flops when examined. E.g., there is no way that “pleading hallucination” could
work to hold as true “Sand is food.” And what “adjustments elsewhere in the
system” could make it possible to reject the laws of logic? Those laws, includ-
ing Excluded Middle, are the hierarchical base of all conceptual functioning.
The very idea of “revision” implies that the revision is what it is, that it is not
what it is not, and that it is not a nothing. What “adjustments” could preserve
algebra while denying arithmetic? Or could hold onto the concept “orphan”
while denying the concept “parent”? Quine is here denying the patent fact
that some knowledge uses and builds upon prior knowledge — i.e., he is

112 For one of the most stunningly anti-hierarchical passages in the history of thought, see
Nelson Goodman’s claim that one could form the concepts “green” and “blue” on the basis
of (impossibly) first forming his (absurd) concepts “grue” and “bleen.” [Goodman, 79–80]
 H OW W E K N OW • 6 : L ogic : T heory 208

denying hierarchy. His motive is to “espouse a more thorough pragmatism.”

That means: to negate the axioms of existence, identity, and consciousness.
Quine’s opposition to hierarchy represents an epistemological egalitar­
ianism: all propositions are held to be equal, so that we can deny any
one of them and avoid the consequences by revising, at whim, any of the
others. In terms of the construction industry, this would mean: we can remove
the ground floor and support the second floor by hanging it from some of
the higher floors. But what supports those higher floors? Can we deny the
evidence supporting a given conclusion and still “hold onto” that conclusion?
Only by making things up to support the preferred conclusion. And in some
cases, such as denying axioms, even that is not possible. Try denying that
existence exists and holding onto, say, “Wines can be red or white” when one
has just claimed that wine, along with everything else, does not exist.
Another 20th-century philosopher, Wilfrid Sellars, launched a highly
influential attack on the very foundation of hierarchy: the perceptual base of
all knowledge, which he describes as “the myth of the given.” This is a colos-
sal stolen concept: the concept of “myth” denotes a product of imagination in
distinction from that which is based on obser­vation — i.e., on the perceptually
given. To say it is a myth that perception is the given is to say that basing ideas
on observation flouts the need to base ideas on observation. If the self-evident
status of perceived fact were a myth, what, then, would distinguish myth from
non-myth? What would it mean to say, “Centaurs are mythical creatures”?
Nothing. And it means precisely nothing to say “the given is a myth.”

The Spiral Process of Knowledge

Although knowledge cannot be acquired out of hierarchical order, the
growth of knowledge is not just a unidirectional, linear progression. There is
also a process of spiraling back — the use of hierarchically later knowledge
to add depth and content to hierarchically prior knowledge. The hierar­
chical order may be A–B–C, but relating C back to A will nonetheless add to
one’s understanding of A. Here, I distinguish “understanding” from “grasp.”
The hierarchical order states the order in which items must be grasped.
That means: to grasp C, one must have already grasped B, which in turn required
a prior grasp of A. But once a fact has been grasped, one’s understanding
of it can grow in all directions, expanding with each integration one makes
to other items of knowledge, even to the hierarchically later knowledge
that it grounds.
 T he Spiral P rocess of K nowledge 209

For instance, grasping the concept of “wine” is hierarchically prior to

making distinctions among types of wine — red vs. white, dry vs. sweet,
Burgundy vs. Merlot — but coming to know each of these subsequent facts
adds to one’s understanding of wine, enhancing and enriching it.
Grasp is an all or nothing affair — either one has taken secure cogni-
tive hold of a fact or one has not — but understanding admits of degrees.
One understands a topic better and better as one learns more and more about
it. (But omniscience is not the standard: gaining an expanded understanding
of something does not mean that previously one had understood nothing
at all about it.) Understanding an item comes from connecting it to other
knowledge, and the more connections made, the fuller the understanding.
As Rand observes:

To understand means to focus on the content of a given subject

. . . to isolate its essentials, to establish its relationship to the
previously known, and to integrate it with the appropriate cate­
gories of other subjects. Integration is the essential part of
understanding. [RP, 68]

For example, take the simple recognition that the sun moves across the
sky from east to west. Grasping that fact was hierarchically required in
order to reach the knowledge (discovered by Copernicus, et al.) that the
Earth rotates on its axis while it orbits the sun. But having reached that
later knowledge adds to one’s understanding of the sun’s apparent motion.
For instance, in conjunction with the later knowledge of the tilt of the
Earth’s axis, it explains why the sun’s path across the sky changes with
the seasons, why the seasons are reversed in the Southern Hemisphere,
and it makes intelligible the observed fact that sunspots move across the sun
(the sun is rotating on its axis, too).
In the case of concepts, there is another important order besides hier­
archical order: the order of generality. The hierarchy of concepts is the
order in which concepts must be formed. Starting from the perceptual level,
the conceptual hierarchy proceeds in two directions, “up” (wider inte-
grations) and “down” (subdivision) (see Chapter 4). In order to learn
“collie,” one must have grasped the concept “dog”; and in order to go “up” from
“dog” to “animal,” one must have conceptualized some other animals — e.g.,
“cat,” “horse,” and “pig.” But once all these concepts have been formed, they
are organized in one’s mental filing system according to their generality,
which gives a different order: collie-dog-animal.
 H OW W E K N OW • 6 : L ogic : T heory 210

Whatever the order in which concepts are formed, a given concept, once
formed, subsumes its subdivisions, and when a file folder is stored inside
a larger one, it “inherits” the knowledge that the larger folder contains.
Conversely, the larger folder gains in content when subfolders are added
inside it. Again, although one cannot grasp what it is for something to be
a Burgundy before knowing what it is to be wine, one gains a fuller, richer
understanding of wine when one learns that wines can have the color, flavor,
and body of a Burgundy in distinction to those of a Merlot.
Spiraling back through the hierarchy also enriches the narrower concepts
that were used to form wider ones. Consider how forming “animal” adds
to one’s understanding of what it is to be a dog. Even if “dog” was one of
the concepts that a given child used in forming “animal,” his realization
that a dog is an animal gives him a wider knowledge of dogs. For instance,
it enables him to see how dogs differ from plants and are similar to other
animals — including their similarity to the very animals from which dogs
were differentiated when he formed the concept “dog.”
Going even wider, the realization that dogs are living organisms requires
that one see dogs as being similar to trees and grass but different from stones
and automobiles. In other words, here too, the later concept illuminates
the earlier concept on which it depends.113
For a more interesting example of the spiral, take the child’s progression
in understanding money. “Money” is an integration wider than the concept
of some specific form of money — dollars, francs, yen, etc. But consider
an American child who has learned “money” from his experience with
only American bills and coins. He understands that these bills and coins are
things universally accepted as payment (i.e., serve as a medium of exchange)
and that they can be saved (i.e., are a store of value). Later, he learns that
in other countries, money takes different forms — euros, pounds, pesetas.
Still later, he learns that in other times gold, silver, and even salt or sea shells
served as money. He did not need this advanced information in order to grasp
“money,” but learning it enhances his understanding of money.
The general point captured by “the spiral process of knowledge” is that
integrating an item into an expanded context of knowledge enhances
the understanding of each item in that context — whatever its hierarchical

113 I am not sure that it is quite accurate to say that a wider concept depends on any particu-
lar one of the narrower concepts used in forming the wider concept, since one could have
used some other narrower concept to make the widening. But this subtlety is unimport-
ant to the overall discussion of spiralling.
 T he Spiral P rocess of K nowledge 211

provenance. But gaining that understanding presupposes that one has grasped
the item in the first place, which requires the hierarchical progression.
Logic is not a desiccated concern with the rules of deduction or a game
of manipulating symbols. Logic is the means of keeping conceptual cognition
connected to reality. Since conceptual cognition is a process of ­progressive,
stepwise integration, the two essentials of logic are context and hier­
archy. Logic demands that one ground each conclusion in perceived facts
or in knowledge derived from perceived facts, and that one integrate
the conclusion into the full context. What this means in actual practice
is the subject of the next chapter.
H OW W E K N OW 212
 LOGIC: PRACTICE
7
U p to this point, the material covered has been descriptive
rather than normative. I have described how we know — when we do
indeed attain knowledge, i.e., when the process of cognition has succeeded.
But the purpose of describing what successful cognition consists of is to
define the oughts — i.e., the norms and standards to which our thinking
must adhere if we are to acquire actual knowledge of reality. Theory is for
practice, and I turn now to the practical application of all the preceding
theory: the rules of logic. This is not a textbook, however, and complete-
ness is not my aim in the following presentation. I have sought to restrict
the topics I cover to those that are essential to logical thinking, giving
more attention to topics that are likely to be new to the reader — either
because classical, Aristotelian logic has come to be neglected in college
courses on logic, or because the material was originated by Ayn Rand
(or, in one or two minor instances, by me). It is widely assumed today that
logic deals only with inference, but logic exists to provide guidance for all the
conceptual functions that are subject to volitional control. Thus, logic covers:
concept-formation, propositional judgment, and inference. The present chapter
takes up each of these, in turn.
 H OW W E K N OW • 7 : L ogic : P ractice 214

Logic and Concepts

In the preceding chapter, I stressed the need to recognize the hier­archy of
concepts — the fact that some concepts (e.g., “sophomore”) can be grasped
only on the basis of first grasping prior concepts (“student”). This gives rise
to a question: since the hierarchy represents a necessary order of forming
concepts, why does logic need to tell us to obey it? Why isn’t the alterna-
tive simply: either one goes through the proper, hierarchical progression
or one is flatly unable to proceed?
The first answer is that obeying the hierarchy is necessary — to form
actual concepts. On the higher levels of abstractions, many people form only
approximate concepts, words used by imitating the way others use them.
Such half-baked, semi-formed concepts cannot be applied accurately to their
units. Accordingly, Rand calls them “floating abstractions.”
Floating abstractions are cognitive cripplers. They represent, in Rand’s
words, “condensing fog into fog into thicker fog — until the hierarchical
structure of concepts breaks down . . . losing all ties to reality.” [ITOE, 76]
Consider the vague, woozy manner in which many people hold and use such
higher-level concepts as “love,” “freedom,” and “justice.” Such people are clear
on the application of these concepts to a few simple concretes; they know,
for example, that a slave held in chains is not free. But, not having gone through
the hier­archical steps necessary to attain a full, clear grasp of these concepts,
people can use them in bizarre and contradictory ways, resulting in romantic,
political, and moral chaos, respectively.
There is a second reason why the hierarchical progression, though neces-
sary, can yet be violated. Even if one did go through all the required steps when
first forming a given concept, one may not remember that hierarchy years later.
Thus one may not notice the fact that a given proposition uses that concept
in a way that violates the conceptual hierarchy, resulting in committing
the fallacy of the stolen concept.
The main remedy for both floating abstractions and stolen concepts is the
process of definition. “A definition is a statement that identifies the nature
of the units subsumed under a concept.” [ITOE, 40]

The Function of Definitions

A definition serves the function of isolating a concept’s units, thus providing
the concept with a specific identity.
 L ogic and C oncepts 215

Every word, except proper names, denotes a concept. An average ­English

speaker knows something like 30,000 words. How is he to keep each concept
attached to its specific referents in reality, particularly when the same concrete
can be classified under many different concepts? I am a ­referent of “man,”
“animal,” “philosopher,” “American,” “husband,” “stockholder,” “taxpayer,”
“civilian,” and dozens of other concepts, so merely bringing an image of me
to mind would not enable you to distinguish these concepts from each other.
How can one organize 30,000 concepts? What keeps order in one’s mind and
distinguishes a given concept from every other? The definition.
The definition is analogous to the label on a (physical) file folder. The label
indicates, as concisely as possible, what the folder contains — i.e., the nature
of the information it stores — just as a concept’s definition allows one to
quickly recognize the meaning of the concept — i.e., the nature of the units
it integrates. A proper label also permits organizing the file folders, so that
one can place related files together inside larger file folders, put those into
file drawers, and organize file drawers into file cabinets and banks of file
cabinets. With 30,000 such files, this is not a mere “convenience,” but a matter
of cognitive necessity.114
(Certain concepts can be defined only ostensively — by pointing to or
otherwise indicating instances and contrasting them with appropriate
foils. Rand places in this category axiomatic concepts, such as “existence,”
“identity,” and “consciousness,” and concepts of the primary aspects of
experience — “sensations” in her terminology. [ITOE, 40–41] Neither what
it is to exist nor what it is to feel pain or to taste the sweetness of sugar can
be analyzed into anything prior. To isolate or communicate these primaries,
one can only give examples and foils.)

The Rules of Proper Definition

A definition’s vital function can be performed only if it meets certain require-
ments. We owe the rules of proper definition to Aristotle. The Aristotelian
rules have importance far beyond their role in definition; they illustrate
the pattern of conceptual cognition as such: differentiation and integration.
Traditionally, however, these rules are stated in negative terms; I have
reformulated them as positives. I also have reorganized them slightly and

114 Reflecting her stress on the need for definitions, Rand defined over 100 philosophic terms
in her nonfiction writings. See Glossary of Objectivist Definitions, Kunze & Moroney, 1999.
 H OW W E K N OW • 7 : L ogic : P ractice 216

put them into an Objectivist context, giving them new names, when doing
so clarifies their function.
The justification of these rules is that they serve the purpose of a definition:
they spell out the method of identifying a concept’s meaning — a method
adapted to the identity of man’s consciousness.

1. The rule of genus and differentia

One of Aristotle’s great achievements is the idea of defining concepts
by means of genus and differentia. The genus is the wider class containing
both the concept’s units and those things from which they are differentiated.
E.g., for “triangle,” the genus is “polygon,” which is the wider class that
contains triangles, quadrilaterals, pentagons, etc. The differentia is the
characteristic(s) that distinguishes the units from the other existents in
the genus — here “three-sided.” A “triangle” is a three-sided polygon.
(In Objectivist terms, the genus is the class of things having the Concep-
tual Common Denominator, CCD, along which the units are differentiated
from foils; the differentia is a range or category of measurements within
that CCD.)

The definition must consist of a genus and a differentia.

The genus-differentia structure is the most efficient means of isolating

the concept’s units from all other existents. Beginning the definitional
process by identifying the genus narrows one’s focus, so that instead of
considering the entire universe, one is focusing on a class that is only one
step wider than the “species” — i.e., the class of things one is seeking to define.
Identifying the genus gets one into the proverbial “ballpark,” making it fairly
easy to go on to differentiate the concept’s units from their nearest “rela-
tives.” Moreover, the most efficient means of organizing concepts is to set up
a nested progression: species within genus, within still wider genus, and so
on — e.g., “triangle” within “polygon,” within “plane figure,” within “shape.”
The genus-differentia structure of a definition helps one to, in effect,
quickly re-form the concept, and thus recapture the concept’s meaning.
Rand writes:

The rules of correct definition are derived from the process

of concept-formation. The units of a concept were differen­
tiated — by means of a distinguishing characteristic(s) — from
other existents possessing a commensurable characteristic,
 L ogic and C oncepts 217

a “Conceptual Common Denominator.” A definition follows the

same principle: it specifies the distinguishing characteristic(s)
of the units, and indicates the category of existents from which
they were differentiated.
The distinguishing characteristic(s) of the units becomes the
differentia of the concept’s definition; the existents possessing
a “Conceptual Common Denominator” become the genus.
Thus a definition complies with the two essential functions
of consciousness: differentiation and integration. The differentia
isolates the units of a concept from all other existents; the genus
indicates their connection to a wider group of existents. [ITOE, 41]

A proper definition must have both a genus and a differentia. When the
genus is omitted, a mistake often made by uncritical thinkers, the result
is an unorganizable approximation, as in “An ‘automobile’ is what you drive.”
A proper definition of “automobile” must contain its genus: “motor vehicle.”
One popular form of omitting the genus is the barbarous use of “is when”:
“A ‘crime’ is when someone violates another’s rights.” A proper definition
would be: “A ‘crime’ is an action violating another’s rights.”
The other breach of this first rule is the omission of the differentia. “A ‘table’
is an item of furniture.” “ ‘Literature’ is an art-form.” There is nothing wrong
with these statements as statements, but they are not definitions because they
do not isolate the concept’s referents from non-referents within its genus.
A proper definition must not only contain both a genus and a differentia,
it may not add in any extraneous element. A definition consists only of the
genus and differentia.115
“Man is a rational animal.” Not: “Man is a rational animal, like my father.”
Extraneous information distracts from the essentials and violates the rule
of unit-economy (rule 5, below).
Identifying a concept’s genus is comparatively easy but immensely
valuable. It brings clarity to one’s thinking and is indispensable to the inte­
gration of one’s knowledge.

115 In some cases, additional terms are required to meet special cognitive needs. Two such
cases are: 1) including the recognizable, sensory qualities of physical objects (as “flat level
surface” is included in “table”), and 2) including the process by which a product is made
(e.g., for “cake” including that it is baked, and for “knowledge” including that it is based
on perception and/or reason).
 H OW W E K N OW • 7 : L ogic : P ractice 218

A personal example will serve to dramatize the crucial role of the genus.
While teaching college logic, I had one interchange that helped me appre­ciate
the importance of the genus. Having just covered the topic of definitions
and worked through some simple examples, it was time to give the students
a harder term to define. Since inflation was a growing concern at that time,
that was the term I asked the class to define. I began by asking for the genus
of “inflation.” The answer I wanted was something like “economic condition.”
One student volunteered the answer that the genus of inflation is “money.”
But “money” is a kind of thing, an entity, whereas inflation is a condition.
Hoping to lead him to see that he had the wrong basic category (“entity,”
instead of “condition”), I asked the student: “What is the genus of “money”?
Without hesitation, he answered: “survival.” Stunned, but intrigued, I asked
him, “And what’s the genus of survival?” “Food” he answered. We went on
like this through several more concepts, wandering around in his hopelessly
disorganized mental filing system, until I gave up.
This student was connecting concepts by random association, failing to
notice major differences among types of relationships, e.g., subsumption
vs. causation. The organization of his mental files amounted to: inflation
relates to money, money relates to survival, survival relates to food — and
so on. All those relationships do exist, but each is a different type of relation-
ship , and none identifies the genus of its subject.
The failure to specify and differentiate such relationships sabotages
logical thinking. If one has inflation filed under its proper genus — economic
condition — thinking about inflation will activate in one’s mind the proper,
logical context. A logical context will make prominent the data on closely
related economic conditions — in this case: deflation. That is the data that
is most logically relevant. Without that context, one is prey to all sorts of
confusions. For instance, suppose one is considering the popular notion
that inflation is due to greed. If one has filed inflation as the contrary of
deflation, one is primed to notice a problem with the greed-explanation:
if inflation is due to greed, then deflation would be due to lack of greed, too
much self-denial. But that clashes with other facts. For instance, in the last huge
deflation — the Great Depression — people were not less concerned with
(“greedy” for) money.
Identifying the proper genus leads one to ask the right questions — here:
“What distinguishes inflation from deflation?” On a common-sense level,
inflation is marked by a rise in the general level of prices, deflation by a fall.
And that raises a new question: “What is the general price-level?” The genus
 L ogic and C oncepts 219

of “general price-level” is “price.” The genus of “price” is: an amount of money

(the differentia is: “paid per unit in purchase of a good”). A price is thus
a ratio (money spent per good), and the general price-level is the equivalent
aggregate ratio: the economy-wide spending divided by the total quantity of
goods sold. A rise in the general price-level is an increase in that ratio, a fact
drawing our attention to the two things that can make that ratio increase:
an expansion of the money supply or a diminished supply of goods. (More
investigation would be required to get a proper understanding of inflation,
but the foregoing should be sufficient to illustrate the cognitive efficacy
afforded by knowing the proper genus of each of one’s concepts.)
In order to have an organized, efficiently functioning mind, one must have
stored one’s concepts in a nested series of genera. This amounts to having
automatized such connections as: a dog is an animal, an animal is an organ-
ism, an organism is an entity. Or: dollar bills are currency, currency is money,
money is a commodity (or claims to a commodity), a commodity is an entity.
That is the kind of structure of links that brings order to one’s mental files,
making them into a system organized around essentials.
A definition is a proposition; propositions establish a specific relation-
ship between a subject and a predicate, allowing the knowledge stored
in the predicate to be applied to the subject. The species-to-genus relationship
is extremely direct and intimate. It is the relationship of, e.g., tables to furni-
ture, and of triangles to polygons. When one identifies the genus of a species,
the link formed between the species and the genus becomes very strong
— perhaps the strongest form of cognitive link there is, because one of those
concepts (e.g., “furniture”) can be formed by abstracting from the other
(“table”). Therefore, keeping the genus in mind means that the most directly
relevant information is potentiated for entry into conscious awareness.

2. The rule of reference

A definition isolates the units of a concept from all other existents. In order
to do that, its formulation must actually direct the mind beyond mere words,
to what those words stand for.

The definition must specify a group of referents in reality.

There are at least four ways in which a definition can fail to identify the
concept’s referents and thus violate this rule.
 H OW W E K N OW • 7 : L ogic : P ractice 220

1. Synonymy: One cannot define a term by simply giving synonyms, nor

can a definition depend upon the use of a synonym of the term being defined.
Relying on synonyms simply transfers the question to the definition of the
synonym, rather than pointing out referents in reality. E.g., take: “Fear is an
emotion of being afraid.” Since “afraid” is a synonym of “fear,” this attempted
definition does not specify referents in reality but merely switches the bur-
den onto “afraid.” (A valid definition of “fear,” in general terms, would be:
an emotion resulting from the evaluation of something as a threat to oneself
or one’s values.)
2. Circularity: A is defined in terms of B, when B has to be defined
in terms of A. E.g., “Economics is the science pursued by economists.”
But “economist” has to be defined in terms of “economics.” (The circularity
may involve more than just two steps: it is still a circular definition when
A is defined in terms of B, which is then defined in terms of C, which is then
defined in terms of A.)
3. Vagueness: If the terms of the proposed definition are vague
or unclear, referents are not specified. E.g., “Faith is the substance of things
hoped for, the evidence of things not seen.” Or Tolstoy’s definition of “art”:
“Art is a human activity having for its purpose the transmission to others of
the highest and best feelings to which men have risen.” This definition fails
to differentiate art from other human activities — e.g., from oratory and
writing love-letters.
4. Metaphor: Figurative language cannot substitute for literal truth. E.g.,
“Architecture is frozen music” makes an interesting comparison of music to
architecture, but the statement does not serve the function of a definition:
one could not use it to demarcate instances of architecture in reality.

3. The rule of scope

If a definition satisfies the rule of reference, so that it does isolate a specific
set of existents, the next question that arises is: has it picked out the right
set — i.e., exactly the things that are the units of the concept being defined?
The definition’s scope must match that of the concept; the definition must
apply to all of the concept’s (normal) units and only to them. In the negative
formulation: a definition must be neither too broad nor too narrow.

The definition must have the same scope as the concept that it defines.
 L ogic and C oncepts 221

The issue here is truth. A definition that is too broad is false qua defin­
ition: it implies that things are units that are not, or vice-versa. For instance,
“A ‘table’ is an item of furniture with a flat top surface” implies that beds are
tables. A definition that is too narrow implies that some things that are units
are not units. For instance, “A ‘table’ is an item of furniture with four legs
and a flat, level surface for supporting smaller objects” implies that six-
legged tables are not tables. A proper differentia must characterize all and
only the units within the genus.
A proposed definition can easily be both too narrow and too broad.
For instance, “Man is a white animal” is both too narrow, excluding non-
Caucasians, and too broad, including polar bears.
Plato, in his dialogue the Meno, raised a problem that has bedeviled this
rule of definition: don’t we need to know the definition in order to know
what the concept includes? But the rule directs us to judge the definition
by reference to what the concept includes. For instance, how do we know
whether or not a six-legged table is to count as a table, prior to having the
definition of “table”?
The answer to this apparent circularity becomes quite obvious once
one recognizes the basic fact that concept-formation precedes definition.
Prior to having a definition of a concept, we know what its units are.
We start from observation of reality, not from definitions. The defini-
tion is the final step in the process, not the first. The first step is obser-
vation of similarities and differences — the definition merely makes
explicit the similarity involved. We are aware pre-definitionally that
a six-legged table is similar to the other things we are classifying as tables.
The scope of a concept is determined by fundamental similarities and differ-
ences, and these are grasped prior to formulating a definition. The file exists
before one writes a label for its folder.
Sometimes one does indeed learn a concept by hearing another
person’s definition of it, but this is learning from others, not learning from
direct observation of reality. And however many men learn the concept from
another man’s definition, someone must have formed the concept for himself
in order that there be anyone to teach it to others. Furthermore, the stu-
dent must bring to mind examples of the concept’s referents in order to
properly understand the concept. The fundamental order here is: first one
has in mind the units of a concept, then one begins to formulate a definition
of that concept.
 H OW W E K N OW • 7 : L ogic : P ractice 222

Not only is the scope of the concept set before the defining process begins,
the definition, once formed, is not a substitute for the concept. The definition
is merely the summary of the units’ nature — the label on the file folder.
When applying the rule of scope: the “all and only” requirement applies
to the normal units of the concept. For instance, in defining “man” one is
not concerned to draw the line in evolution at which man evolved from
the apes. Borderline cases are borderline cases, not tests of the definition.
This follows from the fact that a concept pertains to a range of measurements.
There is an issue of degree here, and one does not need, and cannot demand,
infinite precision. A proper definition allows for borderline cases.
Likewise, it is no objection to defining “man” as “the rational animal” that
immature, senile, or brain-damaged human beings lack the ability to use
reason. These people are still members of the human species, and that is
the species whose means of survival is the use of reason — which is what
“rational” actually means in the definition. (Moreover, the genus “animal”
already includes the fact of immature, defective, or diseased instances.)
In general, defining is not a mechanical process; judgment is required.
The primary purpose of a definition is not to provide criteria for deci­
ding whether or not a given item is to be subsumed under a given concept.
Though a definition can serve as an aid in classifying an item, a definition is
primarily for use in the other cognitive direction: pointing from a concept
to its referents. Definitions tie a concept to its units, the units that constitute
its meaning.

4. The rule of fundamentality

The next rule becomes necessary when, as often happens, there is more
than one characteristic (or set of characteristics) that could serve to isolate
the concept’s units from the other existents within the genus. For example,
all (normal) men and only men have a certain kind of shape, have the abil-
ity to make tools, have the ability to reason. Which of these distinguishing
characteristics should be used as the differentia in the definition of “man”?
The rule of fundamentality tells us:

The definition must state the fundamental distinguishing characteristic(s).

This rule is traditionally phrased in terms of stating the essential charac-

teristic of the concept’s referents. The essential characteristic is one of the
 L ogic and C oncepts 223

distinguishing characteristics, the one that is fundamental. This fundamen-

tality is a causal issue: the fundamental characteristic is the one that causes
and explains the greatest number of other characteristics. Rand writes:

When a given group of existents has more than one characteristic

distinguishing it from other existents, man must observe the
re­­lation­ships among these various characteristics and discover the
one on which all the others (or the greatest number of others)
depend, i.e., the fundamental characteristic without which the
others would not be possible. This fundamental characteristic
is the essential distinguishing characteristic of the existents
involved, and the proper defining characteristic of the concept.
Metaphysically, a fundamental characteristic is that distinctive
characteristic which makes the greatest number of others possible;
epistemologically, it is the one that explains the greatest number
of others.
For instance, one could observe that man is the only animal
who speaks English, wears wristwatches, flies airplanes, manufac-
tures lipstick, studies geometry, reads newspapers, writes poems,
darns socks, etc. None of these is an essential characteristic:
none of them explains the others; none of them applies to all men;
omit any or all of them, assume a man who has never done any of
these things, and he will still be a man. But observe that all these
activities (and innumerable others) require a conceptual grasp
of reality, that an animal would not be able to understand them,
that they are the expressions and consequences of man’s ratio-
nal faculty, that an organism without that faculty would not
be a man — and you will know why man’s rational faculty is his
essential distinguishing and defining characteristic. [ITOE, 45–46]

Consider, for example, the attempt to define man as “the tool-making ani-
mal.” This definition fails the test of fundamentality, because man’s ability
to make tools is a consequence of his ability to use reason. The same error
is committed, less grossly, by the common definition of man as “the ­animal
that uses language.” What explains man’s ability to speak a language?
His ability to form concepts — i.e., his rational faculty.
Note that both of the inferior definitions create further conceptual muddles.
Is a bird’s nest or a beaver’s dam a “tool”? Is the twitter of birds a “language”?
 H OW W E K N OW • 7 : L ogic : P ractice 224

The “dance” done by bees to indicate the location of a pollen-rich field has been
called a “language.” But an actual language uses propositions composed of
concepts, and concepts are formed by a process of reason. Thus, fundamentality
dictates that the definition be: “man is the rational animal.”
(In general, concepts must be not only defined but also formed in the first
place on the basis of fundamental, not superficial, similarities and differences.
More on this will follow.)
The rule of fundamentality explains and subsumes the traditional rule
against negative definitions (“A definition must not be negative where it
can be positive”). Negatives are grasped in relation to the positives from
which they differ, so negatives are non-fundamental. For example, take this
negative statement: “Man is the animal that is not limited to the perceptual
level.” The “not limited” is a consequence, not the fundamental: man is not
limited to the perceptual level because he has a rational faculty.
It must be noted, however, that some concepts are inherently negative.
For instance, “dry” is the absence of moisture. “Blindness” is the absence
of the ability to see. “Freedom” is the absence of coercion. The rule of funda-
mentality entails that a definition must be negative when the concept being
defined is negative — i.e., specifies the absence of something.
Finally, note that to be fundamental a characteristic need not explain
all the other distinguishing characteristics: the requirement is that it must
explain the greatest number of distinguishing characteristics. It is not
obvious, for example, that man’s possession of an opposable thumb is
explained by his possession of a rational faculty. But man’s possession of
a rational faculty does explain an enormous number of his distinguishing
characteristics, incomparably more than does any other candidate.

5. The rule of unit-economy

A definition is not an exhaustive summary of the units’ characteristics.
Nor is it a substitute for the concept. A definition is a cognitive
tool — a means of quickly recalling the nature of the units, as the label on
a file folder informs one of the nature of the folder’s contents. Accordingly,
the definition must be short — the shorter, the more condensed, the better.
(Shorter definitions also make the genus prominent, which aids organizing
one’s concepts into a series of progressively wider genera.)

The definition must be a single, economical sentence.

 L ogic and C oncepts 225

Consider the following as a definition of “man”: “Man is the animal whose

faculty of awareness possesses, from the time of late infancy through senility,
in normal cases, the capacity for the formation and use of concepts.”
Obviously, that won’t do. It would not serve as a “label” for one’s “file.”
Aristotle’s definition of man is unit-economical: “Man is the rational animal.”
Another example of a violation of this rule (and others) is John Dewey’s
definition of “inquiry”: “Inquiry is the controlled or directed transformation
of an indeterminate situation into one that is so determinate in its constituent
distinctions and relations as to convert the elements of the original situation
into a unified whole.” [Dewey, 104] A unit-economical definition of “inquiry”
would be: “Inquiry” is a systematic process of gathering information about
a delimited subject.
These five rules are not arbitrary dicta. They are statements of what is re­quired
for the definition to perform its function. That function is: to enable one to
know what one is talking about — i.e., which things one is referring to and what
their essential nature is. Definitions allow concepts to function as concepts,
rather than as inarticulate sounds or floating abstractions. Defining a concept
in compliance with these rules, if done rationally and tested against a range
of examples, gives one’s concepts a firm identity in one’s mind and ties them
to their referents in reality.116

Definitions as Contextual
A definition is a condensation of knowledge, knowledge of a concept’s units
and of their place in the entire structure of one’s knowledge. Knowledge is not
frozen in a static sum; it grows in the individual’s development, as he advances
from childhood to educated adult, and it grows with the progress of science.
Definitions, to function as optimal condensers of knowledge, must expand
to keep abreast of an expanding context of knowledge.
Definitions, in other words, are contextual. They are established in a given
context of knowledge, and they are to be judged by reference to the context
of knowledge in which they are used. A broadened or deepened knowledge
of the units requires a corresponding change in the concept’s definition.

116 have found that the best method of developing a definition on one’s own is to: 1) select two
or three clear and simple examples of the units and, as foils, one or two “near relatives”;
2) identify the genus that includes all of the selected examples; 3) ask what fundamentally
differentiates the units from their “relatives” in that genus; 4) package the results into
a clear, economical sentence. (The rules are one’s guide throughout the process and, until
they have been automatized, it is a good idea to check one’s definition against each.)
 H OW W E K N OW • 7 : L ogic : P ractice 226

Take the concept “gold.” In a primitive context of knowledge, the (implicit)

definition of “gold” might be “a hard substance of a particular, yellowish color.”
This definition, I am assuming, is sufficient to distinguish gold from all other
materials known at the time. Later, however, other hard, yellow substances,
e.g., “fool’s gold,” are encountered. At that point, the statement “gold is a hard,
yellow substance” — though still true as a statement — no longer serves to
differentiate gold from other things known. Accordingly, the definition must
be expanded (not contradicted, but expanded) by adding new differentiae.
Suppose the other yellow substances are not as dense and malleable as gold.
Then, at this point, the definition of “gold” would be expanded to become:
“dense, malleable yellow substance.”
That definition may suffice for quite a while, but eventually, as civilization
develops, men will create alloys, e.g., of gold and silver. Since these alloys
look and feel too similar to gold to be distinguished perceptually, the expanded
definition will include more advanced differentiae, such as “insoluble
in such and such acids,” and (per Archimedes’ discovery), “having such and
such a density relative to water.”
As science progresses, the genus of the definition must also be expanded.
When the concept “metal” is formed (from observation of different metals),
the genus of “gold” becomes “metal.” Still later, “metal” will be refined to “element.”
These expansions of the genus are necessary to maintain the definition’s role
in organizing one’s concepts.
When the atomic theory of matter is established, a new change will be
required: “gold” will be defined in terms of its atomic number, which states
the number of protons in the gold atom. The result is our present definition:
“gold” is an element having the atomic number of 79. In our current context
of knowledge, this definition states the fundamental. It is the fact that gold
has 79 protons that causes and explains its having 79 electrons, which, in turn,
causes and explains its possession of its distinctive properties — including
its malleability, its density, and its characteristic yellow color.
Throughout the evolution of these definitions, gold remains gold, with all
the properties that characterize it. Gold does not change, nor does the concept
change. And the earlier definitions remain true as statements: gold is still
yellow, hard, malleable, etc. But reference to these properties no longer satis-
fies the cognitive function of a definition.
In all such cases of definitional revision, the facts stated by the earlier defi­
nitions remain facts, the old definitions remain true as statements, but they
no longer suffice to define the concept. Thus, the change in definition
 L ogic and C oncepts 227

required by the growth of knowledge represents a process of expanding, not

of contradicting, the concept’s earlier definition.
The cases calling for definitional expansion are: 1) the discovery of
new concretes (e.g., “fool’s gold”) that need to be distinguished from the
concept’s units; 2) the gaining of new knowledge about the genus (“element”);
and 3) the discovery of a deeper characteristic (atomic number 79) that
causes and explains those previously used as differentiae.
As an example of definitional expansion, Rand cites the case of a child’s
evolving definitions of “man” (definitions that are, of course, only implicit
for the child). The child’s first implicit definition of “man” might be,
she suggests, “a thing that moves and makes sounds.” At a later stage, after
the child learns more, his (implicit) definition might become: “a living
thing that walks on two legs and has no fur,” and, still later: “a living being
that speaks and does things no other living beings can do.” As an adult,
he reaches a still deeper, and now explicit, definition: “a rational animal.”117
Man’s possession of a rational faculty is the fundamental characteristic
causing and explaining the vast complexity of “does things no other living
beings can do.”
If one recognizes that definitions are cognitive tools, not a substitute
for the full concept, there is no problem in embracing the contextuality of
definitions.

Valid vs. Invalid Concepts

The rules of definition presuppose that one has formed a valid concept.118
Though we live in a “non-judgmental” age, the fact is that the concepts one forms
are either right or wrong: a given conceptualization is either pro-cognition or
anti-cognition. Rand writes:

There are such things as invalid concepts, i.e., words that represent
attempts to integrate errors, contradictions or false propositions,
such as concepts originating in mysticism — or words without
specific definitions, without referents, which can mean anything
to anyone. . . . Invalid concepts appear occasionally in men’s

117 Rand also endorsed the idea that a biologist’s definition of “man,” as opposed to a general
philosophic definition, would include sophisticated terms like “primate.” [ITOE, 235]
118 Aristotle noted that a concept of imagination, like “unicorn,” does not have a definition,
but only a “formula” — i.e., a sentence explaining how the word is used in fantasy.
 H OW W E K N OW • 7 : L ogic : P ractice 228

languages, but are usually — though not necessarily — short-

lived, since they lead to cognitive dead-ends. An invalid concept
invalidates every proposition or process of thought in which it is
used as a cognitive assertion. [ITOE, 49]

The two basic mistakes that produce an invalid concept are: 1) making
a concept for non-existent units, or 2) making a concept that uses an invalid
standard, resulting in misclassifying units.

1. Concepts that Lack Units

Concepts lacking units are those that attempt to refer to the contradictory or
to the arbitrary. The test for the validity of a concept here is whether or not
it can be reduced to perceptual reality. (For an example of this process, see
Peikoff ’s reduction of “friend” through such intermediate concepts as “man,”
“knowledge,” “affection,” and “esteem.” [OPAR, 134–135]) If a concept cannot be
reduced to any perceptual base, it is an error, distortion, or fantasy.
A made-up example of a concept that does not reflect facts would be
“biangle”: a two-sided polygon — which is impossible.
Actual examples of invalid concepts are, of course, more controversial,
since to be common enough to figure as “real-life” examples, many people
must take them as valid. But it is worthwhile to discuss some concepts that
I, at least, regard as invalid, because doing so will illustrate how the rules
of epistemology have important application to real-world issues.
The first case is that of invalid concepts “originating in mysticism,”
in Rand’s phrase above. These would include “god,” “fairy,” “angel,” “devil,”
“afterlife,” etc. These concepts are nothing but arbitrary inventions,
rather than being based in any way on logical processing of perceived facts.
The fact that their origin lies in make-believe becomes clear when one
contrasts them with valid, scientific concepts that designate non-perceivable
existents. “Electron” and “ultraviolet” are concepts reached by logical
inference from what is perceived — whereas “god” and the like are based
on nothing but primitive men’s awe before such phenomena of nature
as lightning, floods, and earthquakes. (Concepts of imagination, like “angel”
and “hobbit,” can have a valid role in imaginative literature, but they are
invalid if used “as a cognitive assertion,” in Rand’s careful phrasing.)
The injunction to reduce a concept to the perceptual level is not the
demand: “Show me referents I can see, touch, or otherwise perceive by my
 L ogic and C oncepts 229

senses.” Rather, it is the demand: “Show me how, by logical steps, start-

ing from perception, one can reach this concept — and why one needs it.”
Our senses cannot perceive cosmic rays or genes or a budget-deficit, but each
of these concepts is definitely reducible, through a long chain of intermediate
concepts and knowledge, to perceptual data.
It is not just religion that is guilty of using mystical or groundless
concepts. The history of philosophy offers an embarrassingly large array of
non-reducible, arbitrarily constructed concepts. Some examples, of many, are:
Plato’s “Forms,” Spinoza’s “intuition,” Leibniz’s “monads,” Kant’s “noumena,”
Hegel’s “dialectic,” Marx’s “forces of production.” All of these terms, and
many more, are introduced without any logical derivation from perceptual
reality. They are all invalid concepts, of interest only to the historian.
There are also scientific concepts, such as “epicycle” and “phlogiston,”
now known to contradict observed data (even if they were legitimate
as hypothetical concepts in an earlier context of knowledge). But even
contemporary science (or pseudo-science) exhibits invalid concepts, such as the
concepts deployed by “para­psychology” (itself an invalid concept): “telepathy,”
“telekinesis,” etc.
In contemporary physics, the accusation raised against “string theory”
is that it is an arbitrary construct, based on playing with mathematics,
rather than supported by observational data. I am not a physicist, but if this
accusation is correct, its concepts, notably including “eleven-dimensional
space,” are invalid.
The examination of concepts for validity is much needed in regard to
concepts used in the social sciences . Freud’s concepts of “Id” and “death-
instinct,” the concept of “socialization” in sociology, and the concepts of
“perfect competition” and “externalities” in economics are all arguably invalid,
most of them on grounds of being arbitrary constructs, lacking units in
reality. But “socialization” and “externalities” are terms commiting a much
subtler type of error: the misclassification of units.

2. Concepts that Misclassify Units

Concepts in this category deal with actually existing phenomena, but
mis-organize them, classifying things in a way that is confusing, mislead-
ing, or otherwise anti-cognitive. How can a concept be anti-cognitive?
Isn’t any “conceptual scheme” just as good or bad as any other, provided
that its units exist? No — no more than any filing system is as good or bad
 H OW W E K N OW • 7 : L ogic : P ractice 230

as any other. Here is an extended example that demonstrates the folly of this
non-judgmental, relativist attitude.
Suppose a biologist were to set up a classification system for living organisms
(a “taxonomy”) based on superficials instead of fundamentals. In making the
highest-level distinction, covering all organisms, he rejects the traditional
division between the plant kingdom and the animal kingdom, distinguishing
instead between “stripes” and “solids” — according to whether or not the surface
of the organism shows stripes. This would mean the “stripes” kingdom
would contain, for instance, zebras and tigers, along with certain snakes, bees,
and plants; these would be divided from the “solids” kingdom, which would
contain horses and lions, along with other snakes, bees, and plants — plus all
single-celled organisms.
Now imagine making further subdivisions along equally superficial lines,
such as whether an organism’s stripes are monochromatic or colored, and
then whether its shape is elongated or not. This conceptual scheme, though
absurd, could be entirely consistent: all and only the “stripes” would have
stripes, all and only the “longs” would have relatively elongated shapes.
So, for instance, both what we now call “whales” and what we now call “oaks”
would be “elongated, monochromatic solids.” Both what we now call “tigers”
and what we now call “honeybees” would be “polychromatic stripes,” while
lions and other bees would be “monochromatic, non-long solids,” and so on.
The problem with this taxonomy does not concern any lack of referents
or any contradiction to observed facts. Instead, the problem concerns how
the referents are being grouped. The classifications are made on the basis
of non-essentials — i.e., superficial similarities and differences rather than
fundamental ones. Classifying by superficials puts fundamentally different
things together and separates things that are fundamentally the same.
The result is the kind of groupings that frustrate the needs of cognition,
turning biological classification into an unintegrable jumble. Rather than
providing cognitive leverage, rather than aiding cognitive specialization (as
between botany and zoology), the stripes-solid taxonomy based on superfi-
cials would make that specialization pointless. What is to be gained by spe-
cializing in zebras and honeybees while leaving horses and wasps to those
who study “solids”? This taxonomy would result in the utter stultification
of biological science — if any biologist could actually stick to it, rather than
covertly reverting to the traditional system of classification.
This example makes it dramatically clear that, contrary to the claims of
conceptual relativism, there is an objective right and wrong in how we classify.
 L ogic and C oncepts 231

The basis of that right and wrong lies in the need to obey the identity of man’s
conceptual equipment. Specifically, it is “the crow epistemology” that gives
rise to the need for economizing on units, and thus for judging classifications
from the standpoint of their unit-economy.
Carving nature at the joints — i.e. on the basis of fundamentals — provides
the most unit-economical system of classification. What makes a characteristic
fundamental is that it causes and explains the greatest number of other
characteristics. Thus, in a classification by fundamentals, the greatest number
of characteristics can be condensed into one unit and treated wholesale.
E.g., horses and zebras have many characteristics in common, allowing for
a condensation of units in dealing with them together, whereas far fewer
characteristics are common to zebras and honeybees, and their differences
overload the “crow.” A concept formed on the basis of superficial similarities,
ignoring fundamental differences, is invalid.
To validate a concept and establish its proper definition, one should ask
oneself: What facts of reality give rise to the need for such a concept?
If no such facts are ascertainable, the concept cannot be considered valid,
and one should not attempt to use it cognitively. But most of the concepts
that we form remain in the language because they do reflect fundamental
similarities. In those cases, identifying the need for the concept brings those
fundamentals to light, helping one to form a proper definition.
Rand provides an intriguing example of this process in regard to the
concept “justice”:

For instance: what fact of reality gave rise to the concept

“justice”? The fact that man must draw conclusions about the
things, people and events around him, i.e., must judge and evalu-
ate them. Is his judgment automatically right? No. What causes
his judgment to be wrong? The lack of sufficient evidence, or his
evasion of the evidence, or his inclusion of considerations other
than the facts of the case. How, then, is he to arrive at the right
judgment? By basing it exclusively on the factual evidence and
by considering all the relevant evidence available. But isn’t this
a description of “objectivity”? Yes, “objective judgment” is one
of the wider categories to which the concept “justice” belongs.
What distinguishes “justice” from other instances of objective
judgment? When one evaluates the nature or actions of inani-
mate objects, the criterion of judgment is determined by the
 H OW W E K N OW • 7 : L ogic : P ractice 232

particular purpose for which one evaluates them. But how does one
determine a criterion for evaluating the character and actions of
men, in view of the fact that men possess the faculty of volition?
What science can provide an objective criterion of evaluation in
regard to volitional matters? Ethics. Now, do I need a concept
to designate the act of judging a man’s character and/or actions
exclusively on the basis of all the factual evidence available,
and of evaluating it by means of an objective moral criterion?
Yes. That concept is “justice.” [ITOE, 51]

Rand’s Razor
In the preceding passage, note her final question: “Now, do I need [such]
a concept?” An answer of “Yes” is not to be simply assumed (particu-
larly not in the case of neologisms). As we saw in the example of “stripes”
vs. “solids,” not every recurring fact warrants the formation of a new con-
cept; the vast majority of such facts are, properly, identified in descriptive
propositions, such as “Tigers have stripes.” In contrast, the fact that there
is a fly on the ceiling theoretically could be, but wouldn’t be, elevated into
a permanent, open-ended concept — say, “ceiling-flies.” But properly, one
handles that fact descriptively by combining into a proposition the existing
concepts “fly,” “on,” and “ceiling.”
Forming a concept is not a free lunch; there are “overhead” costs
involved in storing it, carrying it forward, updating it, etc. One should form
a new concept only when the cognitive gains of doing so exceed the cost.
If the concept would not be thus cognitively profitable, forming and
using it constitutes a waste of mental resources, complicating one’s men-
tal filing system. Needless multiplication of concepts results in decreased
cognitive efficacy.
In recognition of this fact, Ayn Rand formulated an epistemological
version of Occam’s Razor, which has come to be called “Rand’s Razor”:

. . . concepts are not to be multiplied beyond necessity — the cor-

ollary of which is: nor are they to be integrated in disregard
of necessity. [ITOE, 72]

What is meant by “necessity” here? What factors would make a new

concept necessary? Rand names three such factors:
 L ogic and C oncepts 233

“The descriptive complexity of a given group of existents, the fre-

quency of their use, and the requirements of cognition (of further
study) are the main reasons for the formation of new concepts.
Of these reasons, the requirements of cognition are the para-
mount one.” [ITOE, 70]

Rand discusses the kinds of concepts whose formation is necessary,

according to these criteria:

There is a great deal of latitude, on the periphery of man’s con-

ceptual vocabulary, a broad area where the choice is optional,
but in regard to certain central categories of existents the forma-
tion of concepts is mandatory. This includes such categories as:
(a) the perceptual concretes with which men deal daily . . .
(b) new discoveries of science; (c) new man-made objects which
differ in their essential characteristics from the previously known
objects (e.g., “television”); (d) complex human relationships
involving combinations of physical and psychological behavior
(e.g., “marriage,” “law,” “justice”). [ITOE, 70]

Failing to conceptualize “television” or “marriage” would mean depriving

oneself of a file folder one needs.
In contrast, consider the case of flatly unnecessary concepts. The worst
of these are not concepts that merely duplicate existing ones, but rather
concepts formed on the basis of non-essentials, such as “ceiling-flies.”
It is clarifying to consider here the analogy with physical (or digital) files,
say for correspondence. To make files for “letters received when the moon
was full,” “letters opened while I was standing up,” and so on, would defeat
the purpose of a filing system. A policy of making files like these when-
ever one was struck by any random similarity would produce chaos — just
as forming concepts to capture random similarities would produce chaos
in one’s inner, mental office.
The chaos is compounded by the fact that concepts become automatized:

A concept substitutes one symbol (one word) for the enormity of

the perceptual aggregate of the concretes it subsumes. In order
to perform its unit-reducing function, the symbol has to become
automatized in a man’s consciousness, i.e., the enormous sum
 H OW W E K N OW • 7 : L ogic : P ractice 234

of its referents must be instantly (implicitly) available to his

conscious mind whenever he uses that concept, without the need
of perceptual visualization or mental summarizing — in the same
manner as the concept “5” does not require that he visualize five
sticks every time he uses it. [ITOE, 64]

Automatizing needless concepts intensifies their destructiveness. Automa-

tized concepts spring to mind when triggered, leaving less mental space avail-
able for retrieving and using proper concepts (per “the crow epistemology”).
Concepts permit a cognitive division of labor, with all the benefits flowing
from specialization. But specialization is beneficial only under certain con­
ditions — only when it is a specialization in things sharing a fundamental
unity. Imagine a business formed to specialize in products that begin with the
letter “N” — nails, nasal sprays, nasturtiums, etc. That kind of “specialization”
would bring not profits but bankruptcy. In the same way, cognition is not
aided but obstructed by forming a concept on the wrong basis, i.e., according
to superficials at the expense of fundamentals.
In considering the cognitive cost of forming a concept, one must keep in
mind the fact that a concept is something that is to be carried forward, used,
and updated. Concept-formation is only the beginning of an ongoing process,
not something that is over and done with after performing the initial mental
integration, nor even after defining the concept. Concepts are open-ended
because they are tools of cognition, tools that are formed in order to be
used. The raison d’être and purpose of conceptualization is application
— application in forming propositional judgments, in the acquisition of
knowledge, and in planning, deciding, and acting to achieve existential goals.
Concepts are a human being’s basic equipment for successfully coping with the
problem of survival.
It is against this background that the choice to add or not to add an
additional concept must be evaluated. The question is: would forming this
additional concept aid cognition, clarify thinking, add to my life-guiding
equipment — or would it merely add clutter, distract me from fundamentals,
and lead to dead ends in thought and action?
To better understand violations of Rand’s Razor, recall the mechanics of
concept-formation (Chapter 3). A concept is formed by grasping simi-
larity (measurement-proximity) against difference. In what way can this
process go wrong? Not in regard to grasping similarity: there is no problem
in correctly comparing the measurements involved. In contrasting a circle,
 L ogic and C oncepts 235

an ellipse, and a triangle no one makes a mistake about which figures are
close in shape-­measurement; it is perceptually evident that the circle and
the ellipse have similar shapes when contrasted with the triangle. Nor, on
a much higher level of abstraction, does anyone take an inflation rate of 8%
to be more like a deflation rate of 10% than an inflation rate of 10%. In terms
of measurement-relationships, no one mistakes proximity for remoteness.
It is not errors about what is nearer in measurement that create invalid
concepts. Instead, it is the set-up itself that is at fault: one chooses the wrong
items to compare, thereby establishing the wrong CCD. Comparing a zebra
and a striped snake to a lion brings to the foreground the issue of stripes
vs. lack of stripes, which is an extremely superficial characteristic in the con-
text of an overall biological taxonomy. The primary cause of misorganizing
the referents — i.e., of violating Rand’s Razor — is using the wrong CCD.

False division
Concepts that violate Rand’s Razor are categorizable as either “false division”
or “false integration.”
Concepts that I am terming “false divisions” are those that make subdivi-
sions based on non-essential differences. These concepts focus attention on
superficial or cognitively insignificant differences at the expense of funda­
mental similarities.
Rand provides a made-up example: “Beautiful blondes with blue eyes,
5’5” tall and 24 years old.” [ITOE, 71]
Suppose we were to make a concept for this group of women — call
them “blins.” There would be no cognitive gains to be achieved by cognitive
specialization in blins. Virtually all the facts that one could learn about blins
would be equally true of beautiful blondes with blue eyes who are 26 years old
or 5’2” tall — despite the fact that the concept “blin” calls for us to canonize,
in a concept, the distinction among them.
Further, if “blin” is to be formed, by what principle would one refuse to
form concepts for every other combination and concatenation of charac-
teristics? Either one adheres to the principle of unit-economy, or one does
not — in which case one’s mental filing system becomes progressively more
clogged and progressively less functional.
“Blin” is a deliberately absurd concept. But consider a real-life, political
example of false division. Since the 1930s, the political spectrum has been
 H OW W E K N OW • 7 : L ogic : P ractice 236

conventionally divided between two poles: fascism as the extreme “right” and
communism as the extreme “left.” This is a lethal false division. There are no
fundamental differences between fascism and communism. Both are forms
of dictatorship. Both are based on the theory and practice of collectivism.
They differ in their outward forms — fascism demands the submission of
the individual to a national and/or racial collective, communism demands
the submission of the individual to an economic collective — but, both deny
completely the rights of the individual, requiring him to live for the state.
This is not to claim that the two systems are identical; a few minor, quite
secondary, distinctions can indeed be made between fascist dictatorship and
communist dictatorship. Rather, the point is that it is disastrously wrong to
set up as the poles of the whole political spectrum two varieties of dic­ta­
torship. (Note the absurd implication that a free society occupies a “middle”
between two forms of dictatorship, as if it combined elements of each.)

False integration
Concepts that I am terming “false integrations” are widenings based on
non-essential similarities, at the expense of essential differences. For
instance, suppose I were to make up the concept “cutter” to denote anyone
who cuts human flesh with a sharp instrument. “Cutter” would include both
surgeons and Jack the Ripper. Clearly, automatizing and working with
“cutter” would produce cognitive chaos and lead in practice to gross injustice.
It is not that one couldn’t tell who was and wasn’t a “cutter,” but that the CCD
here — flesh-cutting — is grossly superficial.
“Cutter” is a made-up example. In actual practice, false integration
frequently occurs when an existing and perfectly valid concept is given
a “definition by non-essentials,” in violation of rule 4. Defining an otherwise
valid concept by non-essentials converts it into something that “integrates
in disregard of necessity,” making the concept into what Ayn Rand terms
an “intellectual package-deal.”
For instance, Bertrand Russell defines “freedom” as: “the absence of
obstacles to the realization of one’s desires.” Consider some of the known
facts — the wider context — that this definition ignores. To a 17th-century
man in Northern Italy seeking to travel into Switzerland, the Alps are an
obstacle to the realization of his desires. Do the Alps deprive him of his
freedom? To a man who wishes to have sex with a given woman, her
unwillingness is an obstacle to the realization of his desires. Does her
refusal deprive him of his freedom? What about her freedom, which would
 L ogic and C oncepts 237

be violated by his act? And what about historical conditions? In contem-

porary America, with its unprecedented standard of living, there are
— in one sense — far fewer unrealized desires than there were in 1800.
Does that mean that individuals in contemporary America have more
freedom than Americans living in 1800?
Russell’s definition is centered on a man’s desires, using their frustration
or fulfillment as the CCD. But a desire results from evaluations that may
be rational or irrational, correct or incorrect. Desires may even come from
evaluations that flout metaphysically given facts: recall Dostoyevsky’s “under-
ground man” who did not like the fact that two plus two equals four
(“The Formula ‘two and two make five’ is not without its attractions.”)
But Russell’s definition treats desires as irreducible givens, when in fact they
are effects, not primaries. His definition converts “freedom,” a crucial term
of political philosophy, into a “package-deal.”
A definition of freedom grounded in all the facts I have mentioned would
result in a concept quite different from Russell’s; it would be a concept
defined in terms of the absence of physical coercion by others. But Russell’s
concept treats human acts of coercion as if they were essentially the same
as inanimate obstacles, the laws of nature, and even the laws of logic
(which present the “obstacle” that you can’t have your cake and eat it, too).
Russell’s definition converts “freedom” into a massive package-deal, obscur-
ing the life-and-death need to be free from coercion by human beings, who
act by deliberation and choice, not necessity (see Chapter 10).
Several neologisms that have been injected into contemporary ­political
discussion exhibit the same kind of anti-cognitive focus on ­superficials
at the expense of fundamentals. The arch-example, as Rand explains,
is “extremism.”

. . . “extremism” is a term which, standing by itself, has no mean-

ing. The concept of “extreme” denotes a relation, a measurement,
a degree. The dictionary gives the following definitions: “Extreme,
adj. — 1. of a character or kind farthest removed from the
ordinary or average. 2. utmost or exceedingly great in degree.”
It is obvious that the first question one has to ask, before using
that term, is: a degree — of what?
To answer: “Of anything!” and to proclaim that any extreme is
evil because it is an extreme — to hold the degree of a character­
istic, regardless of its nature, as evil — is an absurdity (any garbled
Aristotelianism to the contrary notwithstanding). Measurements,
 H OW W E K N OW • 7 : L ogic : P ractice 238

as such, have no value-significance — and acquire it only from

the nature of that which is being measured. [CUI, 177–178]

In politics, the issue that is fundamental is: freedom vs. state coercion.119
But the term “extremism” puts together, as if it were one basic phenomenon,
positions that are “extremely” pro-freedom and “extremely” anti-freedom.
It packages into one “file” and treats as equivalent George Washington
and Adolf Hitler. Washington was “extremely,” i.e., radically and consis-
tently, dedicated to establishing a nation based on the free exercise of the
inalienable rights of the individual. Hitler was “extremely” dedicated to
eradicating individual freedom and establishing a totalitarian state based
on racial collectivism. Packaging them together under one concept is a crime
against logic (and morality). “Extremist” whitewashes Hitler and Nazism by
asso­ciating them with Washington, and it blackens Washington and liberty
by associating them with Hitler.
It is a fact that both Washington and Hitler held extreme views. And it is
a fact that both zebras and certain plants have stripes. But in neither case is
there justification for selecting these facts, rather than fundamentals, as the
basis for forming a new concept (or mis-defining an existing one).
The formation of a concept has to be justified. Concepts, though not
pre-set by nature, are to be judged by objective standards. Justifying
a concept requires showing that it refers to existents, rather than fantasy
or error, and that it is pro-cognition, i.e., that it satisfies “Rand’s Razor,” rather
than christening an unnecessary division or obscuring fundamental differ-
ences. The rules of proper definition, especially the rule of fundamentality,
serve as checks to ensure that one has not violated Rand’s Razor by sacrifi­cing
fundamentals to superficials.
For completeness, I will discuss briefly the “optional” cases — concepts
whose formation is neither mandatory nor prohibited. Such concepts
repre­sent “borderline cases,” in the sense that they fall between mandatory
concepts and invalid ones.
“Optional” concepts, Rand writes, are mainly “subdivisions that denote
subtle shades of meaning, such as adjectives which are almost, but not
fully, synonymous.” [ITOE, 72] Consider “slender” as a subcategory of

119 What makes this the fundamental are two facts: that reason is man’s basic means
of survival and that only coercion — i.e., physical force — can prevent a man from
exercising his reason in the service of his life. Thus, the alternative of freedom vs. force
reflects the alternative of life vs. death. See Rand, “Man’s Rights” and “The Nature of
Government” [VOS, ch. 12 and ch. 14] and Peikoff [OPAR, 310–324 and 350–369].
 L ogic and P ropositions 239

“thin.” Although “slender” marks a shade of difference from other cases

of “thin” (such as “skinny”), one could not say that one would be bereft of
any important distinction if one lacked the concept “slender” and instead
used descriptive phrases, such as “attractively thin.” In contrast, consider the
cognitive overload that would result if one eliminated the concept “justice”
and tried to rely instead on descriptive phrases, such as “evaluating people
and their actions by taking into consideration all the evidence and only
the evidence.”
The optionality of certain concepts is reflected in the fact that some
languages make distinctions that others do not. English, for instance, has
the two terms “forest” and “woods,” but German describes both as “Wald.”
(Of course, in German one can use descriptive phrases to make the distinc-
tion, which pertains to the size of the wooded area; the issue is whether
a single concept is used.)
Optional cases pose a “problem” only for the Realist theory of concepts,
according to which concepts are based on a “universal” that must be either
present or absent in the concretes. “Where do we draw the line between
green and yellow?” the Nominalists ask in attacking Realism — as if some
precise line were necessary. On the Objectivist theory of concepts, which
accommodates continuous measurement-variation, no such sharp division is
necessary. Borderline cases are handled perfectly well by a descriptive phrase
(e.g., “greenish yellow”); no special concept is required.
The very purpose of forming concepts is to use them to identify facts.
This is the function of judgments, and judgments are made in the form of
propositions, to which I now turn.

Logic and Propositions

As defined in Chapter 5, a “proposition” is “a grammatically structured
combination of concepts to identify a subject by a process of measurement-
inclusion.” It is possible for propositions to err, to assert that a thing has
characteristics that it actually does not have. In other words, with the
making of propositional judgments comes the issue of truth or falsehood.
(Concepts, being tools, are not properly described as true or false, but as
valid or invalid.)
 H OW W E K N OW • 7 : L ogic : P ractice 240

The factor making invalid concepts possible is choice — specifically,

the choice of what to contrast with what. The possibility of false propositions
arises because of choice and because of an additional factor: time.120

Propositions as trans-temporal
Perception is time-bound: it is awareness of the present state of things.
Concepts are open-ended in regard to time. The very purpose of forming
concepts is to have a means of carrying knowledge forward, of applying to
future units what one has learned from studying past units; propositions are
the means of doing that.
To say of a given thing, “This S is P” is not merely to say, “This S appears
to be P from here right now” but to assert what S is in reality — that is,
across all conditions of perception — past, present, and future. For instance,
to (mistakenly) describe the distant hills as “blue,” in effect, predicts that from
closer up they will still look similar in color to blueberries and the sky. But that
implicit prediction turns out to be contradicted by later perception, and thus
is false. Similarly, in the case of a stick half-submerged in water, to call it “bent,”
predicts, in effect, that when it is pulled out of the water, its perceived shape
will still be like that of other bent sticks — in contradiction to what one actu-
ally will perceive.
Propositions, of course, do not deny the fact that things can change.
“That is an ice cube” is not shown to have been false if the ice cube later
melts. But a stick does not change its shape by being lifted out of the water,
so “The stick is bent” was false all along.
The use of propositions to make judgments is fallible because to judge
is to apply a predicate, which is a concept, and concepts are open-ended with
regard to time. Concepts are trans-temporal: they integrate, summarize, and
imply how a kind of thing is, was, and will be — across all conditions.
Because concepts are trans-temporal, in using them to make a conceptual
identification one needs to integrate what he perceives at one time with what
he perceives at any other time. A judgment that contains or implies a contra-
diction is false. In fact, the false is precisely the contradictory.

120 Concepts themselves involve time, since they are open-ended in time, applying to all
existents of a certain type — past, present, or future. But the alternative of a concept’s
validity or invalidity is only indirectly related to this fact. The invalidity of the concept
“extremism,” for example, is due to its being based on a superficial characteristic at the
expense of fundamental differences, rather than that “extremism” will run up against
a contradiction in future observation.
 L ogic and P ropositions 241

There are no contradictions in reality. One never perceives a contradiction,

an A that isn’t A. When we talk about contradictions in one’s thinking, what
we actually mean is inconsistency: at a given time, one thinks, “This is A,”
but at another time, in regard to the same, unchanged thing, one thinks,
“This isn’t A.” The inconsistency means that one of the two thoughts is in error.
On the perceptual level, there is no such phenomenon as “inconsistency.”
Inconsistency, and thus error, is a phenomenon of the conceptual level:
the open-ended, trans-temporal nature of concepts is what makes incon-
sistency possible. To subsume something under a concept implies that it is
relevantly the same as the other units of that concept; if it is, in fact,
relevantly different from them, the subsumption is in error.
(In Chapter 2, I rebutted the Argument from Illusion, which attacks
perception by appeal to cases like that of the stick half-submerged in water.
I said that the error lies in the judgment made, not in the perception itself.
The trans-temporal nature of conceptual identification provides the deeper
explanation of this point: the conceptual identification uses open-ended con-
cepts and thus extends into the future. That is why a conceptual identification
can contradict what is later perceived.)
Thus, consistency — non-contradictoriness — is the standard of truth
for propositions. To paraphrase Rand’s statement quoted in Chapter 6:
No proposition man forms is true if it cannot be integrated without contra-
diction into the total sum of his knowledge.121

Defective propositions
The truth or falsehood of a proposition is distinguishable from the validity
or invalidity of its formation. Truth and validity are highly interrelated:
the truth of a proposition depends upon its having been validly formed,
and that validity usually depends upon the truth of prior propositions.
Truth and validity are given separate discussion in what follows, but one
must bear in mind that they are interconnected.
A precondition of considering the truth or falsehood of a proposi-
tion is that the proposition be meaningful; that, in turn, depends upon
how the proposition was formed. A proposition is formed by combin-
ing concepts; a validly formed proposition consists of valid concepts,
validly combined. If the parts are invalid or if they are combined illogically,
the resulting proposition is defective. If the defect is severe enough, what

121 Rand’s original statement was: “No concept man forms is valid unless he integrates
it without contradiction into the total sum of his knowledge.” [AS, 1016]
 H OW W E K N OW • 7 : L ogic : P ractice 242

one is dealing with is not a proposition at all but a pseudo-proposition,

a string of words that fails to make an intelligible statement.
To have a definite meaning, a proposition must ascribe a properly con-
ceptualized predicate to a clearly designated and properly conceptualized
subject.122 Further, to be cognitively meaningful, the proposition must be
based not on fantasy but on some evidence. Sentences uttered on the prem-
ise “I just feel it” or “Why not believe it?” or “It came to me in a dream,” are
pseudo-propositions, not cognitive assertions. Such utterances are outside
the domain of cognition. They are neither true nor false, but arbitrary.
(The topic of arbitrary assertions will be discussed at length in the
next chapter.)
Let’s look closer at the two cases: propositions that are defective due to
using invalid concepts and propositions defective on account of invalidly
combining legitimate concepts.

1. Propositions using invalid concepts

Every concept used in a proposition must be valid; an invalid concept
renders defective any proposition that uses it. Take the earlier example of
“biangle,” which supposedly stands for two-sided polygons. Since “biangle”
is an invalid concept, the sentence “Biangles can be large or small” is not
a proposition at all. The issue of whether it is true or false does not even arise,
because “biangle” involves a contradiction.
The same is true when a proposition uses a less obviously invalid concept.
The false integration “extremism” renders defective any proposition like:
“George Washington was an extremist.” As a matter of logic, one must reject
the proposition, rather than attempt to decide whether Washington was or
was not an extremist.
There are concepts that would be valid if given a proper definition but
are invalid as defined, implicitly or explicitly, by the person using them.
Propositions using such invalidly defined concepts are defective as used
by that person. There are two cases here: concepts that are misdefined and
concepts that are undefined. “Freedom” can be given a logically valid defini-
tion, but since Bertrand Russell defined it as “the absence of obstacles to the
realization of desires” that invalidated any use he made of the term. If Russell
ever said, “Freedom is a good thing,” that was a defective proposition for him,
because of how he defined “freedom.”

122 If the subject term is a proper name, it must clearly specify a particular thing.
 L ogic and P ropositions 243

The same applies in the case of someone who uses terms while lacking
an adequate understanding of them, rather than operating with an explicit
but illogical definition, as Russell does. Suppose that, for instance, someone
says: “Love is the solution to all problems.” There is no point in trying to
assess the truth or falsehood of this statement, because it clearly reflects the
speaker’s unclarity about what “love” means. “Love” in his mind is a floating
abstraction, or it is a “package-deal” that combines actual love (an emotional
response to values) with something fundamentally different: a temporary
relief from guilt, resulting from being granted an unearned forgiveness.
Having clear, unambiguous, properly defined concepts is a precondition
of making cognitive assertions by combining those concepts. If the compo-
nents are defective, so is the proposition combining them.

2. Invalid combination of concepts

Even when each of the concepts used in a proposition is valid, a defective
proposition will result if these concepts are improperly combined.
To be valid, a proposition must meet three criteria: it must be a) grammatical,
b) consistent, and c) referential.

a) The grammaticality requirement

The first requirement of valid proposition is that it be formulated in a gram-
matical sentence. The following is not a proposition: “Is of in in.” But less
radically ungrammatical propositions are still malformed. For example:
“Tom told his uncle that he was working too hard” — was it Tom or his uncle
who was said to be working too hard? Or, “Nearing home, the familiar drive-
way was seen.” (A dangling participle: the driveway was not nearing home.)
Of course, such minor issues are often merely linguistic sloppiness and
can be fixed by editing (though re-writing frequently requires doing real
work to clarify one’s thought). But grammatical issues extend to the philo-
sophical level as well. Consider the following argument for the existence
of God: “Every event is caused by something, and the something that
causes every event is what we call God.” The grammar here equivocates on
“caused by something.” The conclusion states that the same one thing (God)
causes each different event that occurs. But that is not the meaning of the
premise (or else the premise begs the question). The conclusion that follows
 H OW W E K N OW • 7 : L ogic : P ractice 244

from the assumed meaning of the premise would be: each event has its own
particular cause — a statement carrying no theistic implications.123
A more interesting class of cases concerns what I call “philosophical
grammar.” While ordinary grammar concerns the right use of the parts of
speech, philosophical grammar concerns the right use of the metaphy­sical
categories — entity, attribute, action, relationship, etc. To take a pedantic
example, there is nothing linguistically ungrammatical in the statement,
“Saturday is in bed.” But “Saturday” is not an entity and so cannot be in or
out of bed.
Amateur writing often exhibits philosophically ungrammatical
construc­tions. Consider this statement from a Wikipedia entry on the
philosophy of education:

Instead of being taught in philosophy departments, philosophy

of education is usually housed in departments or colleges of
educa­tion, similar to how philosophy of law is generally taught
in law schools.

The writing in this passage is philosophically ungrammatical: it switches

confusingly between metaphysical categories. Where (in what department)
philosophy of education is taught cannot be compared to “how” (in what
manner) philosophy of law is taught. It is beside the point that one can still
reconstruct the idea the author had in mind: “how” is a word from the wrong
category of existents, so using it requires the reader to do the mental work
of figuring out the correct comparison, which takes mental resources away
from grasping and integrating the point the author wishes to make.
The writing in this example can be easily salvaged by doing a little editing;
the important cases are those in which it is the thought behind the words
that is invalid, especially when violations of philosophical grammar occur
in philosophic writing. It would be ungenerous to cite vaguely poetic state-
ments, like “God is love,” so take the following passage penned by the early
20th century philosopher Henri Bergson.

. . . philosophers agree in making a deep distinction between two

ways of knowing a thing. The first implies going all around it,
the second entering into it. [Bergson 1946]

123 The equivocation is the same as the one in the old joke: “Every man loves some woman
— who is she?”
 L ogic and P ropositions 245

Knowing a thing is a mental activity; it is not a matter of movement in

space. Nor can all the objects of knowledge be treated as if they were physical
entities to be moved around or entered into (whatever that would mean).
Take the knowledge that 2 + 2 = 4. How does one move around it or enter
into it? Perhaps the words are to be taken metaphorically. But if so, it is
wholly improper to introduce an allegedly basic philosophical distinction
by means of a metaphor — especially not such an opaque one. Nor does he
follow up with any literal statement of what this “deep distinction” actually
consists of. (Lest you think I’m quoting Bergson out of context, this passage
comes from the very first sentence of the chapter titled “An Introduction
to Metaphysics.”)
A more famous sin against philosophical grammar is displayed in Martin
Heidegger’s pronouncement, “The Nothing nothings.” (“Das Nichts nichtet.”)
[Heidegger, 1929] Likewise, his question: “Why is there something rather than
nothing?” misuses the term “why,” ignoring the fact that it is a request for
the identification of a cause. The expanded meaning of his pseudo-question
would be: “What is the something that causes there to be something rather
than nothing?” — which is unintelligible. The same statement also reifies
“nothing,” treating it as a thing rather than as the absence of a delimited
positive (as in “There is nothing in my pocket” — i.e., none of the palpable
physical objects that might have been there.)
Unintelligible language is not the only result of philosophically ungram-
matical writing; such misuse of categories can also produce monumental and
tragic errors. A notorious example is the following passage by John Stuart
Mill, in his alleged proof that morality requires the sacrifice of the individual
to the collective (under his ethics of utilitarianism):

The only proof . . . that an object is visible is that people actually

see it. The only proof that a sound is audible is that people hear
it; and so of the other sources of our experience. In like manner,
I apprehend, the sole evidence it is possible to produce that any-
thing is desirable is that people do actually desire it. [Mill 1957, 44]

Mill here uses “desirable” in a philosophically wrong way. “Desirable”

does not mean “capable of being desired”; if it did, it would include immoral
desires, such as the desire to murder. Rather than denoting a capacity,
“desirable” pertains to an entirely different category — the category of norma-
tive concepts. “Desirable” denotes what ought to be pursued or obtained, in
marked distinction from whatever people just happen to desire.
 H OW W E K N OW • 7 : L ogic : P ractice 246

Compounding the felony, Mill goes on, two sentences later, to this philo-
sophically ungrammatical horror:

. . . each person’s happiness is a good to that person, and the

general happiness, therefore, [is] a good to the aggregate of all
persons. [Mill 1957, 45]

What is “the aggregate of all persons?” Mill’s statement implies that this
aggregate is some collective entity floating above and beyond the individuals
residing in a given region. The “aggregate of all persons” is treated as if it were
a sentient organism that had its own pursuits, goals, and emotions. It is not
merely the linguistic formulation but the thought behind it that does vio-
lence to philosophical grammar. Yet such pseudo-thought has characterized
most ethical-political theorizing over the last two centuries; the same
philosophical muddling occurs in the near-universal notion that “society”
is an entity capable of providing benefits to its members, and that “the com-
mon good” or “the public interest” is something requiring laws that sacrifice
the interests of individuals (who are thus excluded from “the public”).124
All such collectivist utterances, from Mill’s on down to those of today’s can-
didates for office, are defective propositions, pseudo-propositions actually;
they cannot be treated as expressing something definite that could be true
or false. Sometimes the person putting forward a defective proposition
has an actual thought in mind that he can put into a new (and different)
statement that is logical. But more often such statements either are irreme-
diably confused (as Mill’s are), or, when made intelligible, stand revealed
as patently false.

b) The consistency requirement

The second requirement of valid combination is: consistency. First, the
proposition must be internally consistent, as opposed to such illogical
combinations of concepts as “That circle is square.” But the more interesting,
and oft-flouted, requirement is that a proposition must be consistent with
its own hierarchy: it must not use concepts in a way that contradicts the

124 We know where this collectivist road leads: “We recognize only two Gods: A God in
Heaven and a God on earth and that is our Fatherland,”Adolf Hitler, Volkische Beobachter,
Sept. 23, 1928 [Cohen, 411] and: “[T]he dictatorship of the proletariat is the rule — unre-
stricted by law and based on force — of the proletariat over the bourgeoisie,” Josef Stalin,
The Foundations of Leninism [Cohen, 173].
 L ogic and P ropositions 247

knowledge required to form them. That kind of contradiction is involved in

utterances that commit the fallacy of the stolen concept. I have noted this
fallacy several times earlier in the book; now I can treat it in full detail.
The stolen concept fallacy is a form of hierarchy inversion: it consists of
the attempt to use a derivative concept in a way that contradicts its own
presuppositions — i.e., that negates or ignores a prior concept that is required
in order to grasp and use the concept in question. Suppose someone were to
announce, “I know there are desserts, but there is no such thing as a meal.”
Here the stolen concept is “dessert”; a dessert is a dish that follows a meal.
The concept that is said to be “stolen” is the one that is used without a logi-
cal right to that use, in analogy with using someone else’s property without
a legal right to do so. In this case, “dessert” is the stolen concept.
When a concept is “stolen,” it is being used in a way that severs its connection
to perceptual reality and thus deprives the concept, as used, of meaning.
“Dessert” could not have its present meaning if meals did not exist. The same
type of contradiction is contained in each of the following examples.

• I reject the existence of consciousness.

“Rejection” is an action of consciousness.
• Logic is a Western prejudice.
A “prejudice” is that which is pre-judged, in advance
of logical evidence.
• Life is all a dream.
A “dream” is meaningful only in distinction to wakeful perception.
• Property is theft.
“Theft” is forcibly taking property from its rightful owner.
• The laws of logic are arbitrary.
The “arbitrary” is distinguished from the logical.
• You can’t prove reason is valid.
“Proof ” can be grasped only as a certain process of reason.
• Physics is defined as: what physicists do.
“Physicist” can be grasped only in relation to “physics.”
• The universe is moving.
“Motion” is change of place; “place” is the surrounding entities;
the universe is everything; nothing surrounds it, and it has no place.

In the above, the stolen concepts — the concepts rendered meaningless

— are, in order, “reject,” “prejudice,” “dream,” “theft,” “arbitrary,” “prove,”
“physicist,” and “moving.”
 H OW W E K N OW • 7 : L ogic : P ractice 248

The last three examples involve not denying but ignoring the stolen
concept’s roots: ignoring the dependency of “proof ” on “reason,” of “physi-
cist” on “physics,” and of “motion” on “place” and “entity” (or “existent”).
All of these statements deprive the concept that is stolen of its hierarchical
base, rendering the concept as meaningless as the nonsense sound “slatch.”
Sentences containing a stolen concept express no thought, make no judg-
ment, and are therefore only pseudo-propositions.
Concept-stealing is arguably the most frequent and most destructive
fallacy in the history of philosophy. The mother of all stolen concepts,
one rampant in post-Cartesian philosophy, is the primacy of consciousness:
the attempt to use concepts of consciousness while denying or ignoring
that consciousness is consciousness of something, something that exists
(see Chapter 1).

c) The referentiality requirement

The third requirement of making a valid combination of concepts is that it be
referential. The proposition must succeed in designating a subject. The trivial
case of a failure in this regard is a proposition based on a false presupposition,
such as the proposition Russell devised: “The present king of France is bald.”
There is no present king of France, so the combination of concepts in the
sentence’s subject fails to refer, and the sentence is thus not a proposition and
is neither true nor false. (The non-referential term could be in the predicate,
as in: “That man is the king of France’s nephew.”)
A more interesting way of violating the requirement of referentiality is by
making statements which commit what I call “the fallacy of pure self-reference.”
The simplest example of the fallacy is:

This statement is false.

On inspection, it seems that if this statement is true, then it is false.

If it is false, then it is true.
That statement is representative of a whole family of self-referential
paradoxes. These would remain parlor-tricks were it not for the fact they
have spawned disastrous conclusions in the fields of mathematics and logic.
Morris Kline’s definitive book on the history of mathematics includes
a history of the effect of these paradoxes, and the book is appropriately
titled: Mathematics: The Loss of Certainty. Kline reports that the discovery
of these paradoxes, “and the realization that similar paradoxes might be
 L ogic and P ropositions 249

present, though as yet undetected, in the existing classical mathematics, caused

mathematicians to take seriously the problem of consistency.” [Kline, 216]
These paradoxes gave skepticism and subjectivism a boost. “The pride of
human reason was on the rack.” [Kline, 257]
Take a closer look at “This statement is false.”
The solution to the paradox is to realize that the sentence is neither true
nor false; it makes no judgment; it is not a proposition at all, but a pseudo-
proposition, because it fails to refer to anything. To see this, take the easier
case — the same sentence minus its unnecessary twist:

This statement is true.

What does that assert? That it is true? But that what is true? The sentence
has no content. To be true or false, a statement must first refer to something.
Only then can we evaluate its content as either corresponding to or con-
tradicting the facts. In both “This statement is true” and “This statement is
false,” there is no statement, and therefore nothing to be either true or false.
Neither sentence says anything. A proposition declares what its subject is
or does. But the subject of these sentences is “This statement” — meaning
its content, not its words as shapes on the page. Since there is no content,
no statement is made. Thus, both sentences are pseudo-propositions.
Such sentences attempt the impossible: to refer only to their own act of
referring. Their own referring — to what? To their own referring to their own
referring to . . . ? We are caught in an infinite regress. There is nothing capable
of being either true or false until after the sentence has content — a stage we
can never reach in cases of pure self-reference.
The fallacy of pure self-reference is an instance of the fallacy contained
in any version of the primacy of consciousness: “A consciousness conscious
nothing but itself [an act of pure self-reference] is a contradiction in terms:
terms: before it could identify itself as consciousness, it had to be conscious
of something.” [AS, 1015]
“Pure self-consciousness” is a contradiction in terms. One can certainly
be self-conscious, but self-consciousness can occur only after and on the
basis of first being conscious — i.e., being aware of something that exists.
Extrospection must precede introspection. Before extrospection, there
is no consciousness to introspect. First there is awareness of something,
say of a tomato; only then can one be aware of being aware of the tomato.
As I put it in the first chapter: consciousness precedes self-consciousness.
 H OW W E K N OW • 7 : L ogic : P ractice 250

This necessary order applies to statements as well. A statement can refer

to and include itself, but only after and on the basis of having first referred
to something. For instance, “All propositions refer to something” is a valid
and true proposition, and it includes itself: it has a subject to which it refers
(namely, “propositions”). But it includes itself only because and on the basis
of first referring to propositions other than itself — such propositions as
“Dogs are animals.”
Consider a contrasting case: “This sentence is grammatically correct.” That
sentence is grammatically correct — and true. It contains no pure self-reference:
the sentence refers to its own grammatical structure, but that is a linguistic
matter. Its grammatical structure is independent of its conceptual content
— i.e., of the particular state of affairs it refers to. It remains grammatically
correct if we reverse its assertion: “This sentence is not grammatically correct.”
That is also a grammatically correct sentence; it asserts something, but what
it asserts contradicts the facts, so it is false.
The principle that reference must precede self-reference also resolves
more convoluted cases, such as that of the Russell paradox. That paradox
is based on “the set of all sets that are not members of themselves.” Is that set
a member of itself or not? The answer is: no, it is not self-including. But that
does not imply that the set fits its own definition and thus is self-including.
In fact, the set’s definition cannot be applied to itself. A simplified version of
the paradox will make this point easier to grasp.
Let’s call a man “well-named” if his name correctly describes him. Thus,
Mr. Black is “well-named” if he is black, but “ill-named” if he is white. Like-
wise for Mr. Rich and his financial status. But what about Mr. Ill-Named?
The answer is that Mr. Ill-Named is neither well-named nor ill-named.
He is in the same boat with Mr. Jones. “Jones” does not ascribe any feature
to its bearer, and neither does “Ill-Named.” Pure self-reference is impossible.
A name cannot refer only to the inaptness of its own reference — to what?
Mr. Ill-Named is not well-named. Period. Mr. Well-Named is not well-named.
Period. And the set of all non-self-including sets is not self-including. Period.
One cannot go on to assert that this implies it is self-including: self-inclusion
presupposes that the predicate has meaningful content, which is not the case
when self-application involves an infinite regress.125

125 The hardest case is perhaps: “This sentence is meaningless.” But the principle is the same:
that sentence is meaningless, and one cannot object: “But that’s just what it says it is!”
It does not “say” anything. All paradoxes of pure ­self-reference dissolve when one realizes
that they either attempt to refer to only their own referring, and thus fail to refer,
or attempt to evaluate only their own evaluation (of what?) and thus fail to say anything.
 L ogic and P ropositions 251

Finally, consider Gödel’s theorem in mathematics; it is believed to show

that there is a certain “incompleteness” in logical-mathematical systems.
Carl Boyer says of Gödel’s theorem, “it appears to foredoom hope of math-
ematical certitude.” [Boyer, 656] To the extent that, beneath the mathemati-
cal formalism, this theorem trades upon “This statement is unprovable,”126
the same fallacy is involved: there is no proposition there to be proved and
thus no “incompleteness” in logic or mathematics for being unable to prove
a non-statement.
The emptiness of “This statement is unprovable” shows nothing about
proof or completeness; the same circularity is exhibited by “This statement is
magnificent.” No statement can refer to or evaluate only itself.
The Russellian paradoxes arise because philosophers have attempted
to treat truth as if it were a matter of the correspondence between words
and facts, and to turn proof into a mechanical procedure in a purely formal
system. The validity or defectiveness of a proposition is reduced to an issue of
typography — of accepted symbols and physically defined criteria for “well-
formed formulas” — in the belief that one can thereby avoid such “mentalistic”
ideas as “concept,” “meaning,” and conscious reference to reality.
In fact, these paradoxes constitute a refutation of the formalist approach
to logic. Purely physical criteria provide no means of distinguishing vacuous,
self-referential sentences from those with cognitive content. To distinguish
propositions from pseudo-propositions, one must recognize the fact that
concepts are words given a cognitive content by a mind seeking awareness
of the facts of reality.
A validly formed proposition has a clear, unambiguous meaning. It asserts
that its subject possesses the characteristic(s) conceptualized by its predicate.
This presupposes that the subject and predicate terms are valid concepts,
validly defined, and that they are organized in a way that makes the whole
they constitute both intelligible and graspable.

Truth
Having outlined the requirements for the logical validity of a proposition,
we are now in a position to discuss the nature of truth, and what makes
a proposition true.

126 Gödel himself denied he was trading upon this circularly self-referential statement.
[Gödel, 41 n13] This remains a disputed issue.
 H OW W E K N OW • 7 : L ogic : P ractice 252

Truth, as Rand characterizes it, is “the recognition of reality.” [AS, 1017]

Although a true proposition is often described as one that “corresponds to”
the facts, truth actually pertains not to some match-up but to an awareness,
a mental grasp, of the facts.
Awareness is not a series of isolated responses to isolated stimuli;
awareness is a global activity of differentiating and integrating. Products
of that activity, such as concepts and propositions, are organic outgrowths
of the activity, not self-standing items; concepts and propositions are
better analogized to the branches of a tree than to automobiles coming out
of a factory. Propositions grow out of a context, and their meaning depends
upon that context. No proposition exists or has meaning out of context.
The statement’s background context is infused into that statement and shapes
its meaning.
Part of that context is the hierarchically prior items grounding the
proposition. A proposition is formed out of concepts, and the vast major-
ity of propositions depend hierarchically upon prior conclusions. If these
hierarchically prior items were different, the proposition, though expressed
in the same words, would have a different meaning — i.e., it would be a
different proposition.
As an illustration, take the proposition, “Lying is wrong.”127 Its meaning
depends on the content of the ethics and metaphysics that informs that
statement. For a religionist, “wrong” means “against God’s commandments,”
for Rand, “wrong” means “destructive of man’s life on this earth.” And for
the religionist, the basis of the moral judgment is the unaccountable will
of a supernatural being; for Rand, the basis is the natural causal order.
[PWNI, ch. 10] A Humean, Kantian, or Existentialist philosophy would fuse
still other ideas into “Lying is wrong,” so that the same words would actually
express very different propositions.
To judge a proposition as true or false, one must know its meaning, and
its meaning exists as part of the total cognitive context of which the assertion
is an outgrowth.128 In considering a proposition’s truth one cannot detach the
proposition from the fabric of one’s knowledge. Truth is the relationship that

127 I am indebted to Peter Schwartz for making this point, using a similar example,
in his lecture “Contextual Knowledge” (https://estore.aynrand.org).
128 This does not mean that truth itself is “contextual” — i.e., the advance of knowledge does
not change a true proposition into a false one or vice-versa. Sometimes future knowledge
shows that what one believed to be true had never been true, but if a proposition is true,
it is timelessly true. (Even a truth about what was the case at a certain time, such as that
the Civil War ended in 1865, will remain true forever.)
 L ogic and P ropositions 253

a part of a cognitive whole bears to the facts of reality, when that proposition
expresses in conceptual terms a recognition of those facts.
A false proposition is one that contradicts something. The contradic-
tion may be internal, as in “This circle is square,” or the contradiction may
be to other knowledge, as in “Pears grow on vines.” A proposition is false
if it contradicts any fact.
An infrequently noticed form of contradiction occurs in philosoph­ical
statements that commit what Rand calls “the fallacy of self-exclusion.” This
­fallacy is committed when the act of asserting a proposition contradicts its
own content (thus the speaker is illicitly excluding his own utterance from
what he is claiming). For example, “There are no absolutes” — asserted as an
absolute. (Even if one says, “Probably, there are no absolutes,” that is being
asserted as an absolute.) Or, “Man can know nothing for certain” is asserted
as something known for certain.
The fallacy is also committed in fields outside of philosophy. Freudian
psychologists like to assert: “All thinking is just the rationalization of
unconscious impulses” — but then is that statement itself just a rationalization
of unconscious impulses? Marxist economists claim: “All thinking is deter­
mined not by objective considerations of logic but by the material factors
of production” — but then is that statement itself determined by the material
factors of production, not logic? Neuroscientists exclude their own state-
ments if they claim that the genetic structure of the brain dictates all human
activities — what, then, about the activity of pursuing truth in neuroscience?
Such statements are guilty of self-exclusion no matter who makes
them or how, and their content is thus shown to be false. These are the
most important instances of the fallacy, but in other cases whether or not
the fallacy is committed depends on who makes the statement and how.
“I do not exist” exhibits the fallacy of self-exclusion no matter who says it,
but “Harry Binswanger does not exist” is a self-exclusion only if I say it. “There
is no such thing as the English language,” commits the fallacy because it is
itself stated in English; but “There is no such thing as the French language,”
though obviously false, does not commit this particular fallacy.
The manner of making an assertion can contradict the assertion’s content.
If someone says in an angry, denunciatory tone: “That’s a value-judgment!”
he is making a value-judgment himself. (In contrast, the mere observation
“That’s a value-judgment,” without an implied moral condemnation, commits
no fallacy; it might even be said as a compliment.) A statement is false when
it contradicts any fact, including facts about its own utterance.
 H OW W E K N OW • 7 : L ogic : P ractice 254

The next question is: when do we achieve truth? Is being logical in every
respect enough? Normally, it is. Normally, if a proposition is fully logical
— valid in form and reached by an unbroken chain of logical processing
grounded in perception and integrating with everything one knows — it will
be true. But logical validity alone does not guarantee a proposition’s truth.
There are unusual cases in which a proposition that one is fully justified in
taking to be true, based on all the available evidence, nevertheless turns
out to contradict facts not available to one at the time, and thus is false.
(Such cases will be discussed in the next chapter.) Thus, “logical” and “true”
are not fully coextensive. Moreover, there is a distinction in perspective
between “logical” and “true.” The focus of “logical” is on the process, of “true”
on the product — specifically, on the product’s relationship to reality.
A true proposition is one that is both logically valid and expresses an
awareness of fact. To achieve that awareness, a proposition must be based
on prior observations, conceptualizations, and, in many cases, inferences.
I have already discussed perception (Chapter 2), concept-formation
(Chapters 3 and 4), and the general nature of propositions (Chapter 5).
The remaining level to be discussed is that of inference.

Logic and Inference

Inference is the mental process of deriving a proposition from observation
and/or other propositions. Inferences are either deductive or inductive.
Deduction is the simpler process, so I will take that up first.

Deduction
Deduction is the application of the general to the particular (or to the less
general). The most elementary act of deduction is the syllogism, which is
the “atom” of deduction. In the standard “Socrates” syllogism, the premise
“All men are mortal” is the generalization that is applied to the particular fact,
“Socrates is a man,” to produce the conclusion that he is mortal. The syllogism,
with its two premises and three terms, is the smallest unit of deduction; all
the more complex and extended processes of deductive reasoning can be
reduced to chains of syllogisms.
Almost all of everyday reasoning is deductive. Deduction is the pro-
cess used when one goes from “It is raining” to “I should take an umbrella”
 L ogic and I nference 255

(using the suppressed premise: “When it rains, I should take an umbrella”).

Deduction is the form of inference that a jury uses to reach its verdict. (One
such deduction might be: “The murderer had type B blood” combined with:
“Of the people then in the vicinity, only the defendant had type B blood.”)
Even the act of making a simple statement, like “That is a dog,” is deductive,
at least in pattern, because it applies a general term (dog) to its subject.129
Because a deduction represents the application of general knowledge that
one already possesses, the conclusion was implicitly contained beforehand
in the generalization used. But the deduction is hardly in vain: the purpose
of the deduction is to make that implicit connection explicit — i.e., to bring it
into conscious awareness. Deduction draws to mind the implications of what
one knows. Deduction makes the implicit explicit.
Deduction has been well understood ever since Aristotle identified the
syllogism and analyzed which forms are valid and which invalid.
There are 256 theoretically possible variants of the syllogism, according to
different combinations of the factors known as Quantity, Quality, and Figure.
Only 17 of the 256 are formally valid (i.e., have a structure that actually
succeeds in applying the general to the particular). But of the 17, only two
or three valid forms are used in real-life thought. As to the invalid forms, in
my experience only two are apt to occur in ordinary reasoning: equivocation
and undistributed middle. Consequently, little needs to be said to improve
people’s use of syllogisms, and that little can be quickly gleaned by consulting
any traditional textbook.130 (There is much merit, however in familiarizing
oneself with the so-called “informal f­ allacies,” such as Ad Hominem, Appeal
to Authority, Begging the Question, and a dozen or so others; these are
“informal” not in being casual but in pertaining to content not form, and to
the validity or relevance of the premises used. It was Aristotle who identified
almost all of the informal fallacies.)

Induction
The other type of inference is induction: the process of generalizing from
particulars (or from the less general). Where deduction applies the more
general to the less general, induction moves from the less general to the
more general.

129 See Chapter 5 and ITOE, 28.

130 I recommend older textbooks, such as Logic: An Introduction, by Lionel Ruby (re-issued
by Paper Tiger).
 H OW W E K N OW • 7 : L ogic : P ractice 256

Induction has been under attack for centuries, and is now regarded as
something uncertain or subjective. But the validity of induction cannot
in fact be denied or even questioned: induction is the fundamental means
of acquiring conceptual knowledge. Without induction, there would be no
general premise for a deduction to apply. Deduction presupposes induction.
Since all inferences either are inductions or require inductively reached
premises, the attack on induction is an attack on all inference — which is
self-refuting, since the attacks themselves require making and applying gen-
eralizations reached by induction. In other words, the statement “Induction
is invalid” commits the fallacy of self-exclusion. The same is true of milder
attacks, such as “Induction cannot give certainty” which is itself a claim to
certainty about a generalization.
Though there is no “problem of induction,” there are legitimate questions
regarding how induction works and what is required for an inductive gen-
eralization to be valid. (In the same way, there never was any “problem of
deduction,” but Aristotle’s discovery of the principles of deduction marked
an historic advance.)
A revolutionary understanding of induction has recently been provided
by Leonard Peikoff. Rejecting the conventional statistical approach, Peikoff ’s
treatment of induction focuses on the crucial role of concepts:

One must grasp how the constituent concepts of a generalization

are related to reality before one can grasp how the generalization
itself is related to reality. The theory developed here is based on
Rand’s theory of concepts. . . . [V]alid concepts, in her definition
of “concepts,” not only make possible but also guide our search for
true generalizations.131

The two essentials of Peikoff ’s theory reflect two points of Objectivist

epistemology emphasized in this book: 1) the hierarchical nature of know-
ledge, and 2) the open-endedness of concepts.
1. Hierarchy applies to generalizations just as it does to concepts.
Accordingly, when seeking to understand induction, we cannot plunge in
at any random stage of the hierarchy. We could not understand concept-
formation if, in ignorance of hierarchy, we began with and focused only on

131 Leonard Peikoff, Introduction to The Logical Leap, by David Harriman. [Harriman 2010,
xi] This book, prepared under Peikoff ’s guidance, presents in written form the theory that
Peikoff gave as lectures; Harriman illustrates the theory by discussing, philosophically,
a series of the key discoveries in the history of physics.
 L ogic and I nference 257

higher-level concepts, such as “inflation,” and attempted to discover how

these concepts could be formed directly from perception. Likewise, we can-
not understand induction by taking higher-level generalizations, such as
“Water boils at 212 degrees Fahrenheit,” as our paradigm case, since this is
a higher-level generalization and is not formed directly from perception.
A sensitivity to the hierarchical nature of induction enables us to recog-
nize that advanced generalizations depend on less advanced ones, so that
the right cases to examine first are those at the hierarchy’s base. The base is
first-level generalizations, generalizations that do not presuppose any prior
generalizations. Peikoff ’s examples include: “Fire burns paper,” “Drinking
water quenches thirst,” “Pushing a ball makes it roll.”
The facts that first-level generalizations formulate are self-evident — i.e.,
available to direct perception. Moreover, the concepts used to formulate
those facts are concepts validated by direct perception, and not subject
to error. E.g., the similarity in the rolling of a baseball, a golf ball, and an egg
is perceptually given, so the concept “roll” needs no justification or reduction.
(The same is true of concepts of perceptually given attributes, relations, etc.
One cannot make an error or violate Rand’s Razor in forming such concepts
as: “blue,” “soft,” or “on.”132)
All these cases of first-level generalizations — “Water quenches thirst,”
“Fire burns paper,” and “Pushing a ball makes it roll” — are identifications
of causal connections. This is true of inductions in general, on any level.133
Some cause-and-effect relationships are complex and can be identified
only on the basis of abstract concepts, but the causation operative in these
simple cases is available on the perceptual level.134 A toddler can see and feel

132 The Logical Leap refers to concepts like “roll” and “thirst” as “first-level,” which has
generated some confusion. One must distinguish between two somewhat different ways
of assigning a concept’s “level”: the level pertaining to learning-order, and the level
pertaining to justification. (On this issue, see Chapter 4, p. 153, footnote 74.) A child
directly perceives the difference between the motions of things that roll and the motions
of things that slide or tumble, and thus he cannot get “roll” wrong (as he can get wrong
“vector” or “freedom”).
It is self-evidency (not place in the order of learning) that is required of the concepts
used in one’s first inductions. A generalization, to be first-level, must contain only con-
cepts that are perceptually self-evident, not subject to error, and need no justification.
That is indeed the case for all the concepts in “Pushing a ball makes it roll” and the other
examples of first-level generalizations.
133 A possible exception is the case of generalizations about part-whole relationships,
e.g., “All matter is composed of atoms,” and, “All living organisms are composed of cells.”
But even these could be construed as identifying what Aristotle called “the ­material cause.”
134 Peikoff credits this point to Gregory Salmieri.
 H OW W E K N OW • 7 : L ogic : P ractice 258

the effects of giving the ball a push; he can see the fire burning the paper, he
can feel the water quenching his thirst. The causality here is nothing like that
involved in such abstract generalizations as: “An increase in supply causes
a fall in price” or, “Stars generate their light and their heat by nuclear fusion.”
These two statements do identify causal connections, but not ones that are
perceptually given, as is the causal connection identified in the statement
“Pushing a ball makes it roll.”
To make a valid inductive generalization, one must know the operative
causality (either perceive it directly or reach it conceptually). To general­ize
from “This particular S is P” to “All S is P,” one must know that this S is P
because it is S, not because of some other fact about it. One needs to know
that to be S is to be P.
For example, Archimedes discovered that an object immersed in water is
buoyed up by the weight of the water it displaces. This is a fact about physical
objects as such, not about, say, the material of which they are made. To be an
object is to have weight — thus, to displace water (which is incompressible).
It is not qua stone or wood or metal that a thing is buoyant. An immersed
object qua object displaces its weight in water.
“S qua S is P” implies “To be an S is to be P.” That fact is the basis and the
meaning of the generalization “All S is P.”
Concepts of characteristics, like “weight,” enable one to isolate and identify
causal factors and thus to know qua what a given entity acted as it did.
That knowledge is necessary for one to reach a higher-level generalization.135
2. The second part of Peikoff ’s theory is its explanation of the mechanism
of generalizing — e.g., of the mind’s movement from “This (paper) did that
(burned)” to “Fire burns paper.” The factor permitting such generalizing
is the open-ended nature of the concepts the child applies. In conceptually
identi­fying the fire as fire, the burning as burning, and the paper as paper,
the generalization follows perforce. The open-endedness of concepts is what
permits the application of concepts to new concretes, and the potential for
making that application is the essence of what it is to generalize.
We must bear in mind that first-level generalizations, like all propositions,
are contextual. A child observing a newspaper burning in a fireplace can
form the open-ended proposition “Fire burns paper” without having the
responsibility of testing the scope and conditions of this generalization:

135 This analysis in terms of S qua S, and the formulation “To be an S is to be P” are not taken
from Peikoff or Harriman, and the responsibility for any errors in this way of conceiving
higher-level generalization is my own.
 L ogic in a N utshell 259

the context of his knowledge does not permit doing so, and no valid standard
requires the impossible. Here, as in ethics, “ought” implies “can.”
Consider the case at hand: going from seeing a newspaper burn up in
a fireplace to “Fire burns paper.” Our beginning inducer is perhaps two or
three years old. At his rudimentary stage of knowledge, the child, as Peikoff
points out, could not and does not need to ask himself such questions
as, “Is it only the first edition of The New York Times that acts this way?”
“Is it only in wood fires that a sufficient temperature is reached?” He knows
of no such issues. When he learns them, years later, he may add any needed
contextual provisos to his induction. The key point (to be pursued in the next
chapter) is that his later knowledge does not overturn his earlier knowledge.
On the contrary, the later discoveries depend on and expand his original
discovery that fire burns paper.
Take an extreme case: the fact that paper can be treated with flame-
retardant, in which case it will not burn. This fact, which is way beyond the
young child’s ken, in no way contradicts or invalidates his generalization;
in fact, the invention of flame-retardants was made possible by the prior
knowledge that fire burns paper.
The point is that the later statement, “Fire burns paper, unless counter-
acted by ­flame-retardant,” constitutes more knowledge than “Fire burns
paper.” Learning about flame-retardants moves the child from having one
item of knowledge on the topic to having two, rather than from one item of
(apparent) knowledge back to zero.
There is more work that needs to be done regarding induction, especially
with regard to the precise standards for inducing and validating higher-level
generalizations in science. [See Harriman, 2010] But from what has already
been sketched in here, it should be clear that Peikoff ’s theory answers the
basic questions, and that the key to understanding induction is the application
of logic — i.e., of context and hierarchy — to the topic, on the base provided
by Ayn Rand’s theory of concepts. (Of special importance is Rand’s point:
“No concept man forms is valid unless he integrates it without contradiction
into the total sum of his knowledge.” [AS, 1016])

Logic in a Nutshell
Logic is the reality-based method of conceptual functioning. To be ­logical,
a process must be grounded in and derive its content from observation
of reality. Logic is based on the basic fact of reality: the Law of Identity,
 H OW W E K N OW • 7 : L ogic : P ractice 260

with its two corollary Laws: Non-Contradiction and Excluded Middle.

Logic consists of integrating and differentiating on the basis of the identity
that things are observed to have — i.e., according to their observed character-
istics and the measurement-relationships obtaining among them.
Logic also requires obeying the identity of man’s means of cognition.
This means recognizing that knowledge is gained in a context and is built
up in a hierarchical progression from perceptually given data. And it means
accepting the principle of unit-economy, in recognition of the limit on the
number of distinct units a consciousness can hold in focal awareness and
deal with at one time.
To summarize: logic is observation-based, non-contradictory, and unit-
economical.
 PROOF AND
8
CERTAINTY

T he preceding chapters have focused on the acquisition

of knowledge. Now I turn to the issue of validation — the process of
ensuring that an idea is correct, does constitute knowledge. Some ideas
can be validated by direct perception, but other ideas require a multi-step
process of validation: proof.
The need for proof arises not only in regard to ideas one hears asserted
by others but also in regard to any of one’s own past conclusions whose
logical basis one no longer clearly recalls. The question then is: is the idea true
or false? Is it derivable by logical means or not? What is its epistemic status:
is it certain, likely, possible — or is it invalid?
The epistemic status of an idea depends upon the evidence supporting
it and on the logical validity of the reasoning premised on that evidence.
If someone is asserting an idea, and its basis in reality is not obvious,
the first thing one should do is to ask him directly: “Why do you say that;
what’s your evidence?” — a question asked too rarely. In validating a belief
of one’s own, a logical first step is to ask oneself: “How did I arrive at this
 H OW W E K N OW • 8 : P roof and C ertaint y 262

belief, by what steps?” — not as autobiography, but as a lead to identifying

the steps by which one could validly reach this conclusion. Thus, one attempts
to “reverse engineer” the idea, seeking to determine if it can be reached
logically from prior knowledge.
To prove an idea, one needs to link it back to perceived fact. The Objectivist
term for this process of going back “down” the hierarchy to prove an idea
is: reduction.

The responsibility imposed by the fact that knowledge is hier­

archical is: the need of reduction. . . . Reduction is the means of
connecting an advanced knowledge to reality by traveling back-
ward through the hierarchical structure involved, i.e., in the
re­verse order of that required to reach the knowledge. “Reduc-
tion” is the process of identifying in logical sequence the interme-
diate steps that relate a cognitive item to perceptual data. Since
there are options in the details of a learning process, one need not
always retrace the steps one initially happened to take. What one
must retrace is the essential logical structure. [OPAR, 132–133]

Heraclitus said that the way up and the way down are one and the
same — referring to a staircase. The same can be said of the staircase of
knowledge. Contrary to contemporary notions, there is only one logic,
not both a “logic of discovery” and a “logic of proof.” Instead, there are two
different directions of motion along the same logical, hierarchical structure:
derivation moves “up” from the perceptually given, while proof moves back
“down” to the perceptually given. (There may be intermediate upward steps.)
A derivation moves in thought from what is closer to perception to
what is farther from perception. Reduction is the same process in reverse,
­moving back down the same hierarchical structure, with the hierarchy termi­
nating in the self-evident data of perception. Derivation is from perception;
proof is back to perception.
The two directions — derivation and proof — apply to both deduction and
induction. In deduction, we take knowledge gained from past observation
and apply it to a new instance. For example:

Lassie is a dog.
Dogs are animals.

Lassie is an animal.
 P R O O F A N D C E RTA I N T Y 263

The conclusion is more abstract — farther from perception — than the first
premise. One gains a wider knowledge of Lassie by means of applying to
Lassie the prior generalization that dogs are animals. (For this to be new
knowledge about Lassie, the prior generalization would have to have been
formed from observation of dogs other than Lassie.)
Now consider a somewhat different case of deductive derivation:

There is smoke coming from that hill.

Where there’s smoke there is fire.

There’s fire on that hill.

Here, the conclusion is no more abstract in itself than the first premise,
but it is a conclusion not perceptually available now, from here; it is further
removed from currently available perception than is the starting premise,
“There is smoke coming from that hill.”
In any deductive derivation, we use a general premise (“Where there’s
smoke there’s fire”) as a bridge to make a new application. We reach thereby
not new general knowledge, but new knowledge about the concrete (viz., that
particular hill).
In contrast to derivation, proof goes in the other direction, to check the
validity of the conclusion (or to prove a hypothesis). The thought process
would be something like the following.
“Is Lassie an animal? What is an animal? It’s a living being with conscious-
ness and locomotion, like pigs, snakes, and dogs, as opposed to plants.
Since Lassie is a dog, and, like all dogs, she is conscious and moves herself
around, Lassie is an animal. Q.E.D.”
(One is not by this process checking the validity of the syllogism
used in the deduction; in simple cases, that does not need to be checked;
instead, the process is one of reducing “animal” back to perception,
to see whether there can be an application made. Of course, the example
is artificial, chosen to illustrate the pattern of a reduction rather than
for realism.)
For a case of inductive derivation, consider the following schematized
(not complete) example. It starts from Peikoff ’s example “Fire burns paper,”
and goes one level more abstract:

Fire burns paper.

Focusing sunlight on paper with a magnifying glass burns it.
 H OW W E K N OW • 8 : P roof and C ertaint y 264

Putting paper in a very hot oven burns it.

The common causal factor in these cases is high heat.
Therefore, high heat burns paper.

(Let’s put aside any doubts as to whether this derivation is iron-clad:

my point here is the hierarchical progression — the logical structure — not
the certainty of the result.)
The corresponding reductive proof goes in the other direction:
“Does high heat burn paper? Yes, because fire burns paper, highly focused
sunlight burns paper, and a very hot oven burns it, but applying less heat in
each case does not, so the common factor is high heat.”
This artificial example is too simple to be a process one would ­actually
need to step through, but it shows the pattern that would be involved
in more complex reductions. Take, for instance, the reduction of the Law
of Demand in economics: “The quantity of a good that will be demanded
varies inversely with its price.” This can be reduced by first getting clear
on what is meant by “good,” “demand,” and “price.” (Roughly: a “good”
is an item with exchange value — i.e., which is bought and sold; “demand”
is money spent on the purchase of a good; price is the amount of money paid
per unit of the good when it is bought.) Next, one identifies the intermedi-
ate generalizations on which the Law of Demand is based, such as the fact
that more people can afford to buy a good if its price goes down, fewer if
it goes up. Continuing the reduction “downwards” to concretes: when Ford’s
Model T lowered the price of an automobile, a vastly greater number
of people could afford to buy one, and did. Likewise, falling prices for
big-screen televisions have greatly expanded their sales. To have a logically
complete reduction, more would be needed — for instance, reducing “afford”
to its base (a step unnecessary in most real-life cases) — but the example
should make clear the pattern of reduction.
The complete validation of an idea — the justification for accepting it as
a grasp of fact — requires both that it be proved by reduction to perceptual
reality and that it be integrated into the sum of one’s knowledge. A conclusion
has been fully proved only when it has been related, step by step, back to
perceptual data (the task of reduction) and has been checked for consistency
with the rest of one’s knowledge (the task of integration). Between these two
processes, man achieves a “double check,” as Peikoff terms it, on the validity
of his conclusions. [OPAR, 138]
 P R O O F A N D C E RTA I N T Y 265

The integration required is a positive act, not merely the failure to find
a contradiction, such as the failure to find any contradiction between
the observed price of tomatoes and the inferred size of the planet Jupiter.
Instead, a true idea will positively “fit into place” with the rest of one’s know-
ledge, unifying it into a coherent whole. For instance, the inference to Jupiter’s
size coheres with: its brightness in the night sky, the orbits of its moons,
its gravitational effects on the other planets and the sun, and even the
comparative rarity of asteroids hitting the Earth (Jupiter’s huge gravitational
field, on one theory, acts to sweep asteroids away from the inner planets).136
Since reality is an interconnected whole, an adequate grasp of reality will
capture the connections found in reality. In fact, we grasp reality by means
of becoming aware of relationships. Consciousness is an awareness of differ-
ences and similarities.
For instance, consider the knowledge that the Earth rotates on its axis. This
knowledge is reducible through a series of steps back to what one sees in the
nighttime sky. But this knowledge also explains facts that are not among those
used to reach it. As noted earlier, it explains why weather in the Northern
Hemisphere moves predominantly from west to east. Also, the Earth’s rota-
tion on its axis coheres with the facts, observable by the naked eye, that
sunspots move in only one direction across the sun (because the sun, too,
rotates), and with the fact that ordinary objects, such as a thrown rock,
virtually never move without rotation.
New conceptual knowledge is derived and validated by making logical
connections to antecedent knowledge. Implied thereby is an important
principle: no new knowledge contradicts old knowledge.
New knowledge can contradict old mistaken beliefs, but not old know­
ledge. Knowledge is a mental grasp of the facts of reality. Contradictions do
not exist in reality; newly grasped facts cannot contradict previously grasped
facts. Yes, the facts under consideration can change, as when a rock that
had been cool is warmed by sunlight; but a change in the object of know-

136 One can even, through sufficient intermediate steps, find significant relationships
between the price of tomatoes and the size of Jupiter. E.g., the price of tomatoes is stud-
ied by economics, the size of Jupiter by astronomy, so one belongs to the social sciences,
the other to the natural sciences; one fact is “man-made” in contrast to the other’s status
as “metaphysically given” (see Chapter 2); both the price and the size are measurements
and thus involve a standard of measurement (the dollar and the meter); both standards
of measurement have to be based on a unit adapted to the human scale of perception
[ITOE, 8], etc.
 H OW W E K N OW • 8 : P roof and C ertaint y 266

ledge is not what the principle refers to. Now, in the noonday sun, the rock
has become warm, but it remains eternally true that the rock was cool at
dawn. One’s knowledge of its state at dawn is not contradicted by learning
the changed state of affairs. (In addition, the change is lawful and proceeds
in accordance with the nature of the rock and of the incident sunlight;
if tomorrow, under the same conditions, the rock did not warm, that would
be a contradiction.)
The principle that new knowledge does not contradict old knowledge
rules out notions such as, “For the medievals, the world was flat; for us it is
round.” Such attacks on knowledge and certainty depend upon treating as
equivalent fact and error, truth and falsehood, science and fantasy — as if the
Earth had actually been flat until Columbus’ voyage somehow curved it. But
it is the error of the primacy of consciousness to imagine that the nature of
existence depends upon men’s beliefs, wishes, or social practices.
A slightly different line of attack on certainty is found in the popular
notion that the discovery of new conditions demolishes conclusions that did
not countenance them, as in the idea that “Swans are white” was demolished
by the discovery of black swans in Australia, or that Newton’s law of gravity
was refuted by Einstein’s theory of general relativity. But neither discovery
destroyed any previous knowledge; in fact, both discoveries expanded the
previously existing body of knowledge.
The generalization “Swans are white” could not logically have ­warranted
making the assertion: “There are no black swans anywhere in the world.”
That is not what was known at the earlier stage. The new knowledge is:
“Swans are white, except in Australia where some are black.” Thus, the end
result is more knowledge, not less. Similarly, Newton’s law of gravity was
supplemented, not refuted, by Einstein, because that law never extended
to the kind of conditions (super-huge masses and/or submicroscopic
distances of separation) that Einstein’s theory embraces.137
Knowledge is not an assemblage of out-of-context absolutes, as the skep-
tics assume in their attacks on knowledge. Any conclusion exists in and
depends upon a context, the context of the antecedent knowledge used
to establish it. Peikoff writes:

137 Einstein’s theory applies generally, even to the range of masses and distances that Newton
did refer to; but within this range, the deviation from Newton’s inverse-square law is too
small for Newton to have observed. This too is accommodated by a proper statement
of his law, which incorporates the proviso, “to the limits of accuracy in measurement
presently attainable.”
 P R O O F A N D C E RTA I N T Y 267

If a fact is inherent in human consciousness, then that fact is

not an obstacle to cognition, but a precondition of it. . . . In this
approach to philosophy, there is no “problem” of the senses, of
concepts, of emotions — or of man’s non-omniscience.
Man is a being of limited knowledge — and he must, t­ herefore,
identify the cognitive context of his conclusions. In any situa-
tion where there is reason to suspect that a variety of factors is
relevant to the truth, only some of which are presently known,
he is obliged to acknowledge this fact. The implicit or explicit
preamble to his conclusion must be: “On the basis of the available
evidence, i.e., within the context of the factors so far discovered,
the following is the proper conclusion to draw.” Thereafter, the
individual must continue to observe and identify; should new
information warrant it, he must qualify his conclusion accordingly.
If a man follows this policy, he will find that his knowledge
at one stage is not contradicted by later discoveries. He will find
that the discoveries expand his understanding; that he learns
more about the conditions on which his conclusions depend;
that he moves from relatively generalized, primitive obser-
vations to increasingly detailed, sophisticated formulations.
He will also find that the process is free of epistemological trauma.
The advanced conclusions augment and enhance his earlier know-
ledge; they do not clash with or annul it. [OPAR, 172–173]

Knowledge is a mental product. If one forms that product by employ-

ing the right method (logic) on the right material (prior knowledge), one is
fully entitled to claim the result as knowledge. Yes, there are occasional cases
in which one will be entitled to claim as knowledge a conclusion that later
turns out to have been a mistake, but that is not grist for the skeptics’ mill.
Man is neither infallible nor omniscient, and epistemic standards must
embrace this fact. The purpose of any standard is to guide men in how
to proceed within what is possible. There is no justification for erecting
a fantasy standard of judgment. [See Binswanger, 1981]
Using rational epistemic standards permits us firmly to distinguish
between an actual error (“Swans are pink” or “Gravity depends upon the
cube of the distance”) and a truth that has come to need qualification
(“Swans are white, except in Australia” or “Gravity depends upon the square
of the distance, and on relativistic factors”). In making this distinction,
it is helpful to consider the causes of error.
 H OW W E K N OW • 8 : P roof and C ertaint y 268

Cognitive errors result from a defect in the thought process or in its input
material. More concretely, cognitive errors result from one of three causes:
1) illogic, 2) false premises, or 3) incomplete information.
1. Illogic. When one departs from logic, the conclusion one reaches does
not follow from the evidence and premises used. A simple example is that
of making an error in adding up a long column of figures. If one follows the
rules of arithmetic, one will not err, but we sometimes slip up, depart from
the rules, and thus get the wrong sum.
In very simple, one-step reasoning, error is not possible. No one can
make a mistake in adding 2 + 1, or in combining “That’s ice,” and “Ice is
cold.” But logical missteps are not uncommon in complex calculations
and in the complex, multi-step reasoning used in everyday decision-
making.138 The possibility of error looms much larger in inductive inference,
which requires integration on a much wider scale. (See Harriman, 30–35.)
A person’s departure from logic may come from wanton irrationality,
sloppiness, slight negligence, ignorance, or just an innocent slip-up;
but whatever the illogic’s cause, its product is not knowledge.
2. False premises. Falsehood used as “input” to the process of inference
cannot result in a grasp of fact. Truth cannot be built upon error. One can-
not become aware of reality by weaving in elements of the unreal. (Even if
a given conclusion could have been reached from true premises, it does not
represent knowledge — i.e., a grasp of fact — when it is reached from false
premises.139)
The truth or falsehood of a premise is not a primary: any idea now being
used as a premise was itself reached as the outcome of an earlier conceptual
proce­ss. And the concepts composing the premise were formed in earlier
processes of concept-formation. The ultimate starting point of the conceptual
is perception, and perception cannot be in error (see Chapter 2). Thus, the
truth or falsehood of a conclusion depends upon the nature of the concep-
tual processes one has used to reach it — not just the current processes but
also the earlier processes, which led to the premises and to the concepts used.

138 Slip-ups become possible when the task exceeds the limits of the “crow,” and material has
to be shifted in and out of short-term memory.
139 Thus, the contemporary logical doctrine that “false implies true” is a serious error.
Although one can draw (narrowly) logical inferences about the consequences of believing
falsehoods, a conclusion based on a falsehood does not itself represent truth. A conclu-
sion based on a falsehood is not a recognition of reality. (However, the hypothetical form
of the proposition may be true: “If this were true, then such and such would follow.”)
 P R O O F A N D C E RTA I N T Y 269

So, errors resulting from false premises usually reduce to the first cause
of error: illogical processing (remembering that the standards for determin-
ing what is logical must be based on what is possible, not on an impossible
omniscience). These two causes of error reduce to: current illogic and past
illogic resulting in false ideas now being used as premises. (For completeness:
there is also the rare case of false premises resulting from the next factor.)
3. Incomplete information. Although this occurs very infrequently,
there are cases in which, despite being flawlessly logical, one reaches a false
conclusion because the data were both insufficient and apt to mislead,
because of their similarity to other things known. The simplest kind of case
is erring in identifying perceptual concretes — e.g., thinking the ­distant
hills are blue, that the straight stick semi-submerged in water is bent,
or taking a man’s twin to be him. Though these conclusions resulted from
(we may assume) logical conclusions concerning what was seen, ignorance
of certain facts result in one drawing conclusions that are errors nonetheless.
The judgments “Those hills are blue,” “That stick is bent,” “There’s Joe,”
are simply false, as further observation reveals.
As noted in the previous chapter, being logical does not guarantee
the truth of one’s conclusions. Nevertheless, standards of cognition must
be restricted to what is possible within the context of the information
available to the individual. The standards cannot require omniscience, which
is an impossibility. Thus, conclusions resulting from logical processing
(of material that is itself actual knowledge) are entitled to be claimed as
knowledge. That remains true even in the rare cases when later informa-
tion shows that the conclusion was an error. If it was fully rational to judge
as sufficient the information available when the conclusion was drawn,
the conclusion, though erroneous, was logical, and it would have been illogical
and irrational to draw any other. (This raises the topic of certainty, which will
be discussed shortly.)
A method of cognition is right if it is adapted to the requirements of
success; the fact that it does not make failure impossible does not weigh
against the basic method, only against the specific procedure that led to the
failure. (More on this topic will follow.)
The discovery of an actual error — even of a huge, fundamental error
— always represents cognitive progress. One knows more after discov-
ering the error than one did previously. And, determining the cause of
the error, and resolving to be on the lookout for its recurrence, makes
one less likely to err in the future, thus strengthening one’s cognitive position.
Some discoveries of error are painful, as when one discovers that what one
 H OW W E K N OW • 8 : P roof and C ertaint y 270

took to be a wart is in fact a cancerous growth. But the horrible shock is due
to the nature of the facts discovered, not to having been in error — as shown
by considering the reverse case: when what one took to be a cancer turns out
to be merely a wart, no one’s reaction in this case is: “Damn! I was in error!”
Leaving aside the content of what was discovered, the discovery of error is
always a gain in knowledge and, in that respect, is to be welcomed.
As one learns more about reality, one’s context of knowledge grows.
Contrary to the skeptics, it is not an error but an expansion of knowledge
when one discovers the existence of a new factor to be taken into account.
It is expansion, not refutation, when one learns in economics that prices
can rise due to inflation rather than to real changes in supply and demand
(as opposed to nominal changes), or when one learns that blood types
thought to be compatible with each other (type A with type A) are incom-
patible if the newly discovered Rh factors do not match. [OPAR, 173–174]
Likewise, it is not an error but an expansion when greater precision
or a wider scope of measurement results in a more complete formulation
of a mathematical relationship. (E.g., Tycho Brahe’s careful measurements
of the orbits of Mars enabled Kepler to identify that the planetary orbits
Copernicus thought to be circular were actually slightly elliptical.140)
Localized errors may arise on occasion, but if one adheres to logic,
know­ledge taken globally is a growing sum. Even when limited informa-
tion results in a (justified) error, the processes of reduction and integration
mean that errors are limited and are correctable in the light of new data.
Also, identifying the cause of any error allows one to guard against it in the
future. One learns that the color of hills is not to be judged as blue on
the basis of how they look from a distance, that a stick half-submerged
in water is not bent, or, to take a recent scientific correction, that some radia-
tion does escape a “black hole.”
(There are also lessons one can draw from cases of illogic; these involve
a refinement of one’s methodological tools — e.g., by becoming aware of and
guarding against logical fallacies, such as that of the stolen concept.)
Rather than being traumatized by the fact that logic does not immunize
one against all error, one must recognize that the identification of error as
being error is itself an illustration of the fact that we do know what we know.
For it is only on the basis of subsequent knowledge that an earlier conclusion
can be identified as erroneous.

140 Usually, it is not only greater precision in measurement but also a re-conceptualization
that is involved.
 C ertaint y 271

Instead of bewailing the fact that automatic omniscience is unavailable,

we must establish and adhere to cognitive standards that distinguish
between the right and the wrong ways of using one’s mind volitionally.
Cognitive standards cannot require the impossible; only what is under
one’s volitional control can be judged.

Certainty
Not all claims to knowledge are elaborate or in need of proof. In the case
of “There goes a fire truck,” one makes the observation, applies the relevant
concepts, and that’s that. Similarly, for knowledge reached by a chain of rote
steps, as in arithmetical calculations, one need only check that the steps
were followed. But there are other cases, cases in which an item of know-
ledge is acquired and validated by the accumulation of evidence over time.
For instance, at a jury trial, the presentation of the evidence may take months,
with an objective verdict becoming possible only at the end. In geology,
the theory of plate tectonics was first proposed in 1912, but it took half a cen-
tury of investigation before the theory qualified as having been proved.
Such cases, in which the evidence for a conclusion grows over time, give
rise to the idea of an evidentiary continuum and to the concepts that mark
off ranges along that continuum — notably: “possible,” “likely,” and “certain”
(and informal subdivisions, e.g., “barely possible” and “quite likely”).
In measuring evidence, the standard is set by the goal: proof. A given posi-
tion on the scale can be measured by what (approximate) portion of the jour-
ney to proof has been completed. Thus, the unit of evidence is fractional: the
ratio of the evidence at hand to the total set of evidence required for proof.
An idea is “certain” when the evidence for it is conclusive: one has acquired
and integrated all the evidence needed for proof, and contrary ideas have
no supporting evidence. An idea is “possible” when the evidence in its favor
is small in comparison to what would constitute proof, and there is some
evidence for contrary ideas. “Likely” is the middle case: one has a lot of
supporting evidence, but at least one contrary idea still qualifies as possible.
The relationship between certainty and knowledge needs clarification.
“Certainty” and “knowledge” are closely related but distinguishable concepts.
Knowledge is primarily differentiated from ignorance; certainty is primarily
differentiated from states that are less than certain: the possible and the likely.
“Certainty” refers to the cognitive status of an idea, which means it is a purely
epistemological concept; “knowledge,” in contrast, has both a metaphysical
 H OW W E K N OW • 8 : P roof and C ertaint y 272

and an epistemological component: to know something, the thing known

must exist (must be a fact) and one must have a mental grasp of it. “Fact,” in
contrast, is a purely metaphysical term: facts are facts whether or not anyone
knows them or has any evidence of their existence.
But do we actually need the concept “certain”? After all, “proved” already
names the status of being established by conclusive evidence. If “certain”
merely duplicates “proved,” then we should eliminate it, in accordance with
Rand’s Razor. But “certain” does add something to “proved”: the implications
for action. Proof establishes truth; certainty denotes the confidence that one
can have in acting, existentially and cognitively, on the idea. To identify an
idea as “certain” is to recognize that one can act on it without hesitation,
without doubt, without needing further deliberation or investigation.141
An important implication of this understanding of certainty is: one can
be certain but mistaken. A man can be objectively certain of a conclusion
that is, unbeknownst to him, false (although this occurs but rarely). There
is no contradiction in saying, “I was certain but wrong.” The meaning is:
“Given the state of the information I had at the time, I was fully justified
in believing it — the evidence required me to — but in the light of what
I now know, my conclusion was mistaken; the facts were not as I thought
they were, what I took to be knowledge turned out to be a mistaken belief.”
Thus, certainty is contextual. “Conclusive evidence” means “evidence that is
sufficient within a given context of knowledge,” not with an impossible omni-
science as its standard. But certainty’s contextuality does not make it ­subjec­­­­tive:
the standard of “sufficient” is not defined in terms of personal feelings
or social conventions, but in terms of logic. In a given context of knowledge,
the evidence for a conclusion is conclusive, rendering the conclusion certain,
when the totality of the evidence logically requires it. That is, within a prop-
erly delimited field of alternative hypotheses, only one conclusion explains
and is consistent with all the data, all the other hypotheses being excluded as
contradicting some or all of that data.142 In such a case, the evidence, taken
as an integrated sum, supports and is consistent with one and only one conclu-
sion. That conclusion is then contextually certain.

141 Any further information that comes along will be useful cognitively, will add to the fabric
of one’s knowledge. But, with regard to action, there comes a point at which gaining more
information cannot be expected to change one’s action-decision.
142 What I mean here by a “properly delimited field of alternative hypotheses” will be
explained shortly.
 C ertaint y 273

Note that the evidence must logically imply the conclusion, not the other
way around. Certainty is not achieved merely on the grounds that a given
hypothesis, if true, would explain the observed facts. What one needs for
certainty is that only a given hypothesis can explain all the observed facts.
And that “only,” in turn, assumes that one knows enough about this kind
of phenomenon to be certain that one of a relatively few hypotheses has
to be the correct one.
For example, suppose one observes tracks in the soil that have been made
by some passing animal. Suppose that the tracks are not too clear but have
a general shape consistent with the animal having been a fox. One cannot
reason: “Foxes produce tracks shaped like these, so it is certain that a fox
walked by,” because tracks with roughly similar shapes can also be pro-
duced by dogs, wolves, and coyotes. The track shapes, given the appropriate
context of knowledge, do establish that it is possible that a fox walked by, but
to claim certainty would be to commit the fallacy of Affirming the Consequent:
“If a fox walked by, then tracks with these shapes would be present; tracks
with these shapes are present, therefore a fox walked by.” Again, what is
needed is: “Only if a fox walked by would this kind of track be present.” That
statement is equivalent to: “If this kind of track is present, then a fox has
walked by” — which allows for the valid form of the hypothetical syllogism
known as “Affirming the Antecedent.”143
Now let’s add to the example other facts about the observed tracks beyond
their shape: their distance apart, their depth in the soil, the implied nature
of the gait, etc. If one has a knowledge of animal-tracking and knows what
animals inhabit the locale, and if only foxes produce tracks with all those
character­istics, then the evidence is conclusive that a fox made the tracks.
There is then no logical alternative hypothesis. The conclusion is certain.
The mere fact that one is ignorant of any alternative possibility is not
­sufficient grounds for claiming certainty; certainty does not flow from
­ignorance. Nor can one base certainty on the psychological issue of what alter-
native explanations one can or cannot think up. A presupposition of attaining
­certainty on a given topic is that one knows enough to make a rational
delimitation of the hypotheses, so that one knows that the true hypothesis is
within that delimited set.

143 Affirming the Antecedent: If A, then B; A; therefore B — valid. Affirming the Consequent:
If A, then B; B; therefore A — invalid. E.g., “If there is traffic, he will be late; there is traffic;
so he will be late” — valid. “If there is traffic, he will be late; he will be late, so there is
traffic — invalid: he could be late for other reasons.
 H OW W E K N OW • 8 : P roof and C ertaint y 274

These issues arose in the Workshops on Objectivist Epistemology, when Rand

was asked a question about the so-called “hypothetico-deductive” method
in science. Here is the exchange, as reprinted in the Appendix to ITOE:

Prof. M: Would you consider the following method of con-

firming a scientific principle to be valid? One formulates the
principle being guided by one’s knowledge of fact. Using the
­principle, one next deduces how entities under certain conditions
should act. Then, if one observes such action, and within
the context of one’s knowledge can account for it only by the
­principle which predicted it, it follows that the principle has been
confirmed. In summary, one induces the principle, deduces its
consequences, and if only that principle is known to give rise
to those consequences, which in turn exist, then the principle is
confirmed as a contextual absolute.
AR: This is outside the province of my book; this is the theory
of induction. But within this context, I would say, no, this would
not be the right procedure, and there is a danger of a very, very
grave error here. Because if you follow the procedure you outline
here, and you make certain predictions on the basis of a hypo­
thesis, and the entities do act accordingly, you conclude that you
can hold as a contextual absolute that it was your hypothesis
that was operating and that it is therefore true. You are assum-
ing an omniscience that contextual knowledge cannot permit.
Because since you are not omniscient, within the context of
your knowledge you cannot say that your particular hypothesis
was the only possible cause of the entities acting the way you
predicted. You would have to say this offers great confirmation
of your hypothesis, but it still remains a hypothesis and cannot
be taken as knowledge. Why? Because so many other possibili-
ties are involved. And I don’t mean unknown or unknowable
factors — I mean that it would be impossible, for any complex
principle of science that you are trying to establish, to elimi-
nate, even within your own context of knowledge, all the other
possibilities.
What I would question is this part of the procedure: “if only that
principle is known to give rise to those consequences” — that’s the
mistake of arrested knowledge, right there.
 C ertaint y 275

Prof. M: Even though it is relative to what you know at

that time?
AR: Even though it’s at that time and it’s your full context of
knowledge. Because you cannot conclude that something which is
not fully known to you can be produced only by one ­hypothesized
factor. On the basis of that same context of knowledge, any
number of hypotheses could be constructed. Which is why we
need hypotheses. If it were otherwise, then your hypothesis
to begin with would almost have to be a certainty. [ITOE, 301–303]144

In a scientific induction, or in an extended investigation to determine

a concrete fact (e.g., crime detection), one uses background knowledge to
establish a kind of “genus” that delimits the field to cases that warrant inves-
tigation. This narrowing of the field to a genus of possibilities must be based
on knowledge; one must be certain that the truth lies in one of the possibili-
ties in that genus. Only that delimitation enables one to conclude objectively
that certainty is reached when only one hypothesis supports and is consistent
with the totality of the evidence.
Normally, the available evidence is consistent with only a small genus
of possibilities and rules out the rest of the universe, as it were. If only three
people had access to the crime-victim during the time when the crime
was committed, then these three are ruled in and the rest of the world’s
population is ruled out. These three become “suspects,” in view of the fact
that the available information is consistent with any of them being guilty
and inconsistent with anyone else being guilty.
The assessment of an idea as “possible” or “likely” is not a guess or specu-
lation but a cognitive assertion — a claim to have some actual knowledge,
for instance, knowing that foxes leave tracks of a certain shape. To evaluate
an idea as being either possible or likely, one must know the evidence for
it and know that it is evidence — i.e., that it advances cognition, moving it
forward in the direction of the conclusion. But when alternative hypotheses
are still possible, i.e., still have some evidence supporting them, the idea
remains a hypothesis, not a certainty.
Evidence is data that advances us toward knowledge. This means that
evidence must distinguish among alternative hypotheses, favoring one over
others. In crime detection, for instance, suppose we ask: when the perpetra-
tor’s blood is found at the scene of a homicide, does its blood type constitute

144 See also the discussion of inertia and gravity on p. 370.

 H OW W E K N OW • 8 : P roof and C ertaint y 276

evidence? That depends on whether it discriminates among the suspects.

Suppose we know with certainty that the killer is one of three people,
but they all have the type B blood that was found at the crime scene.
In that case the blood type is not evidence as to which man is the murderer.
But if one suspect has type O blood, then the blood analysis is evidence
because it weighs against the hypothesis that he is the murderer.
Now take a case of induction from the history of science: Galileo’s investi­
ga­tion of the nature of motion. In his era, the dominant view was that objects
move in a circular path unless something forces them to move in a straight
one. The evidence for this idea is the motion of the heavenly bodies, plus
the fact that an object thrown follows a curved path. Galileo considered the
alternative hypothesis: objects move in a straight line unless some force con-
strains them to a curved motion. The evidence for this contrary hypothesis
is that a dropped object moves straight downwards and that a curving
motion involves a continuous change in direction, and changes must have
a cause. But what is ruled out? The idea that objects “naturally” (without
constraints) move in a zigzag, back-and-forth motion. There is no evidence
for this dreamed-up notion. Or, the idea that objects’ unconstrained motion
depends upon their color or the day of the week or someone’s prayers.
In his study of motion, Galileo constructed experiments ­demonstrating
the independence of the horizontal and vertical components of motion. The
experiments expanded the context of knowledge, enabling G ­ alileo to con-
clude that it is motion in a straight line, not circular motion, that is “natural.”
(Later knowledge, particularly of gravity as a force exerted by all matter on
other matter, confirmed Galileo’s hypothesis by explaining the circular motion
of heavenly bodies.) In induction, as in crime-detection, observational
evidence is the only coin of the realm.145
On the basis of this understanding of the role of evidence in assessing
the cognitive status of an idea, we are prepared to rescue certainty from
a long-standing attack.
For a conclusion to be certain, there must be no other rational hypo­
thesis that qualifies as possible. But the concept of “possible” has long
been subjected to a disastrous “package-deal” — a false integration that
packages epistemic possibility with metaphysical possibility. The two are
entirely different. Metaphysical possibility denotes an ability, potentiality,
or capacity. Epistemic possibility denotes the status of the evidence.

145 On standards of certainty in scientific induction, see The Logical Leap [Harriman, 2010],
especially pp. 184–187 and 238–239.
 C ertaint y 277

In English, metaphysical possibility is expressed by “can” and epistemic

possibility by “might.”146 I can shoplift, I am able to do it, but that does not
imply I might. Metaphysical possibility, by itself, does not warrant the asser-
tion of epistemic possibility. Can be does not imply might be.147
To say that something might be the case is to treat it as a hypothesis.
Forming something as a hypothesis is a positive mental act that has to be
justified. The required justification is evidence. Hypotheses require evidence.
For something’s metaphysical possibility to serve as evidence warranting
making it into a hypothesis, one must have grounds for thinking that the
metaphysical possibility has been actualized.
I can whistle “Yankee Doodle”: I know how to whistle and I know the tune.
But my having that ability does not, per se, provide any evidence that I am
whistling it now, as you read this. For you to be justified in thinking that
I might be now whistling “Yankee Doodle,” you would need some evidence
for that idea, some reason to think that I am actualizing my potential to
whistle it. If I am scheduled to perform at a public whistling event at about
the present time, and if “Yankee Doodle” is on the program, then know-
ing all that would justify thinking that I might be whistling it right now.
But the mere fact that I can whistle it — that it is metaphysically possible
for me to whistle it — does not count as evidence that I am doing so now.
And in the absence of evidence, the idea is not to be accepted or entertained,
but dismissed.
Dismissing “X might be the case” does not mean holding X to be impos-
sible; the term “impossible” has only a metaphysical use. To know that some-
thing is impossible is to know that it contradicts the facts. It is ­impossible for
man to fly by flapping his arms — that would contradict the nature of man, the
nature of gravity, and the nature of flight. In dismissing possibilities asserted
only because they are metaphysically possible, one is not saying that the thing
dismissed is impossible. One is simply refusing to entertain it as a hypothe-
sis — because there is no reason to do so. It is not impossible for me to whistle
“Yankee Doodle” — unless I am asleep, I am gagged, etc. But “not i­mpossible”
does not translate into positive evidence, and it would be irrational to
entertain the possibility that I might now be whistling “Yankee Doodle” —

146 “Might” and “may” differ in degree of likelihood. “I may attend” indicates a stronger likeli-
hood than “I might attend.” Thus, everything I say about “might” applies with even more
force to “may.”
147 Equivalently, “possible for” is metaphysical and “possible that” is epistemic. The fact that
it is possible for me to shoplift does not imply it is possible that I will shoplift.
 H OW W E K N OW • 8 : P roof and C ertaint y 278

or shoplifiting, dancing a polka, reading Gray’s Anatomy, or doing any of an

unlimited number of things, merely because my doing them is not impossible.
The “crow epistemology” comes into play here. Treating something as
a hypothesis has cognitive costs: one has to form a mental file for the issue
and carry the issue forward for further pursuit, consideration, and integra-
tion. Having too many hypotheses at once overloads the “crow,” frustrating
attempts at cognition.
When there are too many alternatives for the mind to work with on an
ongoing basis, these alternatives do not qualify as epistemic possibilities.
In the terminology of law-enforcement, one cannot have one thousand
“suspects.” That would be a misuse of the concept “suspect.” And in general,
to grant something the epistemic status of “possible,” one must have evidence
— evidence that is specific to the case at hand and that a mind limited by the
“crow” can work with on a continuing basis.
Thus, when an idea is supported by contextually conclusive evidence,
one’s certainty in that idea is not threatened by the mere metaphysical
possibility of an alternative.

The Law of Rationality vs. The Arbitrary

One’s proper attitude to evidence is summarized in what some logicians have
called “The Law of Rationality.” In my formulation, this law states:

In reaching conclusions, consider all the evidence and only the evidence.

Because evidence is the only means of gaining inferential knowledge,

the rational mind accepts all that which the evidence shows, only that which
the evidence shows, and only to the extent that it shows it. Only evidence
— not someone’s assertion, not feelings, not authority, not faith — can pro-
vide the basis for proceeding cognitively.
The question then arises: what does one do when there is zero evidence
in favor of a claim?
Let me make up a scenario in which there is zero evidence in support
of an idea. Suppose the idea pops into your mind that you will inherit a fortune
from some distant relative whom you have never heard about. I assume that
the idea has no evidence to support it. That means it has no cognitive content.
The notion does not fall on the evidentiary continuum. It is neither possible,
 T he L aw of R ationalit y vs . T he A rbitrary 279

nor likely, nor certain. It has no cognitive status. If such an idea is none­
theless asserted as being cognitive, it has an anti-cognitive status: arbitrary.
The term “arbitrary” does not refer merely to a state of ignorance but to
ignorance taken as an epistemological license, as if the ability to imagine
something made it cognitive. “Arbitrary” means: put forward on the premise
that evidence is unnecessary, that one can assert anything one has dreamed
up, and that this assertion stands until and unless it is refuted. An arbitrary
assertion is one made in defiance of the need for such a thing as evidence.
As such, it represents an assault on logic and cannot be countenanced.
The cognitive status of an idea is an outgrowth of the process by which
it was reached. To be cognitive, that process must be one of observing
facts and drawing logical inferences therefrom. Absent that, the idea is
­non-cognitive. To accept it anyway — even to entertain it as a possibility
— is c­ ognitively toxic. Yet, indulging in arbitrary assertions, possibilities,
and “constructs” is standard operating procedure today in philosophy,
in the culture, and in political debate; it has even infected the field of law
(see the upcoming example of arbitrary “what ifs” raised by the defense
in the O. J. Simpson trial).
Knowledge is a mental grasp of fact. When there are no rational grounds
for an idea, there is no means of achieving that grasp. Nor does an arbitrary
claim constitute even a step on the path to gaining knowledge: it does not
represent an awareness of anything and thus does not represent a cogni-
tive advance. It leaves one in the same cognitive state one was in prior to
its assertion. In my example, the idea that you will inherit a fortune does not
move you closer to grasping any fact on any subject.
To be engaged in cognition, one must accept and implement the means
of acquiring knowledge. The means of acquiring knowledge is awareness
of evidence — facts, implications, relationships, causal connections, etc.
— information that moves one toward one’s cognitive goal. An arbitrary
idea, by definition, has no such means. As such it is a cognitive nullity, and
the Law of Rationality requires that one refuse to consider it; the claim must
be summarily dismissed. (Again, this is not to hold that what is claimed to
be the case is impossible.)
The analogy here is to an action at law. The mere lodging of a complaint is
not sufficient grounds for holding a trial. The plaintiff must provide sufficient
evidence to show that a trial, a costly and onerous affair, is warranted. Absent
such a showing, the complaint is dismissed. This dismissal is not a determi-
nation that no legal wrong occurred, only that there is no basis for holding
 H OW W E K N OW • 8 : P roof and C ertaint y 280

a trial; analogously, in epistemology, the dismissal of an arbitrary assertion is

not a determination that the assertion is false, only that there is no rational
basis for entertaining the notion — not even for undertaking to refute it.
In the cognitive sense, there is no “it” to deal with.
In merely entertaining the arbitrary, one has suspended logic, since logic
deals with evidence. Accordingly, logic cannot be used to guide what one
does with the arbitrary. The anti-logical premise that admitted the baseless
idea into consideration prevents one from knowing what to do with it, how
to integrate it, and what it means or implies. Either one accepts the require-
ments of cognition, or one does not. If one does not, there is no answer to:
what is the logical next step to take in working with this idea?
Although logic is not a game, like a game it has rules. Consider, then,
the analogy to changing the rules of a card game at whim. Suppose that
during a game your opponent announces “My three of clubs is now a king
of clubs.” If you accept that, how do you know how to proceed in the game?
Can you now play your three of hearts and have it accepted as a king of hearts?
What happens when the actual king of clubs appears? Does it become
a three of clubs? Or, if it stays a king of clubs, what is its relation to the other
“king of clubs”: higher? lower? equal? When will the next arbitrary rule-
change be announced, and what will it mean for your play? There is no way
to answer these questions once one accepts the premise that the rules may
be changed at whim. And there is no way to answer questions about the
meaning and implications of an idea once one accepts arbitrary assertions
as being cognitive.
Arbitrary ideas are products of imagination. Imagination is entirely proper
— in fact, indispensable — across a broad range of rational activities, but
imagination is not cognition. Imagination is not limited by logic or evidence;
it is the rearrangement of stored images, ideas, or words, according to internal
considerations, such as wishes or fears, which may or may not be realistic.
One can take the idea of gold and the idea of mountain and combine them
in imagination to project a golden mountain. This may be entirely appropriate
as part of a fairy tale. But to treat such fantasies as acts of cognition is an act
of high treason against logic.
To have cognitive content, a claim must first have content — it must for-
mulate a definite, intelligible idea. The assertion, “Something bad is going
to happen” lacks such content, as does “My computer’s problems are due to
gremlins.” Arbitrary ideas ultimately lack content because they are made up.
The content of an arbitrary idea consists of whatever the person dreaming it
 T he L aw of R ationalit y vs . T he A rbitrary 281

up says it consists of — until he announces that it has changed (“The gremlin

teleported himself to another planet as soon as you started looking for him
inside the computer”).
A content of imagination has only as much identity as its imaginer has
conferred on it so far, with the duration of his whim replacing the stability
of existents. As an experiment, picture a toothbrush in your imagination.
How many bristles does it have? There is no answer to that question; you
might provide a number, by making a decision as to what the number shall
be, but there is no objective quantity of bristles waiting to be counted.
Now, imagine brushing your teeth with it. Did it lose any bristles in the
process? There are no facts to find out here, only a decision to be made as to
what you wish the answer to be. The imagined has no i­ndependent identity.
The arbitrary, because it lacks an independent identity and has no connection
to the rest of one’s knowledge, cannot be dealt with cognitively. The arbitrary
is a cognitive dead end. It stops the mind.148
In saying that the mind is stopped by the arbitrary, I mean the rational,
cognizing mind. Imagination can, of course, proceed. The crucial point is
not to confuse imagination with cognition, not to pretend that one is pursu-
ing cognitive contact with reality when the means of doing so — evidence
— is absent.
The arbitrary is not involved when someone supports a claim by what
he believes to be evidence but turns out not to be. In these cases, there
is a rational way to proceed: one can expose the errors in the assertor’s
reasoning, showing how the “evidence” he offers is not actual evidence for
his conclusion. Here, in order to dismiss the claim, one is logically obligated
to deal with any proffered evidence (unless it is transparently absurd); one
has to show that the evidence has been misinterpreted, or that it is not fac-
tual, or that it doesn’t support the conclusion. One cannot arbitrarily assert
that a claim is arbitrary.
When the alleged basis for a claim has been invalidated, the claim returns
to a non-cognitive status, as though it had never been brought up in the
first place. But if the assertor does not accept this dismissal, if he continues
to assert the idea in defiance of its demonstrated baselessness, there is no
way to deal with him or his claim. Lacking cognitive content, the idea must
be dismissed without consideration. Otherwise one is surrendering logic,
abandoning one’s cognitive guide.

148 This is a point made in several lectures by Leonard Peikoff; see the general discussion
in OPAR, 163–171.
 H OW W E K N OW • 8 : P roof and C ertaint y 282

The consequences of granting even a “maybe” to arbitrary assertions are

deadly. Consciousness is a biological faculty. Its survival role is to guide an
organism’s actions toward that which serves its life and away from that which
threatens it. Man’s control over his conceptual faculty gives him the ability
to create in imagination new combinations of his mind’s stored contents;
but if he confuses the content coming from reality (evidence) with whatever
he chooses to rearrange in his mind, if he confuses combinations dictated
by logic with those generated by his imagination, his mind’s output is simply
junk, not a means to guiding his actions and succeeding in his life.
Again, one needs imagination in regard to planning, self-­motivation,
literature, amusement, and other purposes; but in these cases, one is not
holding that the imagined content is true.
Once granted cognitive standing, the arbitrary cannot be rationally con-
fined, because reason has been dispensed with. Any “limits” set would
themselves be arbitrary — which means they could be adjusted, extended,
or revoked arbitrarily. Since reason never endorses the arbitrary, since the
acceptance of the arbitrary is always counter to reason, a compromise with
the arbitrary surrenders the supremacy of reason. Henceforth, feelings
dictate where reason is to operate and where it is not. Reason and logic then
function only by permission of feelings.
One must have a principle of methodology: either one recognizes that
imagination is imagination, or one does not. If one compromises the prin-
ciple here, the result is that reason is no longer supreme in one’s mind.
The gates to the realm of the arbitrary bear the inscription: “Abandon all
hope of cognition, ye who enter here.”
One cannot object: “But sometimes the arbitrary idea will turn out to have
been true.” There is no such thing. Whatever transpires, it cannot represent
an arbitrary claim “turning out to have been true.” Non-awareness cannot
“turn out to have been” awareness. Yes, the words used to make an arbitrary
utterance may be the same as the words used later, in a cognitive context,
to express a true proposition; but that does not mean the arbitrary utterance
was an act of awareness or that one judged correctly. When an idea is put
forward without any grounds, it is not expressing a judgment at all, merely
engaging in imaginative projection.
Since the arbitrary is actually fantasy, it is instructive to compare it with
dreams. Suppose a man asleep dreams that he meets a beautiful red-headed
woman, and suppose that the next day he does. It is not the case that his
dream “turned out to be true.” The woman he meets is not “the woman
 T he L aw of R ationalit y vs . T he A rbitrary 283

from his dream.” His dream had no actual referent in reality. A reference to
r­ eality requires an awake mind with mental content that has some ­cognitive
­connection to reality, i.e., some awareness on which it is based. Where that
awareness is absent, there is no reference at all. Dream images are merely
a succession of internal experiences; the images may resemble things
encountered in the external world, but they are not about the world;
they make no cognitive reference to reality. To paraphrase the old legal
preface to fiction stories: any resemblance between the dream’s content and
actual persons or events is purely coincidental.
To see the vacuousness of an arbitrary assertion, take the following case.
Suppose a parrot squawks, “It is raining now.” Sometimes when it squawks
that, it is raining, sometimes it is not. But the parrot does not go from
uttering truth to uttering falsehood with the changes in the weather.
The parrot squawks are, for it, just squawks. Words do not have an intrinsic
connection to reality: that reference depends on the mental actions of
the speaker or listener. (In terms of just the sounds, “It is raining now”
is identical to “It is reigning now” and “It is reining now,” so clearly any
­reference of parrot-sounds to the weather depends on the meaning supplied
by the human being listening to them.)
Accordingly, an arbitrary claim is not to be taken as true, nor as even
possibly true. But neither is it false. To be false, a statement must first
say something: it must attribute some definite characteristics to a clearly
designated subject.149 An idea’s falsehood is established by reference to a body
of knowledge — knowledge contradicting the idea. [OPAR, 166] When you
know that something isn’t so, you know it by reference to your awareness of
what is. But arbitrary assertions, lacking evidence, provide no such aware-
ness, so they do not reach even the level of being false.
As explained in Chapter 5, negative terms like “not” mean “different
from.” Take a reasonable but false statement, such as: “There is a soup bowl
on Joe’s dining room table.” (Assume that it is now dinner time and that one
knows that Joe often has soup at dinner, making this a reasonable statement.)
To learn that this claim is false, one could look at Joe’s tabletop and see what
is — i.e., the wooden expanse and whatever other objects are on it; what one
sees has a different identity from that of a soup bowl. The evidentially sup-
ported (but false) statement “There is a soup bowl on Joe’s dining room table,”
is definite and intelligible. Contrast it with: “There is an invisible aura sur-
rounding Joe’s dining room table.” That has neither intelligible content nor

149 See Chapter 5 on the meaning of a proposition.

 H OW W E K N OW • 8 : P roof and C ertaint y 284

cognitive standing. We don’t know what it even means; we don’t know what
would serve as evidence for it or against it.
The Law of Rationality, again, states: In reaching conclusions, consider all
the evidence and only the evidence. The Law does not merely demand that
you justify your conclusions, it also demands that you justify your ­cognitive
actions — justify even spending time considering a claim. It is irrational
to consider that for which there is no evidence.
Thus, arbitrary ideas are not even to be entertained as hypotheses,
as “maybes.” Things that are on the way to being established as true or false
(hypotheses grounded in evidence) are different from notions that are just
dreamed up. Arbitrary assertions are not propositions at all, but pseudo-
propositions: words with the linguistic form of a proposition, but without
cognitive meaning.
There are an unlimited number of imaginary notions that one can
construct, once evidence is regarded as unnecessary. For any given arbitrary
idea, one could construct 10,000 alternative ideas on the same non-basis.
Take the earlier example of the arbitrary assertion that you will inherit
a fortune from an unknown relative. If that is to be granted cognitive respect,
why not: “You will inherit a crushing debt”? Or, “You will inherit a dog,”
“You will inherit a cat,” “You will inherit a baseball autographed by Mickey
Mantle,” “You will inherit a lamp with magic powers,” and so on ad infini-
tum? There is no logical way to prefer any one of these notions to any other.
And one cannot deal with a deluge of alternative hypotheses.
Newton famously wrote “hypotheses non fingo” (I feign no hypotheses).
In light of the Law of Rationality, one could formulate “Newton’s Razor”:
hypotheses are not to be multiplied beyond the evidence, nor are they to be
dismissed in disregard of the evidence.
The rational response to an arbitrary assertion is to dismiss it — i.e., not to
assume its truth, not to assume its falsity, not to take it as a hypothesis, not to
try to refute it, but to recognize it as fantasy and turn one’s attention to reality.
It is important to realize that for an idea to be disproved, for it to qualify
as false, the principle of dismissing the arbitrary must already be accepted
and adhered to. Otherwise, there can be no disproof and nothing can be
established as false (or as true). The refutation of any claim presupposes
that the arbitrary is not logically entitled to a refutation, that it has no cogni-
tive standing. Otherwise, there is always the comeback, “Maybe you erred,”
or “Maybe you’re dreaming all this.”
 T he A d I gnorantiam Fallac y 285

Thus, the idea that dismissing the arbitrary is only tentative, compared
to having a disproof, is contradictory: dismissing the arbitrary is a precon­
dition of disproof.

The Ad Ignorantiam Fallacy

Consider what qualifies as evidence and what does not.
The mere fact that someone has asserted something is not evidence for its
truth. (Testimony from a credible source about what he observed is funda-
mentally different from raw assertion.150)
The fact that no counter-evidence has been offered is not evidence. The
absence of evidence against is not evidence for. By the same token, the
absence of evidence for is not evidence against. Yes, there are cases in which
one has looked for positive evidence, where such should have been avail-
able, and found only facts different from those asserted (as with seeing only
Joe’s tabletop where one would have seen a soup bowl if it had been there).
But this is awareness of facts, facts from which an implication can be drawn
contrary to a rational hypothesis. The logical point is: sheer unawareness of
evidence is not itself a form of awareness, is not evidence. To assume other-
wise is to commit the logical fallacy known as Ad Ignorantiam — the appeal
to ignorance. Nothing follows from ignorance.
I must register another caveat regarding what counts as evidence. The
epistemological issue of the amount of evidence is not to be confused with
mathematical probability. A judgment of an event’s probability uses known
facts plus a rational theoretical framework to predict a relative frequency,
as in calculating how many times out of 100 a fair coin can be expected to
land heads up when tossed. The epistemic question is different: how well
established is a given idea?
For instance, there is a mathematical probability assignable to winning
a lottery. But your holding one ticket out of a million sold does not count as
any evidence that you will win. Your winning is not impossible, but neither
is it a “maybe.” It is valid to say “The odds of my winning are one in a million,”
but “I might win” is not a valid hypothesis for you to form.

150 In the case of another’s testimony, the evidence one needs is not of the content of
his report (beyond it being consistent with what you know) but of its credibility as
a report — i.e., knowledge of the honesty, reliability, and accuracy of the reporter.
The reporter, if thus qualified, is serving as your eyes and ears. This is why courts do not
allow as testimony conclusions that a witness drew, but only the facts that he observed.
 H OW W E K N OW • 8 : P roof and C ertaint y 286

If you were to regard winning as a hypothesis, you would, by the same

principle, have to form hypotheses for all other “not impossible” events,
such as contracting a rare tropical disease, getting hit by falling space-debris,
happening upon the U.S. Ambassador to Guatemala around the next corner
— and so on, without limit. But one cannot hold an unlimited number of
hypotheses. Attempting to do so would merely make the word “hypothesis”
into an empty sound, and a new word would have to be used to refer to ideas
having some actual evidence, not just a mathematical probability.
The lottery example spotlights the absolute difference between refusing to
consider something, even as a hypothesis, and claiming to know that its oppo-
site is true. To dismiss the idea of winning the lottery is not to say, “I know
for a fact that I won’t win.” You do not know that. To dismiss is to refrain
from doing something — i.e., it is nonaction. Dismissing the arbitrary is
refusing to squander limited cognitive resources on pursuing notions merely
on the grounds that they cannot be ruled out as impossible. Dismissing
X is very different from claiming not-X. They share the complete rejection of
X, in the one case by rejecting it as arbitrary and in the other by rejecting it
as contradicting known facts. The difference is the same as that between not
investing in a given stock and selling the stock short.
However, to know that something is false is not a “stronger” rejection than
dismissing it as arbitrary. The person who rejects the idea of God as arbitrary,
as imagination, does not have a “weaker” rejection of that idea than the per-
son who rejects the idea of square circles. Whether an assertion is groundless
or contradicts known facts makes no difference to one’s epistemic attitude
toward it. In both cases, one totally refuses to accept it. There’s no “strength”
added to one’s rejection of an idea that comes with seeing that it is impossible,
as if one could say: “Previously, I merely rejected the idea, now I really reject
it.” That’s nonsense.

The Burden of Proof Principle

One doesn’t need a reason not to consider something a fact. One doesn’t need
a reason not to entertain something as a hypothesis. The reverse is true:
The burden of proof is on him who claims knowledge.
Knowledge is an effect of the operation of certain causes. For the effect
to be present, the causes must have been present. Ignorance, not knowledge,
is the default condition. Thus, a claim to have achieved knowledge (even
 T he Burden of P roof P rinciple 287

knowledge of possibility) must be supported by showing that the cause was

present and operative. The cause is awareness — direct perception of the
thing or awareness of evidence logically supporting it. In the absence of such
awareness, the claim to know is arbitrary and thus is to be dismissed.
Some formulations of the Burden of Proof Principle make reference
to asserting a “positive.” But “positive” is an ambiguous term that raises
a tangle of unnecessary questions here. In one sense, any claim is a “positive”:
it is a claim to having achieved knowledge. Even the claim that something
is possible is a claim to knowledge — the knowledge of certain facts and
the knowledge of how far they support the conclusion. In another sense,
only claims that assert the existence of an entity are taken as “positive.”
In a third sense, claims that assert that an entity possesses a given character-
istic or is acting in a certain way are taken as “positive.”151
But these are needless issues. “Concepts are not to be multiplied beyond
necessity.” That canon (“Rand’s Razor”) applies to forming concepts of
“positive” and “negative” statements: there is no need for those concepts.
Instead, the required distinction is between the cognitive and the arbitrary.
The Burden of Proof Principle does not concern the kind of content a claim
has but that it is a claim to be aware of something, to know something.
What has the burden of proof is any claim to have achieved knowledge
(even of possibility).152
To know something, one must have used the means of gaining knowledge:
evidence. That is all that the Burden of Proof Principle states. When there is no
evidence for “S is P,” there is no awareness of S being P, nor of anything
indicating that S is P. That means there are no grounds for hypothesizing
that S is P.
The Burden of Proof Principle also applies to negative statements: “S is not P”
or even “S does not exist.” Take the statement, “There is no life elsewhere
in the universe.” Given the state of our ignorance (or, if you prefer, the

151 Here is a real-life example. I just flipped a coin, and it landed on my desk. It came up
either heads or tails. If you, who don’t know which side came up, assert “It was heads”
is that “the positive”? The proper assessment is: you know, from your knowledge of coins,
that it is possible that it was heads and it is possible that it was tails. To claim that it
was one over the other is arbitrary. Both “It was heads” and “It was tails” are claims to
knowledge and thus both face a burden of proof — which in this case you cannot meet.
(P.S., it came up tails.)
152 The negative propositions discussed in Chapter 5 (e.g., “Lassie is not a beagle”) are
negative in structure, not in content. Such propositions do make claims and thus require
justification. What does not require justification is non-consideration; that is the default
state. One needs a reason to consider, not a reason not to consider.
 H OW W E K N OW • 8 : P roof and C ertaint y 288

ignorance of 100 years ago), there is no evidence for this “negative”

claim — as there is none for the corresponding positive: “There is life
elsewhere in the universe” (probabilistic appeals to the huge number
of stars in the universe notwithstanding). Where there is no evidence
on a subject, the Law of Rationality directs us not to pretend there is
— i.e., not to consider it, one way or another. Thus, there is a burden of proof
for claiming to know that something does not exist, because that means
claiming to know facts that exclude it. Both “I will get leukemia in the next
month” and “I will not get leukemia in the next month” have a burden of proof.
In the absence of evidence, neither idea is to be entertained.
By the same token, the familiar idea that it is impossible to prove a
“negative” actually means: it is impossible to disprove an arbitrary assertion;
it is impossible to proceed under the assumption that any assertion stands
until refuted, or that only a disproof of an assertion would justify not accept-
ing it as a “maybe.”
One does not need evidence against an idea in order to dismiss it — i.e.,
not to entertain it. For dismissal, the most that one needs is to show that one
is informed of the state of the evidence and has found that there is none sup-
porting the idea. (This is not a claim to be aware of non-existence: it means
that one has looked where that evidence should have been but has found
only data different from evidence for the assertion.) And, if there has been
anything put forward mistakenly as evidence, in order to dismiss the claim
one must identify why the purported evidence is not, in fact, evidence.
The Burden of Proof Principle is a corollary of the Law of Rationality;
the principle implies that when a claim to knowledge is unsupported by
evidence, we must not even consider it. Ad Ignorantiam adds the fact that
ignorance is not a source of evidence.
There are no degrees of meeting the burden of proof: it is either met or not,
period. “Proof ” requires conclusive evidence, evidence beyond a reasonable
doubt. (Unreasonable doubts are arbitrary.) When an idea has fully met that
burden, it is certain. Even though a lesser degree of evidence qualifies an idea
as likely or possible, the burden exists for proving conclusively, with certainty,
that it is likely or possible. Possibilities have to be proved. To prove possibility
is to prove that the idea is evidence-driven, that one knows a subset of the
facts required to attain certainty on the issue.
Take the earlier example of seeing certain tracks in the soil. Assume that
one is warranted in saying, “I know that a small animal walked by here;
the tracks are unclear, but it is possible that they were made by a fox,
 T he Burden of P roof P rinciple 289

since the tracks are consistent with fox pawprints and foxes do inhabit
this region.” By pointing out that evidence, one has fully met the burden
of proof for concluding: “It is possible that a fox recently walked by here.”
But now take a case when that burden of proof is not met. Suppose there are
no tracks, and the only “evidence” is that one dreamed last night that a fox
walked by. In that case, the idea has no cognitive content, and The Law of
Rationality demands dismissing it as arbitrary.
Or, take the issue of the existence of God. The burden of proof is on
him who claims to know that God exists, or even that “maybe” God exists.
One doesn’t need a reason not to believe in God. Rather, one needs a r­ eason
to believe. In the absence of such reason, one is logically required not to
believe. This is not agnosticism, but atheism. Agnosticism is the claim that
God is a valid hypothesis — “maybe God exists, maybe not, we can’t know.”
This kind of fence-sitting is precisely what the Burden of Proof Principle
rules out. Given that there is no rational evidence in support of a God, either
one entertains the idea anyway or one does not. There is no “third” or “mid-
dle” position.153
When the burden of proof is not met, the claim must be rejected, no
disproof being required. The process of disproof presupposes that one
accepts logic as an absolute, that one is not granting cognitive significance
to arbitrary assertions. But on the premise that the arbitrary is entitled
to a refutation, that it is good until and unless disproved, nothing can
be proved or disproved. As to disproof, the assertor can always dream
up new arbitrary content to back up his prior claim. The ultimate such
arbitrary backup is: “I say there is a mistake in your refutation; now you have
to prove there isn’t.” Whatever one says in response to that, the assertor of the
arbitrary need only repeat his claim, asserting that you are mistaken about
that. You have agreed to play a game that you cannot win.
On any given issue, there is only one truth but an unlimited number of
non-truths. To spend one’s time eliminating non-truths (if one could even
do that) would mean getting nowhere cognitively. One would be preventing
oneself from spending that time learning about reality.

153 Given a semi-intelligible definition of God — e.g., an omnipotent, omniscient, immate-

rial consciousness — one can show the contradictions in such a notion; but doing that
is not required in order to reject God’s existence, even as a possibility. And note that
such disproofs do not faze committed theists, who immediately take refuge in arbitrary
rejoinders such as, “Such contradictions are resolved from God’s perspective,” or “You
can’t know someone won’t come up with a definition of God not subject to your criticism.”
 H OW W E K N OW • 8 : P roof and C ertaint y 290

The application of the Burden of Proof Principle to the possibility of errors

due to incomplete information deserves more discussion.
Suppose you have a longstanding friendship with a man. By everything
you know, he is an honest, industrious person. But later you discover that
for the past year he has been embezzling from his firm. Let’s assume that
during this year you were objectively certain that he was honest, having
years of evidence of his honesty and absolutely no evidence of his current
misdeeds. Nonetheless, you were mistaken; you lacked the crucial information:
that he had stooped to embezzlement (and of his inner state that made
this crime possible to him). The painful discovery of this man’s dishon-
esty is, however, no grounds for doubting the honesty of your remaining
friends. The fact that a good person can go bad, even coupled with your
experience with the embezzling friend, provides no grounds for doubt of
anyone else. It would be grossly unjust to start entertaining the possi­bility
that another friend is secretly dishonest. Since man has free will, the choices
made by one individual provide no grounds for judging what another
individual will choose. (On free will, see Chapter 10.)
Despite the fact that being logical does not guarantee immunity
from error, one has to go by the available evidence, without engaging in
groundless doubts. First, there are conclusions about which no contrary
data could arise — e.g., that two and two is four, that dogs are animals,
or that the Earth rotates on its axis. But even for conclusions that are, in
principle, subject to correction by new data, such as your judgment of a
man’s moral character, the mere fact that, on rare occasions, new data will
reverse a prior certainty provides no grounds for doubt in any new and
different case. What is supported by conclusive evidence is certain, and this
certainty is unaffected by the fact that it is not impossible for a certainty to
be overthrown by new information.
The arbitrary assertion of possibilities is widespread today. The equivo-
cation between metaphysical and epistemic possibility — between can and
might be — is the stock in trade of skeptics in their attacks on certainty.
The skeptics’ argument here is from the fact of fallibility. Because you can
make a mistake, they argue, you always might be mistaken about anything.
But can does not imply might be. The capacity to make a mistake provides
no evidence that a given conclusion is mistaken, and thus gives no grounds
for doubting one’s conclusions — just as my capacity to whistle “Yankee
Doodle” gives no grounds for thinking I am whistling it now.
 T he Burden of P roof P rinciple 291

“Maybe” is not an epistemological free lunch. One has to have grounds

for entertaining a possibility, even the possibility that a given conclusion
is illogical (when one has conclusive evidence that it is logical) or that it will
be refuted by facts that will be discovered in the future.
Along with crediting the arbitrary claim, “Maybe you’ve made a mistake,”
most people assume that an idea begins with the status “possibly true,” as its
default state. They believe that an idea has cognitive standing until and unless
someone refutes it. For example, in the 1994 O. J. Simpson murder trial, the
defense made many arbitrary claims —about an alleged police conspiracy,
cuts on Simpson’s hands being due to “golf injuries,” and so on. The premise
was that unless the prosecution could somehow refute these baseless claims,
“reasonable doubt” existed as to Simpson’s guilt. But in fact, such doubts were
flagrantly unreasonable.
There is no difference in principle between those unreasonable doubts and
more bizarre assertions of the arbitrary, such as those of astrology or the
Book of Revelation.
Throughout the history of thought, it has been common for philosophers
to take arbitrary claims as their starting points. For instance, Spinoza begins
his entire system with this as his axiom (Proposition I): “By cause of itself
I understand that whose essence involves existence, or that whose nature
cannot be conceived unless existing.” [Spinoza, 41]
This pseudo-axiom treats as irreducible primaries such notions as “cause,”
“essence,” and “conceiving.” These are proper concepts — if understood in
terms of prior concepts (particularly, “existence,” “identity,” “action,” and
“consciousness”). But just plopping them into a “Proposition I” deprives them
of any base in reality, rendering them arbitrary.
Kant launches his Critique of Pure Reason with the entirely arbitrary
distinction between “analytic” and “synthetic” judgments, and between
“a priori” and “a posteriori” knowledge — distinctions without rational basis.154
The second paragraph of his “Introduction” arbitrarily announces:

But though all our knowledge begins with experience, it does not
follow that it all arises out of experience. For it may well be [!]
that even our empirical knowledge is made up of what we receive
through impressions and of what our own faculty of knowledge
(sensible impressions serving merely as the occasion) supplies
from itself.

154 See Chapter 5.

 H OW W E K N OW • 8 : P roof and C ertaint y 292

“May well be” — based on what evidence? (In fact, the idea of knowledge
whose source is purely internal denies the very need of evidence.)
To parody the grounding of philosophic systems in arbitrary claims,
Rand projected a philosopher announcing: since man has only two eyes,
he can see only two things. (This then, in her parody, gives rise to two
subsequent schools: one maintains that only two things actually exist,
the other acknowledges that many things exist but asserts that man actually
has more than two eyes.) [Peikoff 2013, 73]
The first step in judging the validity of an idea is to identify its source:
is it based on fact or fantasy? If the idea is evidence-based, one can check
the interpretation placed on that evidence; but that which is asserted
arbitrarily — proceeding from “what if?” or “why not?” or “it may well
be” — offers no evidence to be interpreted. Such ideas are not in the realm
of logic but of make-believe.

What is Objectivity?
When a mental product results from the deliberate application of logic
to evidence, it has a unique status: it is objective. Ayn Rand provides a new
understanding of what objectivity is and requires.
In its metaphysical usage, “objective” simply means: existing indepen-
dently of consciousness. But now we are concerned with “objective” in
its epistemological meaning — i.e., the objectivity of a mental process
or p­ roduct. What is it for a mental process or product to be “objective”?
To properly understand the concept of “objectivity,” we must pose Rand’s
key question: what facts of reality give rise to the need for such a concept?
Among many such facts, three stand out: 1) the primacy of existence:
existence exists independently of consciousness, and consciousness is aware-
ness of that which exists; 2) the law of identity: contradictions do not exist
in reality, so, “non-contradictory identification” (logic) is the method of
acquiring knowledge of reality; 3) volition: man can choose to regulate his
own cognitive activities, applying his knowledge of logic to guide them and
to check his conclusions.
Do we need a concept that distinguishes between those cognitive activi-
ties that are deliberately guided by logic and those that are not? That is,
do we need the concept of objectivity? Yes, because processes guided by logic
are fundamentally different from those that are not. Only if one knows and
 W hat is O bjectivit y ? 293

consciously applies logic can one warrantedly claim to have knowledge as

opposed to mere belief. Only by reference to logic can one have a standard
of certainty.
One can, of course, be logical before one knows that that is what one
is doing. A child does not need to study logic in order to go from “Lassie is
a dog” and “Dogs chew bones” to “Lassie chews bones.” Adapting a state-
ment from John Locke, men didn’t have to wait for Aristotle to become
logical. Although that is true, men did in fact have to wait for Aristotle in
order to know, beyond the most rudimentary level, what logic consists of and
to therefore be able to make their thinking objective. Before having a decent
understanding of how to validate an idea, men could say no more than,
“My idea just seems right to me.”
Even when one can give reasons to support his beliefs, objectivity about
one’s conclusions requires something more: knowing how to distinguish
valid from fallacious reasoning, what constitutes evidence, what kind or
quantity of reasons constitutes proof.
To be objective involves standing back from one’s mental processes,
viewing them as if they were external, making the form and method of
one’s thinking into the object of awareness (thus the term “objective”155).
An objective thought-process is one that is observed, inspected, judged — just
as are the objects of perceptual awareness. On the perceptual level, one
makes, say, a flower into an object of perception by turning one’s gaze
to it; on the conceptual level, one makes a thought-process into an object
of judgment by turning one’s attention to it. By adopting this self-conscious
standpoint, one can subject his thought to critical examination, judging
its validity by the canons of logic. Some of this can be done in “real time,”
while the thought is occurring, but often it requires retrospection.
Unless one uses logic to critically evaluate one’s own thinking, one is
absorbed by and caught up in the content of the thought, in which case
it is precisely objectivity that one has not yet achieved. Prior to making the
thought into the object, one is thinking “through” the thought; in a sense, one
is the thinking process. By reflecting on the thought, one becomes its judge,
not its advocate. One establishes a certain separation between self and the
thinking, making it easier to avoid distorting factors, such as a too narrow
viewpoint, or psychological defense-mechanisms, such as rationalization.

155 “Objective” also connotes: object-based — i.e., based on evidence not on the wishes of the
subject, the person holding the idea.
 H OW W E K N OW • 8 : P roof and C ertaint y 294

Consider what so often happens when one argues over abstract issues,
such as politics. After one has presented an extended argument that one
takes to be con­clusive, the other party is neither persuaded nor silenced.
Rather, the other party replies with a stream of objections and counter-­
arguments. After several such back-and-forths, no resolution is in sight.
In order to know who made a logical case, one needs to know what making
a logical case consists of. In order to get beyond the mere feeling, “I made
a good case, he did not,” one needs to know how to analyze, criticize, and
check an argument — one’s own even more than that of a disagreeing party.
In order to have an objective assessment of the reasoning offered in support
of an idea, one must know logic explicitly. The explicit knowledge of logic is
what Aristotle supplied.
Just as being “scientific” in one’s thinking requires having some explicit
knowledge of the scientific method, so being “objective” in one’s thinking
requires having some explicit knowledge of logic. In contrast, to be “rational”
simply means to exercise the faculty of reason, something that all men, even
savages and young children, can do when they form concepts, use them to
make identifications, and draw logical inferences. But to achieve objectivity
is to go beyond that; objectivity enters when one consciously applies the rules
of logic to guide and check one’s cognitive processes. That is a higher-order
phenomenon, and to reach that level requires an explicit knowledge of logic.
For reasoning of any complexity, to determine what one knows, as
opposed to what one merely believes or assumes, one must check the reason-
ing against the standards and methods of logic.
(This does not mean, absurdly, that to be objective a person has to
be continually saying to himself any such things as: “I am now using
a syllogism of type AAA-1,” or “I am now testing my definition against rule 2.”
For one thing, the proper methodology becomes second-nature — or would,
granted a proper education. But to be objective, one must be able, if the need
arises, to state at least roughly the reasons for his conclusion, the meaning of
his terms, and what makes his reasoning valid and complete.)
Since logic, for Objectivism, is based on the identity of both the object and
man’s cognitive equipment, Peikoff defines “objective” as follows:

To be “objective” in one’s conceptual activities is volitionally to

adhere to reality by following certain rules of method, a method
based on facts and appropriate to man’s form of cognition.
[OPAR, 117]
 W hat is O bjectivit y ? 295

His shorter formulation describes objectivity as: “Volitional adherence to

reality by the method of logic.” [OPAR, 116]
This theory is a deepening of the colloquial meaning of “objective.”
Colloquially, to be objective is to be unbiased, to go by the facts, to place
no consideration above determining the truth. The opposite is being
subjective, in which case one’s conclusions are produced by internal
factors — wishes, hopes, fears, and whatever happens to pass through
one’s mind. The means of ensuring that one is proceeding in a reality-
based manner is logic. And, to apply the Objectivist understanding of logic,
this includes (once learned) rules derived from the nature of man’s concep-
tual faculty (see Chapter 7).
In our formation and use of concepts, nothing forces us to proceed logically
— i.e., to hold the full context and retain the links back to reality by
following a hierarchical progression. The rules of logic are not pre-
programmed into the brain; they must be discovered. Even once learned,
the rules of logic do not apply themselves: one needs to institute them
by an active process that is not automatic but volitional. In every moment
and issue, one faces the alternative of putting forth the effort to engage in
logical processing of observed facts, integrating one’s ideas into the full
context and checking for contradictions — or passively riding with anything
less than that, as in going by authority, urges, guesses, pretenses, or faith.
An objective process has truth as its goal and its normal result, but
being objective does not guarantee that one’s conclusion will be true.
Truth, like knowledge, has both a metaphysical and epistemological com-
ponent: a true conclusion must both state a fact and reflect awareness
of that fact.
“Truth” and “knowledge” are “win-words” — i.e., terms that apply to cases of
cognitive success. In contrast, “objective” denotes the nature of the process
(and the status of its products), whether or not the process succeeds in
reaching awareness of fact. And objectivity, like certainty, is compatible with
being in error; but objectivity is one’s best protection against such error.
Objectivity is a deliberate, honest, truth-seeking, methodical process
— as opposed to surrendering control of one’s mind to emotional urges
and random associations (or of actively evading facts available to one).
Derivatively, a mental product, such as a conclusion or theory, has the
status of “objective” when it has been reached by and checked by an objective
process, a process of deliberately subjecting one’s thought to the rules of logic.
 H OW W E K N OW • 8 : P roof and C ertaint y 296

A Trichotomy in the History of Thought

The concept of “objective” opens the way to a new and deeper understanding
of the history of thought, including the history of theories of how we know.
Traditionally, theories of knowledge have oscillated between two poles of
a false alternative: either man’s knowledge is gained passively and automati-
cally, or his knowledge is due to his own “interpretation” and is not a valid
grasp of fact. Rand terms these false alternatives intrinsicism and subjectivism.
As against both, she offers a third possibility: objectivity.
Thus, there is not just a dichotomy but a trichotomy in regard to theories
of mental processes and their products. The revolutionary nature of
Rand’s philosophy is to define, in area after area, a position that is neither
dogmatic nor skeptic, neither Platonist nor Humean, neither intrinsic nor
subjective, but objective. That is why she named her philosophy “Objectivism.”
The intrinsic-subjective-objective trichotomy offers a new and fundamental
way to classify theories in philosophy. Moreover, all of the humanities and
social sciences manifest intrinsic, subjective and objective schools of thought.
For an application of this trichotomy to issues throughout philosophy, see
Objectivism: The Philosophy of Ayn Rand. Here, I will illustrate it only in
regard to theories of concepts, the basic issue of epistemology.

The Objective Status of Concepts

The recognition that concepts, if properly formed, are objective contrasts
with the two false theories that have dominated the history of e­ pistemology.
Looking again at the schools discussed in Chapter 3, the Realist school
holds that a concept refers to something non-specific that exists intrin-
sically in the external world, whether as a separate Platonic “Form,”
an Aristotelian “essence” inside concretes, or a Lockean non-specific attribute
(e.g., “whiteness”). The Realist theory represents intrinsicism because it
holds that a concept is the mental representation of this intrinsically existing
“universal.” Concept-formation is treated as something akin to a revelation:
the passive absorption of the pre-existing, pre-packaged universal that exists
“out there,” waiting to be taken in by the mind. Thus, the intrinsic theory
assimilates the conceptual to the perceptual. We just “see” abstract truth with
“the eye of the intellect.”
Describing the intrinsicism of Platonic Realism, Peikoff writes: “In the end,
[Plato] thinks, the mind need merely remain motionless, passive, r­ eceptive,
 A T richotom y in the H istory of T hought 297

and the light of truth will automatically stream in, taking the form of
a synoptic and ineffable intuition.” [OPAR, 142]
Later intrinsicist theories of concepts are less overtly mystical, but retain
the idea that concepts are a percept-like vision of an intrinsic universal that
lies within concretes. For Aristotle and Locke, the process of abstraction is
one of subtracting away from a given concrete all that is specific, after which
the intrinsic “universal” is there to be “seen” in that single concrete.
The intrinsicist theory of concepts results from the failure to recognize
the role of the identity of conceptual consciousness, including its volitional
nature. The premise of intrinsicism is that for concepts to be valid, the mind
must passively mirror something in the external world.
In reaction to the intrinsicism of the Realists, the Nominalists advance
the equally false idea that concepts are subjective. The Nominalist theory
recognizes that, apart from man’s mind, everything that exists is specific
through and through; no non-specific characteristic or “universal” exists.
There is thus nothing in reality for the mind to passively mirror. But from
that the Nominalists conclude that concepts are subjective inventions of
man’s consciousness. Concepts are regarded as invalid, as nothing more than
arbitrary social conventions. Since “nature doesn’t tell us” what concepts to
form — i.e., since concepts are not intrinsic — they are subjective, according
to the Nominalists.
Rand sums up and provides the overlooked alternative:

The extreme realist (Platonist) and the moderate realist

(Aristotelian) schools of thought regard the referents of ­concepts
as intrinsic, i.e., as “universals” inherent in things (either as arche-
types or as metaphysical essences), as special existents un­related
to man’s consciousness — to be perceived by man directly,
like any other kind of concrete existents, but perceived by some
non-sensory or extra-sensory means.
The nominalist and the conceptualist schools regard concepts
as subjective, i.e., as products of man’s consciousness, unrelated to
the facts of reality, as mere “names” or notions arbitrarily assigned
to arbitrary groupings of concretes on the ground of vague,
inexplicable resemblances. . . .
None of these schools regards concepts as objective, i.e., as nei-
ther revealed nor invented, but as produced by man’s conscious-
ness in accordance with the facts of reality, as mental integrations
 H OW W E K N OW • 8 : P roof and C ertaint y 298

of factual data computed by man — as the products of a cognitive

method of classification whose processes must be performed by
man, but whose content is dictated by reality. [ITOE, 53–54]

Objective knowledge is a mental product that one knows one made

in the right way. Objectivity, i.e., being self-consciously logical, comes down
to integration: integrating “downward” to perception (i.e., reduction) and
integrating generally, in all other “directions.”
Thus, to state the point negatively, objectivity means: shun the unintegrable.
Do not pretend that what cannot be integrated into the full context of one’s
knowledge is nevertheless cognitive.
An idea may be unintegrable for either of two reasons: 1) it involves a contra-
diction, or 2) it is arbitrary — i.e., there is no means to integrate it. Being non-
objective means either holding contradictions or entertaining the arbitrary.
The first step toward objectivity is the rejection of the arbitrary — i.e., the
recognition that claims to knowledge have to be justified, that the mere
presence of an idea in one’s mind is no basis for assenting to it or treating
it as cognition rather than fantasy. The need for objective validation applies
not just to propositions and theories but also to their root: concepts.

Objectivity begins with the realization that man (including his

every attribute and faculty, including his consciousness) is an
entity of a specific nature who must act accordingly; that there
is no escape from the law of identity, neither in the universe with
which he deals nor in the working of his own consciousness,
and if he is to acquire knowledge of the first, he must discover
the proper method of using the second; that there is no room for
the arbitrary in any activity of man, least of all in his method of
cognition — and just as he has learned to be guided by objective
criteria in making his physical tools, so he must be guided by
objective criteria in forming his tools of cognition: his concepts.
Just as man’s physical existence was liberated when he grasped
the principle that “nature, to be commanded, must be obeyed,”
so his consciousness will be liberated when he grasps that nature,
to be apprehended, must be obeyed — that the rules of cognition
must be derived from the nature of existence and the nature,
the identity, of his cognitive faculty. [ITOE, 82]
 PRINCIPLES
9
A theme running throughout this book is that knowledge
is not an end in itself, but a means of acting successfully in the world.
At the same time, however, I have warned against Pragmatism, a philos­
ophy holding that practical success requires rejecting absolutes, certainty,
and, above all, principles.
Today, to describe some­one as a “pragmatist” is considered to be paying
him a compliment. Anyone who adheres to principles is attacked as being an
“ideologue.” But, in fact, principles offer the only guide to practical success.
Consider one striking data-point: the fate of Richard Nixon. Nixon was, by
general acclaim, a virtuoso of pragmatism. The lesson of his downfall is clear:
nothing is as impractical as the attempt to function without principles.156
A principle is a fundamental generalization. To understand principles, and
their practical potency, one must understand the nature of fundamentality.

156 The Washington Post, in a 25-year retrospective on Watergate, discusses the transcripts
of secret tapes President Nixon made of his meetings, which reveal his unconcern
with moral principles: “The transcript reveals that Mr. Nixon, on his own initiative,
discussed accommodating blackmail demands on at least a half-dozen occasions during
the meeting without once suggesting that paying the men for their silence would be wrong. ”
(www.washingtonpost.com/wp-srv/national/longterm/watergate/articles/050174–2.htm)
 H OW W E K N OW • 9 : P rinciples 300

Fundamentality
Fundamentality pertains to a certain kind of hierarchical order. In addition
to the hierarchy of knowledge, there are other kinds of hierarchical
order — for instance, there is the hierarchy of composition (part-whole):
subatomic particles, atoms, molecules, macroscopic entities, astronomic
entities (such as galaxies).
Fundamentality refers to causal sequences. For example, the military
chain of command refers to who gives orders to whom, and the Commander
in Chief is the fundamental of the military hierarchy.
In contrast, though the ground floor of a building causally supports all
the higher floors, the causal sequence is too simple for the ground floor
to be called the building’s fundamental. Fundamentality concerns causal
sequences that have a branching, tree-like structure of causation, from trunk,
to major divides, to large limbs, to smaller limbs, to twigs.

Representation
of the relation of
a fundamental
to its derivatives.
(Not captured in this
diagram is the fact
that the derivatives
are usually outgrowths
or versions of the Fundamental
fundamental.)

The existence of this kind of ramified set of relation­ships, stemming

from one root cause, is the fact that gives rise to the need for the concept of
a “fundamental.” As a preliminary definition, a “fundamental” is a causal factor
on which a multi-level, branching series of effects depends. The de­pen­d­ency
here is causal: the fundamental is a necessary condition — a sine qua non —
of the derivatives’ occurrence.
 F undamentalit y 301

(Sometimes the same overall type of result can be produced by a num-

ber of different causes. For instance, when a rock is dropped into a pond, it
causes a series of ripples to spread across the pond, lapping upon the distant
shores, making a faint sound. Similar results would have been produced by
dropping a bottle into the pond. But the series of consequences that did in
fact occur depended upon the dropped rock; that was the root cause of those
consequences. It is irrelevant that a different cause would have produced
a different “tree” of consequences.157)
A “family tree” nicely illustrates the kind of causal relationship a funda­mental
bears to its derivatives. The founding patriarch and matriarch are the funda­
mentals of the line of descent in that their union was a necessary condition of
that whole line. Although the founders brought into existence only their own
children, not their grandchildren or any later generations, the founders are
the fundamentals: but for their union, there would be no ancestral line at all.
Similarly, the division of labor is a fundamental of all economic-finan-
cial phenomena; without the division of labor, there can be no supply and
demand, no credit, banking, stock exchanges, arbitraging, etc. It is only by
reference to the division of labor that we can understand all these diverse
activities, by integrating them as forms that the division of labor takes.
The existence of a domain of causally interrelated phenomena, existing
on different levels, is one fact that gives rise to the need for the concept of
“fundamental.” Thus, a fuller definition would be: a “fundamental” is a factor
necessary to the existence, nature, and explanation of a ramified set of items
in a given domain.
In the biological realm, the fundamental factor is natural selection. Natural
selection causes and explains the whole “tree of life.” It also causes and explains
the adaptedness of the structure of organisms on every level of taxonomy:
the structure of any particular species of flowering plant, of flowering plants
in general, and of the entire plant kingdom. Whether one is seeking to
explain the specific shape of the pistil in the rose, the fact that flowers have
bright colors, or the presence of chlorophyll in plants, one will find that the
contribution made to the plant’s survival is the causal factor that forms the
deepest part of that explanation. Note that natural selection is a necessary
factor: without it, evolution would not have occurred.

157 If one is considering a category of causal trees, not just the particular one caused by the
rock, one must state the fundamental in correspondingly general terms (e.g., “the impact
of a solid body”). The issue is: from what are the effects being differentiated?
 H OW W E K N OW • 9 : P rinciples 302

Note, however, that for the field of biology as a whole, natural selection
is a derivative, because it depends on the fact that life is conditional on
successful action and on the nature of reproduction. Something is funda-
mental only in relation to a specific domain, not per se. Natural selection is
a fundamental in relation to the evolution of organisms, but it is a derivative
in relation to the nature of life as such. Because organisms are metabolizing
systems, capable of reproduction in kind but with variation, there is a natural
selection of variations favorable to descendants.
(The only things that are fundamental simpliciter are the philosophic
axioms — primarily, the axioms of existence, identity, and consciousness — and
the basic essentials of epistemology. These state the necessary conditions
of all knowledge, including the knowledge of any domain of related items.
All other fundamentals qualify as such in relation to a delimited domain.)
A fundamental is a factor causing and explaining the items in a given
domain. More informally, a fundamental is both wide and deep — “wide”
in terms of explaining the ramified set of things in the domain, and “deep”
in identifying their root cause. One arrives at a fundamental when many
concrete phenomena can be traced back to the same overall cause.
For example, in a business: why is the boy in the mailroom doing what he
is doing? Because his boss assigned that kind of task to him. Why did his
boss assign that kind of task? Because of the direction supplied by his boss.
Following the chain back will lead one to the overarching direction supplied
by the CEO. The CEO has the same ultimate authority over the activities of the
salesmen, the legal staff, the accounting staff, and every employee of the firm,
including its executives. The CEO is the firm’s fundamental employee.
Or, the U.S. President, as Commander-in-Chief, has the fundamental role
in directing America’s military; his decisions underlie and explain the orders
given by the generals, which underlie and explain the orders given by the
colonels, and so on, down to the enlisted men.
In a given domain, a fundamental is the factor that integrates and explains
the existence and interrelations of all items within that domain. All the items
in that domain, on whatever level, are subordinate to the fundamental;
they are its consequences, implications, and/or variants.158

158 I say “consequences, implications, and/or variants” to cover different types of dependency.
Although all fundamentals identify first causes, in some cases the causality pertains
to the external world and in others to man’s cognition.
 F undamentalit y 303

What Identifying Fundamentals Accomplishes

Knowing fundamentals is a source of immense cognitive power. That power
results from the unit-economy provided by a knowledge of fundamentals.
Since a fundamental causes and is expressed in everything in the domain,
it is the one factor to be held in mind when dealing with anything in that
domain. For instance, in the military, every soldier’s actions are devoted
to carrying out the fundamental order given by the Commander-in-Chief
(an order that may reduce to a directive as simple as: “Defeat the enemy”).
Every military decision must be judged solely by its potential contribution
to that fundamental goal. (The addition of qualifications, such as “with
­minimum casualties to our soldiers” does not alter the fundamentality of the
goal.) Or, in a business, the fundamental goal is making a profit by follow-
ing the firm’s business model, and that fundamental sets the standard used
to judge every decision on every scale of operations. Losing sight of that
fundamental leads to decisions that may look good out of context, but are
harmful in the long run.
The cognitive power of fundamentals is dramatically illustrated in the
history of astronomy by the shift from the geocentric to the heliocentric
conception of the solar system. Since motion is relative, the geocentric model
is just as factually correct as the heliocentric model. If we take the Earth
as our frame of reference, the celestial bodies do move as described in the
geocentric system. However, the sun’s gravity is the fundamental cause of
all the relative motions, and this makes the heliocentric model objectively
superior. The baroque complexity of the Ptolemaic model, with its epicycles and
deferents, results from taking a non-fundamental (Earth) as the frame
of reference; the simplicity (unit-economy) of the heliocentric model stems
from taking the fundamental as its frame of reference.159
Organizing one’s knowledge on the basis of fundamentals permits a “crow-
friendly” condensation of the whole “tree” of phenomena. This not only
adds clarity, it also allows for automatizing what is most c­ ognitively fertile.

159 The inclusion of relativistic effects, such as the precession of the perihelion of Mercury,
does not change the point: a stubborn geocentrist could handle the precession by simply
adding further ad hoc devices, similar to “epicycles” and “deferents.” In reality, the motion
of A around B is the very same phenomenon as the motion of B around A — i.e., the fact
is that A and B are moving relative to each other.
 H OW W E K N OW • 9 : P rinciples 304

Automatization is essential to building new knowledge on old, because of,

once again, “the crow epistemology.” Rand explains the process.

. . . all learning involves a process of automatizing, i.e., of first

acquiring knowledge by fully conscious, focused attention and
observation, then of establishing mental connections which
make that knowledge automatic (instantly available as a context),
thus freeing man’s mind to pursue further, more complex
knowledge. [ITOE, 65]

By repeatedly tracing the connections of the different branches down

to the trunk (and, conversely, seeing how the trunk ramifies into branches)
one can automatize the whole tree-structure — not as a random pattern but
as theme and variations. For example, the fundamental in chess is the goal
of killing the opponent’s king. Without having that goal as the automatized,
taken-for-granted context, the moves and arrangements of pieces make
no sense. A chess expert, who has automatized many derivatives in relation
to that supreme goal, can look at the board of someone else’s game midway
through it and see at a glance what is going on. Or, in music, automatiz-
ing the relation of each note of the scale to the scale’s keynote (the “tonic”)
is what allows the music to be intelligible. Music is key-relative and the
­keynote is the fundamental of the notes in that key.160
The power of simple, automatizable fundamentals is what accounts for the
decimal system having replaced Roman numerals. The decimal system allows
for automatizing one fundamental: the “place” of the digit; that governs what
power of 10 the digit represents — ones, tens, hundreds, etc. Having long ago
automatized this fundamental, you now have no difficulty understanding the
number 644,012 even though it is very unlikely that you have ever encountered
it before. But in grasping Roman numerals, we must alternate among three
different procedures. First, we increment by conjoining another of the
same symbol: I to II to III; then, to denote the next number, we switch to
subtracting one from the symbol to the right: IV; then, to continue incre-
menting, we drop the subtraction and use just the new symbol: V; after that,
we revert to the first method: VI. And all three procedures are often combined:
XCVII (97), making for a far more cumbersome system, than the decimal one.

160 In the C major scale, the tonic is C, and one automatically perceives all notes in relation to C.
For instance, the notes E and G will be heard in their specific harmonic relations to C: as
the pleasingly consonant 3rd and 5th scale degrees. But C# will have a radically different,
very discordant, musical (and emotional) meaning.
 P rinciples as F undamental G enerali z ations 305

Principles as Fundamental Generalizations

Fundamentality can concern a particular cause in relation to a particular
tree of effects, as with our sun and the particular planets of our solar system.
But the real cognitive value to be reaped comes from the generalization to
a type of cause and a type of tree of effects, as with: “Any star is the basic
causal factor controlling the orbits of its planets.” When such generalizations
reach a certain level of scope, they qualify as principles.
A “principle” is a fundamental generalization that serves as a standard
of judgment in a given domain.
We need standards to guide our thinking, including the thinking devoted
to deciding what to do existentially. E.g., the law of universal gravitation is
a principle guiding the thinking of the physicist, and the principle of honesty
guides (or ought to guide) each man’s thinking about his own conduct.
Principles are differentiated from other action-guiding generalizations,
such as rules of thumb or statements of “good policy.” A principle, by iden-
tifying a fundamental cause, informs us of requirements that are absolute:
one cannot have effects without their causes. In contrast, a rule of thumb
has no such absolutism. “Don’t buy stocks upon the release of good news”
(because the news is probably already reflected in the stock’s price) may be
sound advice in the main, but sometimes stocks do go up on good news.
“Never draw to an inside straight” is a correct statement of the odds in poker,
but the less usual does occasionally happen. Unlike probabilistic generaliza-
tions, a principle can never be flouted with impunity. In this regard, “Honesty
is a virtue” states a principle, whereas “Honesty is the best policy” merely
makes a probabilistic recommendation.
Principles are also distinguished from those generalizations that, even
though absolute, are not wide and deep enough to qualify as principles.
“Don’t shoplift” is absolute, not probabilistic, but does not rise to the level
of a principle — the principle here being: “Respect the rights of others.”
The same depth-requirement applies to principles that guide cognition
rather than conduct. In mathematics, the commutative law (a + b = b + a)
is a principle, but “The sum of two even numbers is even,” though
exceptionlessly true, does not have the requisite scope. Or, in the field
of grammar, “Subject and verb must agree in number” is a principle,
whereas “The word ‘none’ takes a singular verb” is too narrow to qualify for
that appellation.
One must keep in mind that the issue of fundamentality pertains to
a specified domain (and a given context of knowledge). Honesty, for
 H OW W E K N OW • 9 : P rinciples 306

instance, is the fundamental issue at stake in a wide range of situations,

but honesty itself depends upon deeper philosophical generalizations, such
as the law of causality; causality is a fundamental relative to a much wider
domain. The same is true of a (literal) tree: a major branch is the fundamental
of all the sub-branches, twigs, and leaves, but that branch is itself one of several
outgrowths of the trunk of the tree.

The Need for Principles

The need for principles is psycho-epistemological: principles provide a unit-
economical, long-range view of consequences. Because principles deal in
fundamentals, principles integrate and condense all the derivatives — i.e., all
the consequences, implications, and variants.
Life is complex; principles allow us to simplify it. This is a view of
principles very different from the traditional one. Principles have been
looked upon as externally imposed rules aimed at keeping us from doing
what we actually want to do.
The popular, but mistaken, view of moral principles is the paradigm
case in this respect. Moral principles are widely regarded as being a series of
Thou Shalt Nots, commanded from on high. The same shadow of grim duty
darkens other fields — e.g., the principles of grammar are typically viewed
as arbitrary prohibitions, rather than as enablers of clear, precise expression.
Even the principles of logic have been viewed by some as galling limitations
on their “freedom.”161 But in a proper understanding, principles are a means
of cognition — which is why they serve as standards of judgment to guide
cognitive or existential action.
By spotlighting root causes, principles make one aware of a long train
of consequences — not just the immediate effects, but the sum across a life-
time. Principles are thus indispensable for acting long-range. In this regard,
Peikoff makes an apt contrast of man with animals:

An animal cannot grasp or deal with the total of its lifespan and
does not need to do so. . . . Man can and must know not merely
tomorrow’s requirements or this season’s, but every identifi-

161 Dostoyevsky’s “underground man” famously vents his rage at “two plus two makes four”
and “a wall is a wall.”
 T he N eed for P rinciples 307

able factor that affects his survival. He can assess not merely the
­proximate, but also the remote consequences of his choices. It is
not enough for him to consider the chance of a toothache next
week; he also needs to know whether he is courting bankruptcy
next month, an anxiety attack next year, an invasion of human
predators in the next decade, or a nuclear holocaust in the next
generation. [OPAR, 216–217]

Thus, one needs principles because they provide the overview, the road-
map outlining the kind of consequences that follow from choosing one
way or another. Principles identify cause-and-effect relationships. Acting
in defiance of a valid principle means pretending that there can be causes
without effects or effects without their causes. Both attempts are inherently
self-defeating.

Principles as Cognitive Bridges

Since principles are generalizations, they are reached by a process of induction.
And since principles are high-level generalizations, they are formed from
generalizations that are one level less abstract, rather than from cognitive
ground-zero.
For example, in physics, the principle that all matter attracts other
matter is a very wide generalization integrating narrower ones, such as
generaliz­ations pertaining to the rate at which falling bodies accelerate,
Kepler’s laws of planetary motion, and the periodicity of the tides. (Finally, one
“spirals back”: the seemingly unrelated phenomena that one integrated to
grasp the principle come to be seen as manifestations of that principle.)
To reach a principle, as we have seen, is to grasp a general type of root
cause, a factor explaining a whole “tree” of derivatives — as the principle of
identity (A is A) underlies and explains the rules of valid deduction, valid
induction, proper definition. Likewise, the principle of individual rights
identifies a fundamental requirement of moral dealings with others, whether
on a personal, social, or political level.
One grasps a principle by a process of abstraction, just as one does in
concept-formation. And abstraction, we have seen, is not subtraction but
inter­relation. Neither concepts nor principles are formed by “intuition,”
“insight,” or any process other than observing the similarities and differences
among concretes.
 H OW W E K N OW • 9 : P rinciples 308

Principles offer a conceptual approach to cognition and conduct. Like con-

cepts, principles are integrations; by integrating narrower generalizations,
principles provide unit-economy. The condensed grasp of facts achieved
by a principle gives us a mountain-top view of the particulars with which
we must deal, cognitively and existentially. Principles enable us to apply what
we have learned about observed cases to those as yet unobserved.
For example, to oversimplify somewhat, the principle “One must be
honest” codifies, in abstract terms, the lessons learned from observing
and reflecting on the nature of particular cases of honesty vs. dishonesty.
The oversimplification is contained in the phrase “the nature of particular
cases.” To grasp the principle of honesty, one must not only see that the
particular acts of dishonesty led to bad consequences in those circumstances,
one must also grasp what in the nature of those acts caused the results to be
bad and why some form of negative outcome is inevitable. In other words,
the principle must be a valid induction — i.e., one that identifies a causal
connection,
A not justB a statistical correlation.
162
C
Generalizations serve as cognitive bridges leading from cases that have
been studied to those that have not been. As with a literal bridge, the value of
a principle is that it allows one to cross over — i.e., to take knowledge gained
by investigating the concretes from which one generalized and apply it to
concretes not yet investigated.
A B C
The diagram below illustrates the process of moving in cognition from
observed concretes A, B, and C to identify a principle, then moving back
“down” to apply the principle to a new concrete, S.
A D B C

principle

A B C S

concretes

162 Dishonesty is destructive because it is the attempt to fake what cannot be faked:
the actual facts. Moreover, reality is an interconnected whole, not a mosaic of isolated
compartments, so the faking cannot be limited to one domain but, to be maintained in
the face of the growing disparity with reality, must grow ever wider — until the whole
system of lies collapses. This is the lesson of the Bernie Madoff fraud.
 T he N eed for P rinciples 309

Knowledge is not an end in itself. The value of all knowledge, whether

concrete, mid-level, or fundamental, lies in its application — which, ultimately,
means its use in guiding action. Whether one induces “Arsenic is poisonous” or
“All bodies attract other bodies in inverse proportion to the square of the
distance between them,” the value of knowing the generalization lies in its
application — e.g., to avoid dying from arsenic-poisoning or to put satellites
into orbit. The use of principles consists in deductive applications of the
induced generalizations.163
Principles, since they identify fundamentals, are more cognitively powerful
than lesser generalizations: they have a wider scope and isolate a more potent,
longer-range causal factor. For instance, the law of gravi­ty is more powerful,
in this double sense, than “Objects fall downwards.”
Contra Plato, only concretes are real; abstractions, including principles,
are merely man’s method of understanding and dealing with concretes.
A principle, if valid, is an invaluable tool for learning about the concretes
and their long-range consequences; but the principle is only a tool,
not a substitute for study of the concretes with all their particular character­
istics. Principles tell us where to look — what considerations to hold in
mind — when dealing with concretes, but we still have to look.
Even thinkers who reject Platonism may inadvertently adopt a Platonic
approach by converting principles into formulas, which are then imposed
mechanically upon unexamined concretes.164 The proper application of
principles uses them as an abstract frame of reference to guide a diagnosis
of concretes, based upon sensitive attention to their particular character-
istics. A physician observing that a patient is coughing would not reason:
“Tuberculosis causes coughing, so this patient has tuberculosis,” but would
look at all the patient’s symptoms, his medical history, the prevalence
of tuberculosis in the environment, and many other factors, including test
results. In the same way, the proper application of principles requires care-
ful attention to all the facts about the concretes, some of which may modify
the way the principle applies or, in unusual cases, even indicate that the
concrete has special features that place it outside the principle’s domain.

163 Analogies function as temporary bridges, based on grasping a relevant, but isolated,
similarity between some concretes that are well understood and others that are made
more intelligible by reference to that similarity. Analogizing principles to a bridge is itself
such a case. The “bridge” is temporary in that the isolated similarity does not warrant
forming a concept to permanently integrate the analogized items, because they are too
different, overall.
164 In Objectivist terminology, this represents “psycho-epistemological Rationalism.”
 H OW W E K N OW • 9 : P rinciples 310

This procedure also provides a check that one has induced a valid principle,
and has defined its proper sphere of application.
An extreme example of the Platonic, Rationalistic misuse of principles
is Kant’s notorious position that the principle of honesty requires telling
the truth to an armed homicidal maniac demanding to know where one’s
children are — an absurd and outrageous example of context-dropping.165

Principles as Simplifiers
Since one learns principles by observing concretes (a, b, and c in the pre­ceding
diagram), the question arises: why are principles necessary? Couldn’t we
just observe the new concrete(s) as we observed the original ones? Part of
the answer is that having a principle eliminates the need to continually re-
investigate the same issue; the principle is an invaluable shortcut. But there
is a deeper answer. Concretes differ in their complexity. Principles enable the
clear, simple cases to shed light on the obscure, difficult ones.
This is the answer to those philosophers who scorn principles as being
“tautologies” or “truisms.” These philosophers say, for instance, that the Law
of Non-Contradiction is “empty,” citing the fact that we gain no new informa-
tion from being told “It cannot be raining and not raining at the same place.”
But, in fact, holding in mind the principle of non-contradiction reminds one
to check for non-obvious inconsistencies; it directs one to work to integrate
every conclusion into the full context.
It is from simple cases like raining vs. non-raining that we draw the lesson:
check for non-obvious contradictions. E.g., there is a non-obvious contra-
diction in an economy having both full employment and a minimum-wage
law.166 Or, take the idea that man, as a “sinful” being, cannot avoid immoral-
ity; what is unavoidable is not subject to moral judgment: morality exists
to judge choices. Thus, the idea of unavoidable sin implies a contradiction.
A recent study on the sociology of scientific research found that from
half to two-thirds of experiments on the frontiers of knowledge do not
produce the expected results. The study also found that there is a tendency,
especially among junior researchers, simply to ignore the anomalous results.
As one distinguished biologist reported, graduate students and post-doctoral
researchers frequently dismiss anomalous results as being due to “phases of

165 “On a Supposed Right to Lie from Altruistic Motives.” [Kant 1949, 427]
166 Minimum wage laws mean that a man whose services are worth less to an employer than
the mandated minimum wage cannot be employed.
 P rinciples as C onte x tual 311

the moon.”167 In fact, there is much to be learned from anomalous results,

whether about how to avoid errors in procedure or about the existence of
previously unknown factors. Bringing to mind the principle “Every effect has
a cause” is an important reminder of the need for investigating the cause of
all results — expected or unexpected. Likewise, the principle that knowledge
is an integrated whole, not isolated bits, reminds the researcher to bring to
bear on the problem everything relevant, not just the details of one particular
run of one particular experiment.

Principles as Contextual
Like all conceptual knowledge, principles are contextual. The context of
a principle is the knowledge from which it derives and the conditions under
which it applies. For instance, the context of moral principles is the actions
of a volitional being. The fact that there are no moral principles for ants is not
any defect in ethics or moral principles.
Other issues of context are not as obvious (a point that itself illustrates
the fact that principles clarify non-obvious cases by connecting them to
more obvious ones). For instance, in ethics there is the oft-heard challenge,
“What would ethics prescribe for two men in a lifeboat that can hold only one?”
This sort of pretend-philosophizing drops the context of ethics, which is:
normal conditions of existence, in which a long-range course of action is
possible, and in which the survival of one does not threaten that of another.168
Even arithmetic has been attacked by using the same context-dropping
approach. In objection to the principle that 1 + 1 = 2, it is said that one lion
plus one lamb equals one well-fed lion, and that one drop of water plus one
drop of water equals one (coalesced) drop of water. But the context of 1 + 1 = 2
is not physical combination but mental comparison. The physical results of
an interaction are outside the context of arithmetic.
Often an invalid, concrete-bound formulation of a principle is exploited
to attack it. For instance, one philosopher claimed that the existence of air-
planes proves that gravity is not a principle. But the actual principle here
is not “Heavy things fall”; it is “All bodies are attracted by other bodies.”
And an airplane is, of course, subject to that force of attraction — which is
why the plane needs the power of its engines in order to fly.

167 Judith Berliner, UCLA, personal communication.

168 On such “lifeboat cases,” see Rand, “The Ethics of Emergencies,” VOS, ch. 2.
 H OW W E K N OW • 9 : P rinciples 312

The fact that principles apply only within a delimited context does not make
them dispensable or second-rate — except by a Platonic standard of evaluation,
proceeding from the arbitrary idea that knowledge concerns relations
between Forms existing in another dimension. A principle is an abstraction
derived from observing and integrating facts, and its sphere of proper appli-
cation is therefore determined by reference to the context of knowledge
on which it is based.

Principles as Absolutes
Within their proper context, principles are absolutes. This follows from the
nature of a principle: the law of causality has no exceptions. Thus, a valid
principle can never be violated with impunity.
A principle allows for the operation of countervailing factors (as the force
of gravity can be balanced by the force of an airplane’s lift), but there is no
such thing as a cancellation of the law of causality. Within their context,
principles hold exceptionlessly.
Why, then, do people so often violate their own principles? In some
cases, the explanation is that the principle is false. In that case, the principle
tells one to follow a course of action that is demonstrably self-destructive.
In other cases, the problem is one’s failure to grasp and hold a principle in
an authentic, first-hand manner. If a man’s “principles” are merely his inter-
nalization of the blindly accepted assertions of his parents, his teachers,
and his neighbors, then he holds arbitrary, undigested rules, not actual
principles — i.e., not road-maps to the achievement of his own values.
Finally, man is not automatically rational: one does not automatically
activate his abstract, conceptual knowledge. It takes no effort to be aware
of short-range consequences; they are glaringly evident, here and now.
But principles are highly abstract; it takes some conceptual effort even to
realize that a principle is at stake, and more effort is required to apply the
principle to a new, concrete case — and to hold onto the resulting knowledge
under fire. So, in some cases, simple mental lethargy explains why people act
against the better knowledge their principles make available to them.
A frequent rationalization used to justify violating a principle is that
it will be violated “just this once.” But the absolutism of principles cannot
be escaped. First, if the principle is true, then it is true. That means one
cannot succeed — whether one is attempting to acquire knowledge without
­obeying the principles required to do so, or trying to reach an existential
 P rinciples as A bsolutes 313

goal without ­obeying the causal principles that identify the necessary means.
Principles provide the map that identifies the location of a goal that one
cannot directly see. If the map shows that the goal lies to the north, whatever
the temporary attraction of heading south, doing so takes one farther from
one’s goal. Even if one seeks a “compromise” by heading northwest, one will
miss the goal.
Violating a principle does not work; sooner or later, it results in failure.
Failure exacts existential and psychological costs: wasted time and resources,
and weakened self-confidence and self-respect.
Consider a second, and deeper, penalty for violating a principle. There is
a logic to principles, and a logic to what happens when one acts against them.
In acting against a principle, one faces consequences not just in regard to
the case at hand, one is also implicitly endorsing an opposite principle and
beginning to establish it in one’s soul. For instance, if one tells a “white lie”
to spare a friend’s feelings, one is endorsing the (false) principle, “Avoiding
negative feelings is more important than facing reality.” One is also endorsing
certain principles about the nature of friendship, such as that it is based not
on mutual esteem but on pity and shared weaknesses.
To the extent that one is rational, one’s principles define what one judges
to be rationally necessary. Thus, flouting a principle jettisons both reason
and causal necessity. This amounts to endorsing two wrong wider ­principles:
metaphysically, the notion that reality is a fluid, non-absolute realm;
epistemologically, that reason is not an absolute and that one’s rational
judgment is dispensable.
The latter, epistemological point is not just a theoretical implication.
The unadmitted meaning of violating a principle is the dethroning of
reason. One’s operating premise is: “I’ll go by reason — unless I don’t feel like
it.” The meaning of this is a new (and false) principle: feelings trump reason.
It doesn’t matter what one advances as the factor before which reason must
retreat — a hunch, social mores, short-range advantage, God’s will — that
­factor cannot be something endorsed by reason. The idea of a rational limi-
tation on reason is a contradiction in terms: if reason endorses the use of
some consideration, then it is not beyond reason but part of reason; if reason
does not endorse it (which is always the actual case), then it is against reason.
And as explained in the preceding chapter, accepting the arbitrary is as
irrational as accepting the contradictory. What is not rational to believe
is irrational to believe.
The issue is: on what grounds does one mark off an area with the sign,
“Here reason does not enter”? It cannot be on rational grounds: the attempt
 H OW W E K N OW • 9 : P rinciples 314

to use reason to exclude reason from a given domain is a contradiction:

it means holding, “By the nature of the things in this domain, reason cannot
know them.” Yet, supposedly, it is precisely the nature of those things that reason
cannot know. Let us even assume that the things in the domain are not held
to be contrary to reason, only inaccessible to reason. But how does one know that?
Not by reason, because reason has been excluded. It must be that the “limits”
set for reason are imposed by faith, or by whatever non-rational, non-validatable
means of knowledge one claims to possess.
Thus, the claim reduces to: “I have no reason to think that there are things
inaccessible to reason; I accept that on faith.” Which is to say that it is an
arbitrary assertion and therefore against reason.
Whenever someone asserts: “There are areas that the human mind cannot
penetrate,” he has either just penetrated it (in contradiction to his claim),
or else he has made an arbitrary assertion. (There is a third possibility:
he could claim to be superhuman.)169
Since the only demonstrable human faculty that can stand opposed to
reason is emotions, the violation of a rational principle necessarily inculcates
the principle that emotions are superior to reason. Since reason is the fac-
ulty required to identify the facts of reality, taking emotions to be superior
to reason means taking them to be superior to reality — which is the primacy
of consciousness. Beneath all the sophistries and rationalizations, the man
who violates a rational principle is saying: “This will work, because I want
it to work.”

“Just One Contradiction”

A principle identifies an action that is required by the facts at hand. To violate
a principle is to act as if what is required were not required — a contra­diction.
The oft-heard excuse “Just this once” means: “It’s safe to accept just this one
contradiction.” But accepting a contra­diction undercuts the whole structure
of one’s knowledge. It forces a puzzle-piece into a space it does not fit, spoil-
ing the overall picture (and leaving no place to put the right piece).

169 This fallacy is committed in a secular form by the Positivists (e.g., Ernest Nagel) who
assert that the laws of logic are not truths about reality, but merely ways that the human
mind is compelled to think. This entails the self-contradictory claim: “Contradictions
may exist, but we cannot conceive that they exist” — a claim that begins by conceiving
that contradictions do exist. [Nagel, 1949]
 A n E x ample : I ndividual R ights 315

Clinging to the contradiction undercuts method as well as content: one

becomes unable to check ideas for consistency with all that one knows. With-
out that consistency-checking, one cannot distinguish knowledge from mere
belief or feeling. The contradiction leads one into a swamp of subjectivism,
bereft of epistemic guidance. The only way out is to renounce the attempt to
“get away with” a contradiction, repair the damage done, and resume the task
of non-contradictory integration, as a matter of principle.
To see the cognitive consequences of accepting a contradiction, con-
sider the simplest possible case: accepting a contradiction in arithmetic.
Let’s take a hard case: a “small” contradiction not at the base of arithmetic,
but further down the line. Suppose one accepts the contradictory idea that 14
= 15. Can’t one still have arithmetic?
No, because arithmetic is an integrated whole. Consider: if 14 = 15,
then what is the result of 15 − 14? Is it 1? Is it 0? There’s no way to know.
What is 14 + 14? It could be 28, 29, or 30.
Such undecidable questions cannot be quarantined. What is the result of
2 x 14? Is it the same as 2 x 15? Is 14 even or odd? Since 14 is supposed to
be the same quantity as 15, we can’t answer. What happens to the Pythago-
rean Theorem if we have a right triangle with one side 14 inches long and
one side 15?
A contradiction, if maintained, paralyzes thought. One can proceed only
by abandoning logic and just making up an answer as an arbitrary dictum.
Ultimately, the alternative is: adherence to logic or cognitive paralysis.

An Example: Individual Rights

A vivid concretization of the contextual absolutism of principles is provided
by the principle of individual rights. On an objective theory of rights, rights
are moral principles. Rights are the application of the principles of morality
to man’s dealings with others; they demarcate the individual’s proper sphere
of independent action. To say that a man has the right to do X is to say that
he should be the one to choose whether or not he does X; no other person
or group may force their choices upon him. Rights prescribe freedom by
proscribing coercion.170

170 “A ‘right’ is a moral principle defining and sanctioning a man’s freedom of action in
a social context.” [Rand, “Man’s Rights,” VOS, 110] See this essay for a new theory of what
individual rights are and how they are validated.
 H OW W E K N OW • 9 : P rinciples 316

Principles are contextual. Rights are contextual in that they arise only in
civil society; they provide guidance in organizing a proper social-political
system under government. Locke’s “state of nature” is outside the context
to which rights apply. There are no rights on a desert island, or between
two 17th-century fur trappers who cross paths in a deserted wilderness.
Likewise, rights do not apply to the survivors of a shipwreck clinging
to a life raft on the high seas.171 But in civil society, in non-emergency
conditions, rights define objectively the proper sphere of an individual’s
freedom of action.
Within their proper context, principles are absolute. The principle of
individual rights starkly illustrates the absolutism of principles: rights exist
to define what takes precedence over what in cases of conflict, social or indi-
vidual. In such cases, rights define the supreme moral consideration — that
over which nothing can take precedence. The right to free speech, for exam-
ple, cannot be superseded by any other ethical or political consideration. If
one holds that free speech may be abridged for the sake of “promoting virtue,”
preventing “blasphemy,” or for achieving any other “social good,” one has
thereby implied that free speech is not a right, but a permission.
In any decision-making process, two competing considerations cannot
both be supreme. It’s either/or — either rights are inalienable, uncondi-
tional, absolute — or they are not and can be overridden by something else.
If the latter, then they are not rights. Just as reason cannot “leave room” for
faith, so rights are precisely that which cannot be compromised.
The popular idea of “conflicting rights,” which must be “balanced” against
each other, is a contradiction in terms (e.g., the conflict between a home-
owner’s property rights and the public’s alleged right to eminent domain).
There are certainly cases in which it has not yet been determined which party
has the right and which party must yield. But the idea of a conflict among
the very principles used to resolve conflicts is incoherent. A political right
functions like a right-of-way in driving or boating: both are designed to
settle the issue of who may proceed and who must yield. If someone were to
claim that two cars approaching an intersection each have the right of way,
it would be clear that he fails to understand what a right of way is (or else
that the rules of the road have been improperly defined). The same is true for
anyone claiming that political rights can be in conflict. Rights are the

171 In the fur-trapper example, the individuals remain obligated, by individual morality,
to treat each other justly and peacefully (unless threatened); but if one constructs
a lifeboat example in which it is impossible for all to survive, one has placed it outside
the context in which the virtue of justice applies.
 P hilosoph y and P rinciples 317

principles defining who may act independently, without interference, and

who must refrain from interfering. Just as two competing principles cannot
each be supreme, so two disputants in a conflict cannot each have the right
to have his own way.
Rights are thus absolute — within their proper context.

Philosophy and Principles

Philosophic principles, like those of any other science, are reached by induc-
tion from experience. Philosophic principles are the fundamentals of the
widest domains: existence and consciousness. Rand defines “philosophy”
as the science that studies “the fundamental nature of existence, of man, and
of man’s relationship to existence.” [PWNI, 2]
Philosophic principles have real-world application, as do principles in
any other field. People understand that the principles of, say, physics have
immense practical value, but they fail to see that philosophy has even more.
Indeed, philosophy, in identifying the nature of existence and the rules of
cognition, is the base of physics. The formation of any lesser generalization
and the recognition of its value depend upon the implicit acceptance of its
philosophic base. For instance, the generalization “All matter attracts other
matter” could not have been discovered by consulting sacred texts or wait-
ing for revelation; nor could it be reached on the premise that the world is
governed not by natural law but by the decree of an omnipotent super-spirit;
nor could it have any value to those who hold that this life is God’s punish-
ment for sin — all of which is why this generalization was not reached in the
one-thousand years ruled by religious mysticism.172
Every adult, whether he has formally studied philosophy or not, has,
in fact, stored a whole host of conclusions that reach a philosophic level.
It is philosophical ideas that are being expressed in colloquial formulations
such as: “Facts are facts” or “What’s true for one person isn’t necessarily true
for another”; “There’s an explanation for everything” or “Lots of things must
remain as impenetrable mysteries”; “I can understand things” or “I can’t ever
be certain on my own — it’s safer to follow others”; “I’m in charge of my

172 “For in much wisdom is much vexation, and he who increases knowledge increases
sorrow.” (Eccles. 1:17–18) And: “. . . there is scarcely anything so lowly or so simple to
understand that man can thoroughly grasp or fully understand. . . . we should live in the
constant presence of death.” [Innocent III, 12, 24]
 H OW W E K N OW • 9 : P rinciples 318

own life” or “People are as society makes them”; “My life is my own to live”
or “My duty in life is to serve the needs of others.”

You have no choice about the necessity to integrate your obser-

vations, your experiences, your knowledge into abstract ideas,
i.e., into principles. Your only choice is whether these ­principles
are true or false, whether they represent your conscious,
rational convictions — or a grab-bag of notions snatched at
random, whose sources, validity, context and consequences you
do not know, notions which, more often than not, you would drop
like a hot potato if you knew. [PWNI, 5]

Philosophic principles, like those of physics, have real-world applica-

tion — deductive application. One’s actions are shaped by one’s philosophic
conclusions — implicit or explicit, rational or irrational, correct or mistaken.
The influence of one’s worldview on concrete issues and decisions is unavoid-
able, because the worldview is automatized. (Indeed, it forms an integral part
of who one is.) Just as the investor who has automatized the importance of
diversification looks at his investments from that perspective, so one who has
automatized the idea that what’s true for you is not necessarily true for me
will view disputes between men from that perspective.
Since philosophy consists of the highest-level principles, since it is
supported, rationally or irrationally, by the widest range of particulars, since
adults have automatized it and built their lives upon it, philosophy’s effect on
men’s lives — and on history — is fundamental.
Ironically, it is the principles of philosophy men have accepted or
absorbed that explain the current widespread contempt for principles and
for philosophy.
In American culture today, anyone who operates short range, reversing
himself moment to moment, is admired as being “pragmatic” and not “rigid.”
Our leaders continually tell us to adapt to “new circumstances,” and claim
that “the tired, old solutions of the past” will no longer work (ignoring the
fact that that idea is one of the tiredest, oldest “solutions” of the past).
This cultural contempt for principles is relatively new. America’s Found-
ing Fathers were men of principle. (Note that they rose in rebellion against
very light taxes on stamps and tea.) Opposition to principles originated with
philosophers , especially the Pragmatists William James and John Dewey.
They influenced the educators (Dewey launched “progressive education”)
and other opinion-leaders, who then spread the anti-principle attitude across
 P hilosoph y and P rinciples 319

the culture. The average man is not a philosophic innovator; he gets his phil-
osophic framework from the intellectual leadership, as it filters down to the
educators, editorial-writers, artists, journalists, etc.
For over a century, philosophers and the intellectuals they have influ-
enced have opposed principles. Their scorn for principles is a consequence
of their underlying opposition to concepts. Due to the lack of a proper
theory of how concepts are formed and how objectivity is attained, there
was no proper defense of principles. The only (seemingly) pro-principle
voices have come from the religionists (who defend not actual principles
but frozen dogmas). The resurgence of religious fundamentalism, unleashed
by the intellectuals’ evisceration of reason, is only the other side of the
false alternative: commandments vs. whims. The actual alternative to both
commandments and whims is: rational principles, principles reached by
a hierarchically ordered series of abstractions based upon perceptual obser-
vation, principles held as contextual absolutes.
To understand and defend principles, one must recognize that the law
of identity applies to consciousness fully as much as to the external world.
The knower, not just the known, has a specific, delimited nature. Three facts
about the identity of man’s consciousness are fundamental to epistemology:
1) man’s senses provide an awareness of the world, in a form determined
by his sensory physiology, 2) man’s consciousness can hold only a limited
number of units in a single frame of awareness (“the crow epistemology”),
and 3) the conceptual level is volitional (the subject of the next chapter).
Together, these facts explain the cognitive function of concepts — and
thus of principles. Concepts and principles are consciousness-expanders:
starting from the data of perception, they expand the range of one’s awareness,
overcoming the “crow-limitation” by means of condensing an open-ended
multiplicity into a unity. Concepts integrate a limitless number of percep-
tual concretes into a single unit, retained by one word; principles integrate
a limit­­less number of derivative truths into one fundamental relationship,
retained by one proposition. Concepts carve steps into the cliff face; princi-
ples describe the essentials of the scene spread out before one, after mounting
those steps to reach the summit. The summit’s view is not available to those
who reject principles.
Both Pragmatists and religionists hold that “real life” cannot be governed
by reason, that concepts do not integrate percepts, that theory is opposed
to practice, and soul is opposed to body. In all these false dichotomies,
the religionists side with the soul, the Pragmatists with the body. The motto
of Pragma­tism is: “It may be good in theory, but it doesn’t work in practice.”
 H OW W E K N OW • 9 : P rinciples 320

The religionists (or Platonists) tell us, “It’s good in theory, forget practicality ;
that’s materialistic.” The objective (and Objectivist) answer to both is:
“What makes a theory good is that it names the fundamental requirements
of successful practice; there can be no mind-body dichotomy.” If a theory
does not work in practice, that means the theory is wrong or misapplied.
If the practice is to succeed in promoting life on this earth, it must obey the
requirements identified by a valid theory.
Concepts and principles are neither social conventions nor ends in them-
selves. Concepts are the means of grasping knowledge beyond the animal
level of “here now this.” Principles are neither commandments — “Thou
Shalt Nots” set against the needs of living on earth — nor ad hoc rules of
thumb. Rather, they are cognitive maps of the world. Using the map analogy,
Rand summarizes the anti-principle attitude:

The present state of our culture may be gauged by the extent to

which principles have vanished from public discussion, r­ educing
our cultural atmosphere to [one] . . . that haggles over trivial
concretes, while betraying all its major values, selling out its
future for some spurious advantage of the moment. . . . and by
panicky appeals to “practicality.”
But there is nothing as impractical as a so-called “practical”
man. His view of practicality can best be illustrated as follows:
if you want to drive from New York to Los Angeles, it is
­“impractical” and “idealistic” to consult a map and to select the best
way to get there; you will get there much faster if you just start out
driving at random, turning (or cutting) any corner, taking any
road in any direction, following nothing but the mood and the
weather of the moment. [CUI, 144–145]

Principles are a biological imperative for Homo sapiens. To say to a human

being, “Don’t be theoretical” is like saying to a bird, “Don’t fly.”
Principles are the fullest realization of how we know. That is why they are
how we survive and prosper.
 10
FREE WILL

I have emphasized the fundamental difference between the

perceptual level and the conceptual level. Perception is automatic,
unchosen; conceptual activities are volitional — they are subject to choice.
But what does choice mean and imply? Choice operating where? Caused
by what? Affecting what? Now is the time to meet those questions head on
— i.e., to take up the topic of free will.
Free will has traditionally been thought of as the ability to choose among
alternative physical actions, e.g., choosing between going to work and
going to the corner saloon. This view is not false, but it is quite superficial.
Physical actions are not primaries; one’s body does not unaccountably lurch
off in one direction or another. The actions of one’s body are controlled
by the actions of one’s mind: one decides what to do.
A decision, however, is also not a primary: it is the outcome of a decision-
making process, and the input to that process is constituted by one’s beliefs
and values, specific and general.
One’s beliefs and values are, in turn, the products of earlier processes. The
process fashions the product. All conceptual products — all ideas, values,
theories, and convictions — are caused and shaped by the processing that
one employs in reaching them.
 H OW W E K N OW • 1 0 : F ree W ill 322

That processing can be performed rationally or irrationally. One can reach

conclusions by a conscientious, fact-centered process of thought, or by any
irrational substitute, such as emotion-driven leaps in the dark or unthinking
absorption of the beliefs and values of others.
Here we have reached the actual primary: the rationality or irrationality of
one’s mental processes. It is this that is under one’s direct, volitional control.
This is the understanding of free will originated by Ayn Rand,173 and
hers is the first philosophy to recognize that free will is fundamentally an
epistemological issue, that it pertains to conceptual cognition as such.

. . . man is a being of volitional consciousness. Reason does not work

automatically; thinking is not a mechanical process; the con­
nec­tions of logic are not made by instinct. The f­ unction of your
stomach, lungs or heart is automatic; the function of your mind
is not. In any hour and issue of your life, you are free to think or to
evade that effort. But you are not free to escape from your nature,
from the fact that reason is your means of survival — so that
for you, who are a human being, the ­question “to be or not to be”
is the question “to think or not to think.” [AS, 1012]

Man’s free will consists in his sovereign control over how he uses his own
mind. How a man uses his mind determines the rest: the conclusions he
reaches, the goals he sets, the action-decisions he makes. But nothing, in turn,
controls how he uses his mind; that is his “sovereignty”: the causal chain
begins within one’s own mind. Having the power to initiate a rational process
makes man an autonomous, self-regulating being, not a robot programmed
by outside forces.
Sovereign control is not omnipotence; one cannot take actions that would
violate the identity of one’s conceptual faculty. One cannot, for example,
will to be infallible or omniscient. Nor is one immune from the effects
on cognition of drugs or a blow to the head. But given a normal, healthy
brain, one’s sovereign control means that such factors as genes, upbringing,
subculture, and desires do not control whether or not one engages in
rational thought.

173 William James, in The Principles of Psychology, recognizes that free will is psychological,
and his discussion there of will as control of attention is quite similar, though not identi-
cal, to Rand’s idea of “mental focus.” [James, 1890] Also, an Aristotelian of the 2nd century
A.D., Alexander of Aphrodisias, proposes essentially the same idea [Quaestiones III.13].
 FREE WILL 323

The doctrine that denies this sovereign control is known as “determinism.”

Determinism is the theory holding that antecedent factors beyond man’s
control necessitate everything he is and does, including the nature of every
mental process he performs. Determinists claim that one’s sense of control
over one’s own mind is illusory. According to determinism, every event in
one’s consciousness is only a reaction, necessitated by prior events, which
were in turn necessitated by still earlier events, and so on, reaching back in
time to before one’s birth, before human beings evolved, before the Earth
was formed. According to the theory of determinism, whatever happens in
consciousness, as in the mater­ial world, had to happen, with no alternative,
no actual control, no freedom, no genuine choice.
But if a man’s mind were in thrall to some factor that forced ideas upon
him, he could not validate his conclusions objectively, and thus he could not
distinguish knowledge from mistaken belief. Objectivity requires guiding
one’s thinking by logic (see Chapter 6); the operations of a deterministic
mind would be ruled not by logic but by some necessitating factor — genes,
social “conditioning,” brain structures, “confirmation bias,” etc. A deter-
ministic mind would only emit outputs, like a computer; it could not make
objective judgments, distinguish the logical from the illogical, or separate
truth from falsehood. As I show later in this chapter, if conceptual processes
were deterministic, they would be like perceptual processes: metaphysically
given and incapable of error . But denying the existence of error lands one in
a contradiction: “The belief that error exists is an error.”
If men could not think or act otherwise than they do, neither ethics nor
epistemology would make sense: one cannot evaluate the impossible. To say
that one ought to take some action presupposes that one can choose to take it.
For instance, logic instructs one to hold the full context in mind in mak-
ing judgments — which presupposes that one can do so, that nothing blocks
one from consulting all the relevant information available, that no mental
content or thought is barred from awareness, that one is free to consider all
sides, pro and con.
The same sovereign control is presupposed by all the principles of logic.
“Dismiss the arbitrary” presupposes that one can dismiss it, and “reject the
Appeal to Authority” presupposes that one can reject it, that one is capable
of independent thought, rather than being deterministically “conditioned”
by parents, teachers, and social practices.
In short, logic, as a normative field, presupposes that nothing “runs” one’s
mind, that one has sovereign control over its operation.
 H OW W E K N OW • 1 0 : F ree W ill 324

Such control is precisely what Ayn Rand identified as the locus of

man’s free will. Man’s ability to take charge of his mind gives him control over
all the other aspects of his life: how he uses his mind shapes his conclusions,
goals, actions, and character.

That which you call your soul or spirit is your consciousness, and
that which you call “free will” is your mind’s freedom to think or
not, the only will you have, your only freedom, the choice that
controls all the choices you make and determines your life and
your character. [AS, 1017]

Focus
One’s power to take hold of the mental reins is always present — but it is
present as a choice. One does not have to take charge of one’s mind; one is
always free to drop effort, control, concern with reality, and just “space out.”
In Rand’s terms, one has the choice between focusing one’s mind or not:

Thinking requires a state of full, focused awareness. The act of

focusing one’s consciousness is volitional. Man can focus his
mind to a full, active, purposefully directed awareness of reality
— or he can unfocus it and let himself drift in a semiconscious
daze, merely reacting to any chance stimulus of the ­immediate
moment, at the mercy of his undirected sensory-perceptual
mechanism and of any random, associational connections it
might happen to make. [VOS, 22]

Mental focus is wider and deeper than thinking; it is the precondition

of thinking. Focus is the purposeful mental “set” that underlies and drives
the process of thought. (Thought is focus sustained on a specific topic,
to answer a definite question.)
Optical focus is a state of clarity and sharpness. Focusing one’s mind,
as with focusing the image from a movie projector, aims at making things
clear — clear conceptually, in the case of mental focus. As Peikoff aptly
describes it, “ ‘Focus’ (in the conceptual realm) names a quality of purposeful
alertness in a man’s mental state. ‘Focus’ is the state of a goal-directed mind
committed to attaining full awareness of reality.” [OPAR, 56]
 F ocus 325

The essence of focus is purposefulness. To focus is to set and enforce

a goal — the goal of gaining a clear, integrated understanding of the world,
and of oneself.
Purposefulness is not a static state but a process of steering one’s mind,
adjusting its activities to the requirements of attaining full, clear knowledge
of reality. To be in focus means to manage the operations of one’s own mind
in pursuit of a cognitive goal.
But one also can choose to mis-manage one’s mind: to deliberately slam
one’s mind shut in the attempt to deny the reality of something that one
knows, clearly or dimly, is real. This is evasion, which Rand describes as:

. . . the act of blanking out, the willful suspension of one’s

consciousness, the refusal to think — not blindness, but the
refusal to see; not ignorance, but the refusal to know. It is the act
of unfocusing your mind and inducing an inner fog to escape the
responsibility of judgment . . . [AS, 1017]

Evasion has a purpose, but that purpose is non-awareness. To avoid

acknowledging some unpleasant or frightening fact, one steers one’s mind
away from it, working to get it out of awareness.
To focus is to set the goal of awareness, as distinguished from two other
states: evasion and “drift.” To drift is to fail to set the goal of awareness; to
evade is to set the goal of non-awareness.
Either one sets one’s mind to the task of knowing reality, or one does
not. Every use of one’s conceptual faculty — whether to choose a course of
behavior, or choose a value, or make a judgment, or form a concept — can
be done in focus or out of focus. In focus: as a purposeful, self-monitored,
reality-oriented process. Out of focus: either as aimless, unsupervised drift
or as emotion-driven evasion.
Being in focus is like steering a car. Steering has two aspects: observing
the road ahead and turning the steering wheel. Likewise, steering one’s
mind involves both monitoring and directing. Monitoring consists of
observing what is going on in one’s mind; directing is issuing to oneself
the words, or wordless equivalents, that will cause the next step.174

174 The words act upon one’s subconscious, which is embodied in neural mechanisms.
The study of how the conscious mind interacts with the subconscious is the province
of “psycho-epistemology” [see RM, 18].
 H OW W E K N OW • 1 0 : F ree W ill 326

For instance, suppose one monitors a sense of confusion in response

to reading a statement about financial planning. In that case, directing
one’s mind to deal with it might take the simple form of asking oneself:
“What’s confusing me?” Or just: “Huh?” In other cases, some groping might
be involved, as in this sample mental monologue:

“I’m uncomfortable. What’s wrong? I’m confused. What’s confus-

ing here? Oh, the phrase ‘yield to maturity.’ What’s confusing
about it? I don’t know what ‘to maturity’ adds. What can I recall
about ‘to maturity’?”

(Here, I have put the mental process into full sentences, to express it for
the reader. Often one will use a mental shorthand, so that the whole sequence
might occur in a foreshortened form, like: “What? ‘To maturity’?”)
The process of seeking clarity is not automatic. First, nothing makes one
monitor at all: a person can feel confusion but fail to attend to it, fail to iden-
tify that confusion is what he is experiencing. Or, he can be aware that he is
confused but not ask himself: confused about what? Or, he might ask himself
that, and even get the answer “I’m confused about the meaning of ‘yield to
maturity,’ ” yet shrug it off with: “Who knows?” All of these are states of less
than full focus.
(In the last case, an in-focus decision might well be, to again put it in
a full sentence: “I don’t understand ‘yield to maturity,’ but, all things
considered, it’s not worth my time to figure it out, so I will note that
I don’t understand it and table this issue for now.” The issue of focus
concerns taking active control, being “hands on,” not the particular judgment
one makes.)
To be in focus means to make full, unconditional awareness an absolute
of one’s mental functioning: “It means one’s total commitment to a state
of full, conscious awareness, to the maintenance of a full mental focus in
all issues, in all choices, in all of one’s waking hours.” (VOS, 28)

Rationality
Epistemologically, focus means rationality. Rationality is the exercise of
reason; reason is the conceptual faculty. “Focus” names the introspectible
sense or “feel” of the psychological state; “rationality” names the kind
of activities that one performs in that state. The choice to function
 F ocus 327

rationally is the choice to utilize one’s conceptual faculty, activating one’s full
cognitive resources. This choice consists in exerting the effort to take charge
of one’s mind, to set it to the task of understanding what one is dealing
with, to monitor one’s own mental operations and direct them toward what
one judges to be their most effective, rational deployment — as opposed to
anything less than that.
In contrast, sensory perception is automatic, physiologically determined,
and not subject to introspective monitoring. In looking at a table, one is not
aware of the neurophysiological processes that make possible the percep-
tion of the table. In fact, we learn of the existence of the neurophysiological
processes only extrospectively, by scientific investigation.
But on the conceptual level of awareness, the process of cognition is
conscious, deliberate, and introspectible. The elementary act of ­grasping
a new concept requires doing conscious mental work. Even the child learn-
ing the term “table” from a parent has to put forth the effort to attend
to similarities and differences among the objects he perceives, make
conscious comparisons to recollected objects, and grasp that the word “table”
stands for all similar concretes, irrespective of when he considers them.
(Forming such simple concepts is so easy for a human being that even an
out-of-focus child will eventually pick it up, though without the in-focus
child’s sense of achievement.) More mental work is required for an adult
learning advanced abstractions, such as “yield to maturity.” Some people put
forth the effort required, some do not. The choice is up to each individual.175
No effort is required to see the words on this page, but to understand
what is written, to integrate it, grasp its implications, judge its validity
— that requires the choice to do mental work. You don’t have to will
your eyes to see, but you do have to will your brain to think. Perceptual
processes are automatically in contact with the world; conceptual processes
are not. Conceptual processes performed out of focus result in mental
content that is invalid, subjective, out of touch with reality.
To engage in a process of rational thought, one has to take charge of
one’s mind and set it to the task of identification and integration — that is,
one has to focus one’s mind. The alternative is not simply that of full, consis-
tent rationality or a complete retreat to near the perceptual level; there are
various levels of conceptual awareness in between, and (unfortunately) most

175 Choosing not to form a technical concept is not objectionable and doesn’t imply
being out of focus; but using a term without concern for whether one understands
it or not is and does.
 H OW W E K N OW • 1 0 : F ree W ill 328

people drift somewhat passively among them. Full focus is not a momentary
foray into the cognitive realm but an ongoing commitment to full awareness.
The basic volitional choice, to focus one’s mind and proceed rationally,
does not concern differences in intelligence, interest, or prior know-
ledge, but solely the difference in the degree of purposeful effort put forth
to understand reality. E.g., at a university lecture even the slowest, most
uninterested and unprepared student has the power to concentrate on the
lecture and seek to understand, integrate, and judge it. Conversely, even
the most brilliant and well-informed “A” student can simply let his mind go,
tuning out the lecture.

The Reality-Orientation
Metaphysically, rationality consists of instituting the “reality-orientation,”
in Peikoff ’s phrase. [OPAR, 66] Rand writes that rationality, “is the recog­nition
of the fact that existence exists, that nothing can alter the truth and nothing can
take precedence over that act of perceiving it, which is thinking . . . .” [AS, 1018]
Thus, being rational means recognizing that existence has primacy over
consciousness, recognizing that the mere presence of an idea in one’s mind,
no matter how correct it might feel, offers no assurance that it corresponds
to the facts of an independently existing reality. An irrationalist functions
on the opposite implicit premise, the premise of the primacy of conscious-
ness, according to which facts snap into line with whatever passes through
one’s mind or whatever one wishes the facts to be.
Does one go by evidence or by whims and authorities? Does one latch
onto whatever idea first strikes him as plausible, or does he work to inte-
grate it into the sum of his knowledge, checking for contradictions? The basic
alternative here is one’s orientation: does one accept reality and seek to know
it, come what may? Or, does one consider reality expendable when it clashes
with one’s desires, one’s self-image, or the beliefs of others?
The basic choice is a single primary but, as we have seen, it has s­ everal
aspects and thus is describable as focus, rationality, purposefulness,
self-monitoring, managing, adopting the reality-orientation, and in still
other ways.176 Rand writes:

Psychologically, the choice “to think or not” is the choice “to focus
or not.” Existentially, the choice “to focus or not” is the choice “to

176 For more discussion, see OPAR, ch. 2, and Binswanger, 1991.
 M otives and F ocus 329

be conscious or not.” Metaphysically, the choice “to be conscious

or not” is the choice of life or death. [VOS, 22]

Motives and Focus

The basic choice to focus and take charge of one’s mind is independent of
any specific motive.177 If being in focus were conditional upon some specific
motive, we would remain in a daze until something grabbed our atten-
tion, and upon losing interest, we would go out of focus. And if some topic
was unpleasant or frightening, we would automatically evade it. In other
words, if the choice to focus had to wait on and remain subject to some
lure from without or within, then we would be hopelessly passive, irrational
— and mankind could not have survived.
Rather than requiring a specific motive, the choice to focus is uncondi-
tional. You can will yourself to manage your mental operations and adopt
the reality-orientation without needing a concrete incentive to do so.
Even when you feel that you really don’t want to focus, you can still will it.
Not only do you need no specific motive to focus, you can focus in the teeth
of a desire to evade that responsibility.
If man were unable to focus unless he felt like it, he would not be the
rational animal; he would be precisely the irrational animal that scores of
philosophers have claimed him to be.
People can be irrational. They can, by their own choice, make the exercise
of reason wait upon the permission of their feelings. Their operating premise
is: “I look at reality — when my feelings permit it.” But rationality is precisely
the unconditional commitment to perceiving reality, the commitment
to grasping and accepting any relevant fact, whether pleasant or not.
Peikoff makes the point that motives cannot compel consciousness,
because motives presuppose consciousness.

The choice to activate the conceptual level of awareness must pre-

cede any ideas; until a person is conscious in the human sense,
his mind cannot reach new conclusions or even apply previous
ones to a current situation. There can be no intellectual factor

177 Focusing is not an end in itself. It has what can be called a “metaphysical” motive:
the desire to live, which means the commitment to acting in and achieving values
in reality. This fundamental motive is, however, universally applicable, in contrast to any
narrower motive (which presupposes the desire to live and function).
 H OW W E K N OW • 1 0 : F ree W ill 330

which makes a man decide to become aware or which even partly

explains such a decision: to grasp such a factor, he must already
be aware.
For the same reason, there can be no motive or value-judgment
that precedes consciousness and which induces a man to become
conscious. Instituting the reality-orientation must precede value-
judgments. Otherwise, values have no source in one’s cognition
of reality and thus become delusions. Values do not lead to
consciousness; consciousness is what leads to values. [OPAR, 59–60]

This point above applies not only to the alternative of consciousness

versus unconsciousness but also to intermediate, semi-conscious states.
Semi-conscious content need not be accepted as real and binding; it can
be treated just as stray thoughts or imagination.178 In a state of less than
full focus, by definition, one is not committed to the reality of reality;
one is acting as a passenger in one’s own mind, merely spectating upon
snatches of out-of-context ideas. An idea or goal that floats through a semi-
focused mind has only such causal power as one gives it: one is free to choose
whether to stop, consider it, take it seriously (i.e., to deal with it in full focus)
or to let it pass away, as if it were unreal (i.e., to remain in semi-focus or
actively evade).
While in a semi-focused state, a man may happen to feel a desire to
focus and think (or at least to know more about something that has arisen).
But while in that state, the desire is not an absolute commitment; it can be
a faint desire, a wistful longing, a vague sense of guilt — all in contrast to the
active resolve to know the full context and the full truth. Moreover, while not
in full focus, the prospect of “really looking” may sometimes seem unpleas-
ant or even frightening. The primary choice is whether or not to “really look”
despite any such feelings.
The possibility of semi-focus or half-hearted efforts to know reality
does not get around the point that it is either/or: either the actions of one’s
consciousness are surrendered to transient feelings, including feelings about
focusing itself, or one is committed to full awareness of reality.

178 I owe this point to Peter Schwartz (personal conversation).

 T he C hoice of Action 331

The Choice of Action

As I have stressed throughout this book, consciousness is a biological faculty
whose function is to guide action. How, then, does the choice to focus relate
to one’s bodily actions?
First, I must clarify what I mean by “action.” I do not include those local
muscular contractions that occur without conscious initiation and control,
such as the “knee-jerk” reflex. Some borderline cases are sneezing, cough-
ing, vomiting. These may well have an origin in consciousness, but they
“come over us,” rather than being voluntary actions. A greater involvement of
consciousness occurs in brief, trivial acts, of which one is only peripherally
aware, such as absent-mindedly scratching one’s scalp.
Philosophy is concerned with “human action” in the full sense of the term:
purposeful action, action consciously initiated and sustained in p ­ ursuit of
a pre-envisioned end. Although even very minor decisions are subject
to volition and can be made rationally or irrationally, the most fruitful kinds
of examples to keep in mind in thinking about the choice of action are the
more dramatic cases, such as deciding which candidate to vote for in an elec-
tion, choosing between buying a new house and remaining in one’s present
home, choosing a career.
In these kinds of practical, existential choices, and in all lesser variants,
what is the role of the basic choice? Perhaps surprisingly, the answer is:
the same as its role in cognition. In deciding how to act, as in deciding how
to think, the state of focus remains the fundamental, controlling factor. The
conscious ego is capable of controlling both the operations of thought and
the muscles of one’s body; both mind and body can be guided rationally, in
focus, or not, according to one’s choice. The essential causality is the same for
directing one’s mind to take up a particular topic and directing one’s body to
go to a particular location. Both are subchoices, implementing the primary
choice to focus (to the extent one is out of focus, they are not decisions at all
but capitu­lations to internal emotional pressures).
In discussing the relation of volition to action, it is crucial to avoid
Platonism, with its view that the intellect is sundered from action in the
world. Although I have generally presented the primary choice in intellectual
terms — e.g., as the will to know — the premise has been that one is seek-
ing that knowledge in order to guide one’s course of behavior, directly or
indirectly, concretely or abstractly. I have been speaking from a biological,
not a Platonic, orientation.
 H OW W E K N OW • 1 0 : F ree W ill 332

Consciousness is biological equipment. Awareness evolved for its survival

function — to guide the actions of conscious organisms. Directly or in­directly,
the motive for knowing anything is to use it in action. To capture the life-
centered view of cognition, free of any trace of other-worldliness, Rand
crafted a statement that makes a dramatic reversal of Hume’s “is”–“ought”
dichotomy: “Knowledge, for any conscious organism, is the means of
survival; to a living consciousness, every ‘is’ implies an ‘ought.’ ” [VOS, 24]
Fundamentally, consciousness is navigational: it is identification used to
guide action. A prime example is the process of deliberating upon alterna-
tive courses of action. Deliberation seeks to identify the consequences of
alter­native actions, judging which one is best; if successful, its conclusion is
a decision to act. Deliberation is at once a quest to find truth and to direct
action; it is at once intellectual and practical.
The deliberative process need not be performed at all — one can simply
act on emotional urges. Even when one does do the work of deliberating,
the process can be performed rationally or irrationally — i.e., in focus or out
of focus. This is a matter of directly volitional choice.
The relation of the basic choice to a course of behavior is then clear:
one has the choice to deliberate and decide through a rational, in-focus
process — or not. For both cognitive and existential action, the volitional
element — the act of direct will — pertains only to the rationality of the
mental processes that lead to the action.
Some minor decisions, such as to pick up and read the newspaper, require
nothing more than an instant’s thought. Even more momentous decisions
can sometimes be made quickly and easily, because all the evidence points
in one direction and one is free of doubt and conflict — as when one gladly
accepts an unanticipated promotion at work. But what about cases that are
not open and shut, cases in which one has doubts or is in conflict?
For instance, suppose that you have the thought of reading the newspaper,
but it is accompanied by a vague, as yet unidentified, sense of guilt. Suppose
that this feeling’s actual source, also not yet identified by you, is that you had
resolved to start preparing your income tax return. You are experiencing the
result of two contradictory ideas: 1) It would be good to read the newspaper
now, 2) It would be wrong to do that now. Here, you face the three volitional
alternatives: focus, drift, or evasion. The act of pushing the guilty feeling out
of your mind, in order to avoid awareness of what you have only vaguely
glimpsed, would be evasion. Drift would be relinquishing cognitive effort,
 T he C hoice of Action 333

allowing your behavior to be determined by whichever feeling — desire or

guilt — eventually predominates.
The in-focus process consists of seeking to understand — to understand
both the world and yourself. In this example, that means: asking extrospective
and introspective questions — such as (extrospectively) “What is actually best
for me now, in the full context? What’s my schedule?” and (introspectively)
“What is the bad feeling that I’m having now?” (answer: guilt) “Why am I hav-
ing it? Guilt over what?” (answer: not doing my taxes now).
These are only some possible questions and answers; I cannot, of course,
predict what the actual questions and answers would be. The point is that
a rational process is available to you to resolve such conflicts, or to at least
make progress toward resolving them. And that rational process is devoted
to discovering the full truth. Drift, in contrast, is goal-less, and evasion
is devoted to avoiding knowledge.
Deliberation, decision-making, selection among alternatives — all these
acts can be done in-focus, conscientiously, or out-of-focus, i.e., negligently or
by a deliberate act of evasion that attempts to wish away unpleasant facts.179
Whether a decision concerns what to do next with one’s mind or what to
do next in existential action, the decision-making process is simultaneously
cognitive and action-oriented. Cognition is not severed from the action.
One finds out (cognition) what to do (action). And, looking in the other
direction, purposeful human action is the physical expression of having
reached a conclusion about what to do.
Man has the ability to reflect upon and evaluate his own decision-making
processes, and thus to control the processes that control his actions. One can
step back and rationally make a wide variety of judgments, such as:

• The considerations are too complex and/or too confusing

to decide at this time.
• I need to get further information before deciding.
• I am too tired to make a good decision now; I will revisit
the question later, when I am rested.
• I need to check that there is not an option that I have overlooked.
• I have to decide right now, even though I am uncertain,
so I must pick the option that looks best now.

179 In the Objectivist ethics, evasion is the fundamental act of immorality.

 H OW W E K N OW • 1 0 : F ree W ill 334

In cases of confusion or conflict, one can reach full certainty regarding such
“meta-cognitive” judgments.
To reach rational certainty on any topic requires applying a standard
of certainty.180 To be certain one must have conclusive evidence, and that
requires knowing, explicitly or implicitly, what counts as “conclusive evidence.”
Rough, generalized standards are available to anyone who can distinguish
between evidence and assertion, logic and illogic, reason and feelings,
fact and fantasy. (The objective standard, as explained in Chapter 8, is that
the integrated sum of the evidence supports and is consistent with one and
only one hypothesis, within a rationally delimited set of hypotheses.)
Just as thought in general ranges from the easy to the difficult, so do
the development and application of standards. It is trivially easy to judge
an object’s length by reference to the standard of the inch as marked off
on a ruler. It is quite difficult to devise and apply an appropriate standard
for judging a nation’s aggregate economic production. But whether a given
standard allows for precise measurement or only a rough approximation,
the use of a standard of judgment takes decision-making out of the realm
of emotion, of what just “feels right,” and into the realm of objectivity.
Cases of grappling with uncertainty spotlight the need to manage
one’s mind; only a self-conscious, fully focused state allows for stepping back
to judge how best to deal with the uncertainty.

An Extended Concretization
Here is an extended example of managing one’s mind vs. not doing
so — resulting in different actions taken. I project the thought process going
on in two students’ minds, on the night before their final exam in a course
on Ancient Greek philosophy. Each student is deciding whether or not to
study for the exam.

First student’s thoughts:

Am I going to study tonight, or watch TV? I guess I ought to

study — but that course — ugh! Professor Winston is so boring.
I don’t see why I should have to force myself to study something
if the teacher can’t make it interesting. He wants us to do all the
work while all he does is read his lectures from a script. He must
have written that lecture decades ago; there are jokes from the ‘80s

180 I owe this point to Shrikant Rangnekar (personal conversation).

 T he C hoice of Action 335

in them. I bet he hasn’t done any new work in years. Just takes it
easy. He’s probably watching TV right now. What’s on TV tonight?
Oh, that great old Bogart film. That’s the one I saw last year
with Maureen. I remember how she cried at the end of the film.
That was really nice. Maureen is nice. I don’t know why I didn’t
pursue a relationship with her. Hey, I could ask her over right now
to watch that movie on TV. Repeating our old date — that would
come across as real romantic.
Oh, but what about the final? Aw, I guess I’m prepared enough,
and everyone knows you can bluff your way through philosophy
tests. [He reaches for the phone.]

Second student’s thoughts:

Am I going to study tonight or watch TV? What’s at stake?

My grade-point average. If I don’t do well on the final tomorrow,
I’m going to get a C in the course, and that’s not going to look
too good when I apply to graduate school. But is studying tonight
really going to make a difference to my grade? Well, how well do
I already understand the material this exam will cover? Well, what
will it cover? Oh yeah, Professor Winston said it will be mainly on
Aristotle’s ethics. What are the topics of Aristotle’s ethics that we
are going to be responsible for? [Looks at his notes.] Happiness as
an end in itself, the golden mean, pride as the crown of the virtues.
Now how well prepared am I on those topics? Well, I wrote
my term paper on the golden mean — I know I’ve got that
down pat; no problem with the virtue of pride — that’s simple.
But happiness — didn’t that have some technical stuff in it — what
was it? Something about the soul? Yeah, that was confusing.
Professor Winston sure wasn’t clear about that. But then what is
he ever clear on? He expects the class to figure out what he’s saying,
rather than taking the trouble to explain anything. He just reads
his old lectures, and never — wait, that’s beside the point; let me get
back on track.
Okay, if I’m hazy on happiness, I’m not going to do too well.
Winston’s hinted more than once that Aristotle’s view of happiness
is important. I could re-read my notes on happiness and check
a secondary source.
 H OW W E K N OW • 1 0 : F ree W ill 336

But then that’s going to take time, and that great Bogart film
is on TV tonight. I’d hate to miss it. And after all, I’m entitled
to some pleasure in life. No, that’s just an excuse. Could I even
enjoy the movie, with tomorrow’s exam looming over me? No,
I really, really want to get into a good grad school; my whole
future depends on that.
But I want to see that movie! Do I have to miss it? Is there
some other way? Wait — it must be out as a video. Yeah, I’ll rent
the video tomorrow night after the exam as a reward for nailing
down Aristotle’s view of happiness tonight. [He reaches for his
study materials.]

The second student’s focused, rational process of thought, which I tried

to make prosaic rather than extraordinary, led him to the clear, u ­ nconflicted
recognition that studying was the thing to do — the thing he actually wanted
to do. The alternative, as he came to see it, was the investment of one
evening’s work or jeopardizing a long-range value of great personal impor-
tance to him. Plus, his thinking in the last two paragraphs completely
reversed his initial evaluation of seeing the movie that night; it no longer
seemed tempting. But his seeing the issue this way was the consequence
of his maintaining full focus.181
In contrast, the first student’s out-of-focus ramblings led him to the
opposite decision (or pseudo-decision). The choice as he saw it, given his
contracted range of awareness, was between an enjoyable date watching
a good movie while impressing a girl — or a painful and boring evening
carrying out an unjustly imposed duty. His decision in favor of watching the
movie followed from his out-of-focus mental set.
The example concretizes how the alternative of focus vs. non-focus under-
lies all the lesser alternatives one confronts. In the face of the same alternative
studying or not, one student chose rationally, the other irration­ally. That dif-
ference reflected their opposite mental “sets”: one student was committed to
full awareness, the other was not. To be rational, a decision must be made
in focus. If a decision (or a lurch into action) is made while out of focus,
it is not rational. Rational action means action proceeding from the full

181 Notice that he does not deny his emotions (“But I want to see that movie!”). He acknowl-
edges his desire, but he keeps thinking, and thus is able to find an overlooked way to get
the enjoyment without sacrificing his higher value.
 T he C hoice of Action 337

and honest use of reason. And that, in turn, means that one must generate
the effort to keep the full context clear, real, active — as the second student
did — instead of passively letting short-range emotional pushes and pulls
substitute for rational self-direction.
Whether one is functioning in focus or out of focus, the actions one takes
depend on the mental content that enters one’s mind. A series of ideas, images,
value-judgments, and emotions serve as input to the choice of what to do.
But the nature of that input depends on one’s level of awareness: the actively
managed awareness of a man in focus brings to his mind and makes real to
him considerations that would not occur to a man operating on a lower level
of awareness. While out of focus, one is intensely aware of the immediate,
perceptually present factors and only weakly aware of more abstract, long-
range considerations — as the first student is strongly aware of the pleasure of
watching the Bogart movie with his former girlfriend, but only weakly aware,
if at all, of the effects on his future of doing poorly on the exam.
Understanding men’s actions in this way enables us to give a good expla-
nation of the phenomenon that so puzzled the Ancient Greek philosophers:
how people can act against their own know­ledge. People sometimes take
an action while saying, “I know I shouldn’t be doing this.” The explanation is:
they don’t actually know at that moment that they shouldn’t be doing it. They
are not choosing to make that knowledge real and active in their minds.
Thus, the alleged clash of reason and emotion can be reduced to the issue
of context-holding vs. context-dropping. The clash is not between two inde-
pendent “parts of the soul,” but between a wider and a narrower context
of knowledge, between the way things appear (and how one feels about them)
when considered rationally, in the full context — vs. how they appear when
one relinquishes effort and allows the unsupervised, automatic response of
one’s subconscious to set the terms.
One’s emotions are produced by the brain’s automatic application
of one’s stored beliefs and values to the perceived facts. But one’s level
of focus affects which set of material, out of all that one has stored,
is activated and therefore operative at any given moment. Different beliefs
and values may be activated — and different emotions result — according
to whether one works to bring the full context to mind or myopically stumbles
along on the basis of whatever out-of-context beliefs and values get triggered
by what one perceives. In my example, the in-focus student no longer feels
 H OW W E K N OW • 1 0 : F ree W ill 338

the desire to watch the film on TV once he considers the full context, which
includes the fact that he can later watch the video.182
In summary, one’s fundamental choice is between purposeful, rational
management of one’s own mind and purposeless drift or deliberate evasion.

Volition and Character

By adulthood, one has formed a certain character, which includes dispositions
to behave in certain ways and not in others. The question thus arises: what is
the relation of free will to character? Does a man choose his own character?
Does his character, once formed, control his choices?
What is “character”? The term refers to a person’s moral makeup, as mani-
fested in his characteristic way of behaving. Some people are characteristi-
cally hard-working, some are slackers, some are mixed. Peikoff ’s defin­ition
names the fundamental: “ ‘Character’ means a man’s nature or identity inso-
far as this is shaped by the moral values he accepts and automatizes.” [Peikoff,
1976, Lecture 2]
The kind of moral values one accepts depends upon one’s rationality
or irrationality in thinking about issues of morality; one’s automatization
of moral values comes from action — from repeatedly acting h ­ onestly, justly,
industriously, etc. or failing to do so — which, in each instance, depends upon
one’s state of focus. Thus, the moral values that one ingrains in one’s soul,
resulting in a good character or a bad one, proceed from one’s direct area
of free will: one’s choice to be rational or irrational, to maintain full focus
or to drift or evade.
A man’s character is the net product of all his choices, over his life up to
the present. Man does not choose his character directly, but the nature of the
choices he makes adds up to a characteristic way of acting.183
Consider how one’s repeated choices in regard to the basic issue of focus
affect the development of one’s character. Focus concerns the level of aware-
ness at which one functions mentally. At a low level of awareness, only the

182 This is not to say that all conflicting evaluations can be resolved so easily. But the point
remains that one can go by either a rational, in-context evaluation or the snap judgment
produced by the unsupervised actions of one’s subconscious.
183 Of course, one can consciously dedicate oneself to instilling a good character in oneself;
the point is that that choice can be implemented only in each of the concrete choices
one confronts in the course of daily living, and that one’s character is formed or changed
not over hours or days but over years.
 Volition and C haracter 339

perceptually evident concretes and the emotions they automatically generate

shape one’s behavior. To the extent that one refuses to focus, surrendering
instead to short-range urges, one is establishing a weak-willed character. In
making the effort to function on a higher level of awareness, one is establish-
ing a self-disciplined character.
It is in this sense that we indirectly choose our characters — we forge our
characters, step by step, according to the nature of the specific choices we
make, for good or ill. Man, Rand writes, “is a being of self-made soul.” [AS, 1020]
The self-making occurs gradually, in an ongoing process.184
The sum of past choices does not eliminate choice in the present.
The choice to focus cannot be automatized, because it is precisely the choice
to take charge of one’s mind rather than letting the automatized take over.
Character does not eliminate choice. One can act “out of character.” But
one’s character limits how far one can go — at the moment.
A man who is characteristically rational and self-disciplined might
in a given instance wrongly permit himself an out-of-character rationalization
— say, to indulge in a diet-breaking snack. That sort of out-of-focus action
remains possible to him. On the other hand, there are things that would not
be possible for such an individual to do. He could not take it into his mind to
beat up his wife. He could not arbitrarily decide to inject himself with heroin.
He could not suddenly start acting in obvious contradiction to his basic
convictions. While other individuals, with a record of evasion (and the
psychological disintegration it wreaks) can do such things, the rational
man’s character neither permits that scale of self-betrayal nor gives him any
motive to act that irrationally.
Evasion normally sets off an internal “alarm” — a sense or signal from one’s
subconscious that something is wrong. After all, by evading, one is engaging
in a pretense, in a form of self-deceptive double-think: “I mustn’t look at that;
it’s not there.” But if it actually were not there, why the order not to look at
it? The implicit contradictions involved in acts of evasion produce a range of
emotional signals, from the simple sense of confusion and bafflement on up
to acute feelings of anxiety and guilt.
The individual with a rational character, with a long history of conscien-
tiously facing all the facts, pleasant or painful, is unaccustomed to hearing
such alarms. If he evades, the alarm is a striking exception to his normal
sense of clarity and calm. Thus, to continue the evasion would require more

184 By saying “gradually,” I do not mean to imply that all choices are tiny; there can also be
momentous choices that have huge consequences upon one’s character.
 H OW W E K N OW • 1 0 : F ree W ill 340

of him than it would for a less rational person: it would require that he evade
the alarm as well as the original evasion that set it off. In contrast, the individ-
ual who evades frequently is used to these alarms; to him they are the normal
background. One more alarm almost gets lost in the general din. Thus, for
him, one more evasion may not require any special effort.185
In addition, in order to act out of character, one must have a motive.
To be in focus is to place long-range values over short-range emotions.
But a lifetime of doing just that means that one is precluded from even being
tempted to take certain actions. Taking heroin is not emotionally ­attractive
to a man with an unbroken history of relying on and valuing his mind.
Even when momentarily out of focus, taking heroin is not on the table;
both his reason and his emotions scream against it. Or, he may on an occa-
sion get quite furious with his wife, but nothing within him would urge him
in the direction of striking her.
The reverse situation holds as well: it would be impossible for a man with
a long history of evasions to suddenly start acting just as a rational man
would. (Reforming his character and methodology is possible, but only
through a long, gradual process.)
Hume had it exactly wrong when he claimed that an action has no effect
if it is “out of character”:

Actions are, by their very nature, temporary and perishing; and

where they proceed not from some cause in the characters and
dispositions of the person who performed them, they infix not
themselves upon him. . . . [the act] leaves nothing of that nature
behind it. [Hume, II, III, 2]

But the causal nature of man’s consciousness means that every choice
leaves its trace. Choosing a certain way once makes it a little easier and more
natural to choose that way the next time. Giving in to a temptation increases
its motivational pull; dispelling the power of a temptation, by making
oneself aware of the long-range consequences, decreases any subsequent
motivational pull. The volitional aspects remain volitional, but the sub­
conscious is geared toward automatization; like an obedient servant,
it gets the message, the modus operandi, being established by the choices
one makes. Moreover, each choice has existential consequences, which them-
selves are either identified, ignored, or actively evaded. As Aristotle said,

185 The soul of an evader is dramatized and dissected by Ayn Rand in the character of
James Taggart in Atlas Shrugged.
 Volition and C haracter 341

it is by acting virtuously or viciously that we come to have (and maintain)

a virtuous or vicious character.186
Though one’s character limits the range of actions possible to him, that
range is not fixed. It can be gradually shifted, in a better or worse direction,
by the choices one makes in the present. The character-degrading effect
of bad choices is well illustrated in literature, from Aesop’s fables on.
These stories show how giving in to temptation strengthens the temptation
and makes possible actions that are progressively more irrational, and how,
on the other hand, making the right choice serves to reduce the temptation
to take wrong actions, strengthening one’s “character-muscles,” as it were.187
It is not only one’s motivation but also one’s “psycho-epistemology” that is
affected by how one chooses to use his mind. To the extent that one focuses,
one’s mental files are kept in good order. One not only obtains more infor-
mation to put in those files, one also becomes more easily aware of their
logical relationships, which acts to establish and maintain order among
the files. A simple example is that of identifying the genus of something; that
identification, especially if one treats it as important, acts to place its file
folder within the wider folder that represents its genus.
Establishing a track-record of full focus thus tends to make one,
­perhaps wealthy, but certainly healthy and wise — healthy motivationally
and emotionally, wise in the sense of having well-organized, essential-
ized, smoothly functioning files (i.e., concepts) to store one’s knowledge.
Conversely, a history of drift and evasion produces emotional conflict,
psychological problems — and inner chaos, psycho-epistemologically.
The process of character development begins in early childhood, before
one has philosophic knowledge and full self-control. How, then, can one be
morally responsible for a character that begins forming when one is at such
a primitive stage?
First, the young child’s character remains in a very fluid state for many
years, perhaps through late adolescence, the time when he does begin to deal

186 “. . . men make themselves responsible for being unjust or self-indulgent . . . for it is
activities exercised on particular objects that make the corresponding character.”
Nichomachean Ethics, III, 5, 1114a5–7. Aristotle, in this same section, seems to advance,
by implication, a theory of free will similar to Rand’s: “each man is somehow responsible
for the state of his mind” (a point he makes in answer to the claim that how men evaluate
things is not under their control). See 1114a31–b25.
187 The latter is the theme of the movie “Back to the Future” (Part I), which is both highly
entertaining and deeply philosophical.
 H OW W E K N OW • 1 0 : F ree W ill 342

consciously with more abstract issues. Empirically, we do observe that chil-

dren can change character markedly between, say, age four and twenty-four.
Some children remain on the same path, but some do not.
Second, the basic issue of focus or non-focus confronts the child from the
beginning of his entrance to the conceptual level, as he experiences the fact
that he needs to do something, to put forth effort, if he is to understand the
things with which he deals. From late infancy on, there is all the difference
in the world between the mentally active, focused child and the mentally
passive, unfocused one. That difference is one of direct, volitional choice.
It is not necessary for a child to have philosophic instruction or highly
abstract concepts in order to know in some form that it is good to know
things, and that learning things requires that he do something, that he try,
that he put forth effort.
In what specific form does the choice to focus arise in early childhood?
Picture a boy at age three or four, sitting on the floor playing with some
blocks. He is trying to stack them as high as he can. He has piled them about
a foot high, but when he places the next block on top, they all come crashing
down. This may be a frustrating, upsetting experience for him. Suppose he
feels a wave of negative emotions sweep over him. That emotional response is
automatic, but the volitional issue is: what does he then do? Does he dissolve
in tears, kick his blocks away, and run to Mother for solace, shunning blocks
from then on? Or does he, perhaps after a moment, take himself in hand,
look back at the blocks, and try again, thinking about what went wrong and
how to correct it?
Here, on a kindergarten level, the child confronts the issue of reason
vs. emotion, a directly volitional issue. According to how he chooses to use
his mind, according to whether he places his feelings above facts or facts
above feelings, he is making deeper a groove in his brain; he is forming his
character.
Now picture the boy at age seven at the dinner table. A thought has
occurred to him and he raises it: “Daddy, if God created the world, who
created God?” (Children do ask that.) Let’s suppose his father is the worst
kind of lout: he flies into a rage at the “blasphemous” question, and shouts
“You damn smart aleck! You’ll go to Hell if you ask that kind of thing!”
and he reaches across the table and slaps the boy.
The boy is hurt and confused. He bursts into tears. But the issue here is:
what goes on in his mind, after the heat of the moment is over? Does he think
about what happened, or just wallow in hurt, resentment, and self-doubt?
He is not a philosopher, he is only seven, but what can he already conclude,
 Volition and C haracter 343

if he chooses to think? He can conclude: “I only asked a question. A question

that made sense to me. Daddy didn’t answer my question, he only got mad
and hit me, for no good reason. I still want to know the answer to my
question.”
And most importantly, he can judge his father: “Daddy was bad to me;
he shouldn’t have yelled and hit me. He stops me when I hit my brother.
It can’t be wrong to ask questions.” As a practical matter, he can resolve
to keep quiet about this and other topics that might anger his father.
He can resolve not to say such things aloud, but to keep thinking them
in his own mind. The wonderful thing about thinking is that it is private.
Parents cannot hear what their child is thinking. They cannot control
whether he thinks or not. And, as a general rule, parents do not even seem to
know whether their child thinks or not.
According to whether the boy maintains the will to understand or drops
it, he is forming his character — not just in regard to each concrete, but in
relation to values as such. He builds the premise that the world is a welcom-
ing place, where good things await him, that life is an active, joyful adventure
— or the opposite premise, that the world is fraught with unpredictable
dangers, that values are always threatened, and that life is miserable.
One might think that from his father’s brutality the child in my example
would necessarily form a bad premise about people — that they are unpre-
dictable, dangerous. But this is not so, because the boy always has at least one
counter-instance refuting any such over-generalization. The counter-instance
is himself. If he judges his father’s action as wrong, he can, at the same time
think: “I would never do that.” And it is not only possible for him to judge
his parents, it is his moral responsibility to do so.
The difference between being mentally active and mentally passive makes
a huge difference in the developing psychology of the child. The active,
focused child builds self-esteem. He learns that he can figure things out, and
that gives him a sense of pride and confidence. He learns he can trust his
mind. He develops the sense that he is good — because he knows, wordlessly,
that he achieved his understanding volitionally, by trying hard, by doing
work with his mind.
By this gradual process, concrete by concrete, choice by choice, a child
forms over the days of his life conclusions about himself. He comes to regard
himself as competent or incompetent, good or bad. Which way he goes
depends on how he chooses, how he exercises his free will.188

188 For an extended discussion of the role of choice in a child’s cognitive and characterologi-
cal development, see Rand, “The Comprachicos.” [RP, 51–95]
 H OW W E K N OW • 1 0 : F ree W ill 344

Starting in earliest childhood and continuing throughout life, one does

indeed forge one’s own character, as one continuously chooses, in issues
large and small, to function rationally or irrationally.

Validation of the Theory

What is the basis for holding that focus is volitional? Introspection — the only
direct source of information about the nature and actions of consciousness.
My modus operandi in this chapter is to call to the reader’s attention and
name facts that he can directly introspect. The reader, presumably, recognizes
introspectively his own ability to monitor and direct his mind’s operations,
to seek full awareness, adopt the reality-orientation — in other words, to focus
his mind. Thus, my presentation of what focus is, has simultaneously been
a presentation of the evidence that focusing is directly volitional.
Comparing the in-focus state to the out-of-focus state presents a certain
difficulty: to the extent that one is out of focus, one is not i­ ntrospecting. But
upon coming into focus, it is easy enough to recall the preceding moment.
That is what one does when one catches oneself daydreaming. The difference
between a recalled state of non-focus and a present state of focus is striking
and undeniable: being alert, purposeful, actively in charge vs. being passive,
aimless, not in charge (or actively evading).
But there is no introspectible process of deliberation that occurs before
focusing. In fact, one has to already be in focus to be thinking about any-
thing, including about whether or not to focus. Focusing is therefore unlike
other choices, such as that between items on a restaurant menu. One simply
either puts one’s mind in gear or one does not.
What, then, can be said about the form in which one experiences the
specifically volitional nature of focus? Granted that there is a clear differ-
ence between being in focus and out of focus, how can one introspect one’s
volitional control over which state one functions in?
Volition is experienced directly in one’s sense of agency and effort.
One cannot avoid being aware of oneself as the active agent in the cognitive
process. Initiating and sustaining focus is something that one does, not some-
thing that just happens to one. Part of this sense of agency is the experi-
ence of effort. It takes an effort to think; it even takes an effort to engage in
purposeful sensory o ­ bservation. (“Effort” does not mean strain or suffering;
there is effort required to walk, talk, and eat, but these physical activities, like
the mental activity of seeking to know, are normally pleasant.)
 D eterminism : T he D enial of F ree W ill 345

Taking charge of one’s mind, setting the goal of full awareness, directing
one’s attention, asking oneself questions, making judg­ments — these are
things that one does, as opposed to things that happen to one, as when an idea
“pops into your mind” or a feeling “comes over you.” The effort to understand
a confusing issue stands in stark contrast to, for instance, a tune that one can-
not get out of one’s head. Supplying the effort required to understand an issue
is up to you to do, which is to say: it is volitional.
Consider the contrast with emotions. Emotions are an automatic response.
Feeling anger is something that happens to you, not something you do.
One says: “I am getting angry,” not: “I am doing angering.” Emotions are
effortless responses, based upon stored value-judgments which are triggered
super-rapidly by what one perceives. In contrast, naming one’s emotional
responses, identifying their causes, judging their actual appropriateness
to the facts of the situation — these cognitive acts do require effort. Such
reflective analysis does not occur automatically; one must choose to do
mental work.
To further distinguish the automatic from the effortful, consider the case
of memory. For instance, what happens when you read the following:
“Two plus two is . . .”? The word “four” is triggered automatically and effort-
lessly. But now try to remember what clothes you wore a week ago today.
You probably have to do work to retrieve the answer, if you even can retrieve it.
And you can stop trying at any time. But you cannot normally stop yourself
from hearing “four” in your mind when I triggered your memory of that
sum. Some recall is automatic, some takes effort, and we are very aware of
the difference.
These examples should make clear the manifest difference between the
automatic and the volitional, between that which enters your mind without
effort and that which you do in an act of effortful self-direction.189

Determinism: The Denial of Free Will

To say that man has free will is to say that man is capable of making genuine
choices, that the cognitive and existential actions he performs are up to
him, rather than being made necessary by antecedent factors. The opposite
theory, determinism, holds that choice is an illusion, that what we think of

189 Effort is a sign of volition, but not every expenditure of effort indicates a proper use
of volition; evasion, the antithesis of focus, also requires effort — the effort not to know.
 H OW W E K N OW • 1 0 : F ree W ill 346

as being up to us is actually necessitated by the factors, internal and/or external,

operating at the time of the alleged choice.
Man’s actions, determinists readily concede, have much more complex
causes than those governing the movement of inanimate objects, such as
a rock tumbling down a hillside. But, determinists hold, a man’s actions
are just as necessitated as the rock’s. After all, they observe, a computer
is also very complex, but no one is tempted to posit free will in the physi-
cal processes that go on inside the computer and produce its output.
Human action, mental and physical, is held to be similar: just “output” that is
pre-determined by the person’s makeup plus the conditions obtaining.
Determinism implies that if we knew all the facts, we could predict every-
thing about your future — what you will think, what you will feel, what you
will do. Determinism, however, is basically a metaphysical position, not one
concerning what we can or cannot predict. It is defined as the doctrine that
all man’s actions are necessitated by factors beyond his control.
Numerous attempts have been made throughout the history of philos­
ophy to reconcile free will and determinism, but no such reconciliation is
possible. There is no middle ground between: 1) the act of focusing one’s mind
is necessitated by antecedent factors, and 2) the act of focusing one’s mind
is not necessitated by antecedent factors.
Determinism and free will are contradictories; determinism is precisely
the denial of free will.190
Determinists claim that the introspective experience of free will is illusory.
“You feel that you choose freely,” they say, “but only because you are unaware
of the motives and reasons controlling your decision-making.” This idea has
an element of truth — when directed against theories that locate free will in
existential choices, such as whether to study or go to a movie. Such choices
are indeed not causal primaries. But the choice to focus or not is.
The evidence of free will is not that we deliberate but that to deliberate
rationally one must do things that take effort. Determin­ists cannot deny that
we experience agency and effort. Instead, they must argue that a determinis-
tic being could also experience effort and agency, a hard sell.
Rather than rebutting what the determinists might say on this topic, I turn
now to exposing the error in the main positive argument for determinism:
the false belief that free will would contradict the law of causality.
190 “Compatibilist” theories inevitably side with determinism, and re-define volition out of
existence, as in the claim that “free” means “uncoerced.” The absence of coercion is an
issue of social-political freedom, not of metaphysical freedom.
 D eterminism : T he D enial of F ree W ill 347

Causality and Volition

The popular belief that there is a conflict between free will and causality
stems from a mistaken conception of the law of causality. The proper view
of causality, originated by Plato and Aristotle, recognizes that causality
is a relation between the nature of an entity and its actions. An entity of
a given kind has the properties it has, which gives it certain potentialities
and no others. The actions possible to an entity are determined by its identity,
by what it is.
There is nothing in this proper understanding of causality to clash with
free will. Man, by virtue of his nature, has the potentiality of initiating a
process of rational thought, but he does not have to actualize that potentiality.
However, as far back as the 17th century, the proper, Platonic-Aristotelian
understanding of causality lost favor, and came to be supplanted by an
arbitrary construct: the notion that causality concerns “events,” not entities and
their actions, and that every event is a necessitated reaction to previous events.
Unfortunately, it was Galileo who popularized the new event-to-event view,
as historian Wilhelm Windelband notes:

. . . the idea of cause had acquired a completely new significance

through Galileo. According to the [preceding] scholastic concep-
tion . . . causes were substances or things, while effects, on the
other hand, were either their activities or were other substances
and things which were held to come about only by such activities:
this was the Platonic-Aristotelian conception of the aitia [causes].
Galileo, on the contrary, went back to the idea of the older
Greek thinkers who applied the causal relation only to the states
— that meant now to the motions of substances — not to the
Being [identity] of the substances themselves. Causes are motions,
and effects are motions. [Windelband, II, 410; emphasis deleted]

This notion of causality does lead to determinism. But the event-to-event

view is wrong. We do not encounter any such thing as free-floating “events”;
actions are actions of entities. The event-to-event model of causality cannot
be applied even to its proponent’s favorite case: billiard balls. When one
billiard ball collides with another and sets it in motion, the interaction
is causally determined by the nature of the entities involved, including their
 H OW W E K N OW • 1 0 : F ree W ill 348

state of motion. To see this, one need only consider the consequences if either
ball had a different identity. What if one or both balls had a different shape,
say a cylindrical one, or were made of gelatin, or were only a thin, brittle
shell, hollow inside? The actions resulting from collisions would be different
if the entities’ natures were different. Causality relates entity, identity, and
action — not event to event.
The identity of a material object consists not only of its state of motion
but also of the kind of material of which it is made, and the structure in
which that material is arranged — both the macro-structure, overall shape,
and micro-structure, down to the scale of molecules, atoms, and subatomic
particles. What a material object does in interaction with other material
entities depends on all these factors, on its (and their) total identity, not just
their state of motion.
Accordingly, the proper understanding of the law of causality is that the
actions of an entity are an expression of its identity; the interaction of entities
is an expression of the identity of each. What an entity can do is determined
by what it is. “The law of causality is the law of identity applied to action.”191
This understanding of causality represents an integration of three concepts
that are irreducible primaries: “entity,” “identity,” and “action.” We perceive
entities acting. We form concepts for their actions by observing them act.
We form “walking,” for instance, by observing men and animals walking.
The concept of “walking” includes the fact that legs are involved in the action.
Thus, we know that a billiard ball cannot walk — the action would be a viola-
tion of its identity, since it has no legs (and lacks many other requirements).
Or, “burning” is a concept we form from observing the burning of wood,
paper, and other combustible materials. This implies that the miracle of
“the burning bush that was not consumed” is incoherent: for a thing to burn
is for it to burn up — i.e., to be consumed.
At any given time, we may not know what features of an entity are
necessary for it to perform a given action. Pre-scientifically, for instance,
men did not know that combustion is the combination of an oxidant with
the material composing the burning substance. But from their first grasp
of causality as such, men know that the actions possible to an entity are
caused by facts — whether known or yet to be discovered — about the makeup,
the identity, of the entities involved.

191 [AS, 1037] Here I might be in disagreement with Rand’s view of causality, since I see it as
involving what an entity can do, and some of her statements are phrased in terms of what
an entity will do.
 D eterminism : T he D enial of F ree W ill 349

Next, by observation and induction, we discover that the material world

is governed by mechanical causation. Matter as such is inert; it cannot set
itself in motion. As scientific knowledge progresses, we learn that even the
self-generated actions of living organisms, in their physical (non-conscious)
aspects, are essentially like those of inanimate objects — i.e., are deterministic.
The difference is that living organisms possess an internal store of physical
energy, and their structure enables them to route that energy to power
different types of action.192
The motions of material objects, and of the material aspects of living
organisms, are subject to strict necessity: whatever they do, they had to do.
The same matter in the same circumstances will act the same way. But the
form of causation applicable to man’s consciousness is different — because
man’s consciousness is a different kind of phenomenon. The nature of man
causes him to have the power of choice — fundamentally, the choice to
focus his mind or not. By virtue of his makeup, including the makeup of his
nervous system, man has sovereign control over the operation of his concep-
tual faculty. When a man chooses to use that faculty to pursue conceptual
understanding, the action is not causeless — his choice is the cause. The same
applies when he lets himself drift passively or chooses actively to evade.
Free will is a specific form of causation, not a cancellation of causality,
a form of causation that arises from and depends upon the functioning of
one’s nervous system and body, as one interacts with the world.193
Can the choice to focus itself be explained causally? The answer depends
on what one means by “explaining the choice.” The very existence of a choice
between focus and non-focus can be explained causally: man, by virtue
of his makeup, has such a choice; lower animals, which lack the concep-
tual faculty, do not have this choice. But the specific outcome of the choice
— the fact that a given man chose to focus rather than not — can be explained
only in the sense that one focuses for a reason: in order to be fully aware.
But this reason is not a necessitating factor. It is only a potential reason;
the actualization of that potential takes an act of will. At root, the actuali­

192 See my discussion of “self-generated action” in Binswanger 1990, 40–54, where I conclude
that “an action is self-generated when it results from the utilization of an internal energy
source integral to the agent according to a directive mechanism.”
193 The relationship of mental activity to brain function is a difficult and largely scientific
issue. Philosophy can establish only two general points here: consciousness has causal
efficacy in relation to man’s body, and the choice to focus is volitional. But clearly
the mind depends causally upon the healthy functioning of the brain, and the entire
discussion of free will presupposes that we are considering normal, healthy brain states.
 H OW W E K N OW • 1 0 : F ree W ill 350

zation of that potential  — i.e., setting the goal of full awareness  — is what the
exercise of free will consists of.
What one cannot do is “explain” the choice to focus (or not to) in the sense
of specifying some antecedent factor that made one choose the way one did.
This fact constitutes a philosophical problem only for those who equate
causality with necessitation. And there is no way to justify that equation.
Taking causality to mean necessitation is an over-generalization from the
kind of causation applicable to matter. But consciousness is not matter;
it is not inert, but active.194 To insist that consciousness must be governed by
the specific form of causation exhibited by matter is to approach man with
an arbitrary, a priori commitment to materialism.
In an extemporaneous discussion of this issue, Rand said:

The appearance of a conflict between causality and free will is

due to taking causality to be only that which governs the material
world. Consciousness is an existent having a nature different from
that of matter. The law of causality implies, accordingly, that the
type of action consciousness can take will be different.195

Free will means choice, and choice is a phenomenon of consciousness.

The nature of consciousness is different from the nature of matter, and the
law of causality says that different natures entail different forms of action.
Consciousness, on the conceptual level, is self-initiated. The nature of con-
ceptual consciousness is to be volitional, not deterministic — i.e., to be able to
activate its full resources, to set awareness as its goal. This is not an exception
to, but an instance of, the law of causality. The law of causality does not ­prohibit
the existence of choice, of self-initiated action, of starting a causal chain.
In the case of volition, one starts a causal chain by willing the state of full
awareness, as a “first cause” within consciousness.

194 Why is free will possible to consciousness but not to matter? I suspect that the difference
is due to the fact that consciousness is an organic whole, without parts but with a rich
“internal” identity. Material objects are ultimately collections of “atoms” — using this term
broadly for whatever are the ultimate constituents of material things. These ultimate
“atoms” of matter, having no parts, can have no internal structure. If so, then they can have
no unactualized potentialities and can only go on doing that which they have always
done. Material objects, being nothing but arrangements of these “atoms,” could not, then,
act differently in the same circumstances. But consciousness is not a compound of parts;
it is an organic whole having a complex identity, with many unactualized potentialities.
195 From my notes of a personal conversation with Ayn Rand, circa 1980. (I can vouch for
the exact wording of only the first sentence.)
 D eterminism : T he D enial of F ree W ill 351

(Furthermore, the law of causality entails that causes, once instituted,

have necessary effects. This applies to volitional acts as well: initiating rational,
focused cognition necessarily produces different effects, existentially and­
psy­cho­logically, from those produced by drift or evasion.)
The exercise of reason is a causal primary.196 A “causal primary” is a factor
that 1) determines a series of effects, and 2) is not, in turn, determined by
any prior factor. The exercise of reason exhibits both of these relationships:
1) the rationality (or irrationality) of a man’s mental processes determines
the nature of his conclusions, his decisions, his actions, his character, and his
psychology; 2) no prior factor determines whether or when he will function
rationally or irrationally — that is a matter of his direct volition.
Another question arises in regard to integrating volition with causality.
In the case of material objects, the same thing put in the same circumstances
will act the same way. Does this apply to man? Can we say that a given man
placed again in the same circumstances must use his mind in the same way
he did before — assuming, for the sake of argument, that both the external
and the internal, psychological, circumstances could be absolutely identical?
In one sense, the man must act the same way: he must choose. He cannot
shed his volitional nature; as a human being, he cannot avoid the ongoing
choice of how to use his mind: rationally or irrationally. But, contrary to
what I formerly believed [Binswanger 1999], this does not erase the fact that,
from a more concrete perspective, what he chooses can differ. Analogously,
if one man votes Republican and another votes Democratic, they are, on an
abstract level, “doing the same thing”: voting. But, more concretely, they are
doing different things: voting Republican and voting Democratic.
Secondly, we must not reify the “choosing” that is involved in the choice
to focus. The choice to focus is not like the choice of what shirt to wear: there
is no process of deliberation (which itself would have to be done in focus or
out of focus). In the case of focusing, the “choosing” is not a separate process:
one simply either focuses by choice or one does not, again by choice. The
phrase “by choice” expresses the volitional nature of either action; it does not
mean that the two actions are “really” the same.197
Again, the applicable type of causality depends on the entity’s nature:
“same conditions, same outcome” applies to inanimate matter but not to man.

196 I owe this way of putting it to Allan Gotthelf (personal conversation). Some other causal
primaries are: existence, identity, and the nature of the ultimate constituents of matter.
197 I am indebted to Lee Pierson for pressing me on this issue.
 H OW W E K N OW • 1 0 : F ree W ill 352

Focus and “The Problem of Agency”

One recurrent objection to free will is known as “the problem of agency.”
According to this objection, a being with free will would, paradoxically,
not be a rational agent, would not be the author of his own acts. As one
determinist put it: “Either our wills are determined by prior causes . . . or they
are the product of chance and we are not responsible for them.” [Harris, A15]
A rational agent’s existential actions are caused by the content of his mind,
by what he thinks. But, the objection holds, free will would be something
that breaks the connection between thought and action, thereby preventing
one from acting as a rational agent. Personal responsibility, this objection
concludes, requires determinism: acts for which one is responsible are those
caused by one’s deliberative reasoning, not by a blind act of “willing.”
This objection, however, is not applicable to Ayn Rand’s theory of free will:
the objection evaporates once one recognizes that “will” consists in choos-
ing to engage in and implement a rational process. The will is not some
blind force coming in from the dark, disrupting the connection between
one’s thinking and one’s action. Rather, one wills to institute rationality
— or one does not. The “will” is one’s capacity to generate and sustain rational
thought. One is responsible for one’s actions, cognitive and existential,
precisely because one is responsible for the state of mind — in-focus or out-
of-focus — that caused those actions. (And if we could not control the state
of our minds, if what our minds churned out were necessitated by some
antecedent factor, we would not be self-responsible beings.)
It was this wrong view of the will that led A. J. Ayer to proclaim that:
“if it is a matter of pure chance that a man should act in one way rather than
another, he may be free but can hardly be responsible.” [Ayer, 278] But it is
choice, not “chance,” that governs man’s actions — the choice to generate
a state of full focus or to default on that responsibility.
In notes Ayn Rand made in her personal journal, at age 29, she wrote:

The will does not have to be without reason, or motivation,

in order to be free. One’s act may be motivated by an outside
reason, but the choice of that reason is our free will. . . . Doesn’t the
“free will” question come under the general question of human
reason — and its freedom? . . . Does a free action necessarily mean
an unreasonable one? [JAR, 68–69]
 D eterminism : T he D enial of F ree W ill 353

When one chooses to take a certain action, one has a reason for what
one chooses. The choice is not blind or causeless. But that reason does not
force itself upon one; its presence in one’s mind and its motivational power
result from how one is using one’s mind — i.e., from one’s level of awareness.
By choosing the level of awareness on which one functions, one chooses what
kinds of reasons will become operative.
The example of the two students dramatized this. Each student had
a reason, in the sense of mental content, that explained why he acted as he
did. The first student’s reason for not studying was to see an enjoyable movie
and avoid what he felt, while out of focus, would be a painful evening of
studying. The second student’s reason for studying was to get a better grade
in the course and thus improve his chances of getting into a good graduate
school. Different reasons governed their decisions because they exercised
their free will differently: the first student engaged in out-of-focus, associa-
tional wandering, which kept the long-range consequences out of his mind;
the second student took charge of his mind, directing it to the task of finding
out the full truth, which included the long-range consequences.
Each student’s choice was free, not determined, because it was shaped by
his chosen way of using (or misusing) his mind. The decision each made
did not “just happen.” It was not a “chance” outcome. The decision of each
was based on what was in his mind, which in turn resulted from his level of
awareness. Having free will did not mean that either student acted without
a reason, but rather that each chose whether or not to have a rational reason.
Does the determinist have a comeback? Can he say about focus itself:
either you focus for a reason, in which case focus is determined by that rea-
son, or you focus for no reason, in which case, it is arbitrary and irrational?
No. There is always a reason to focus, because focus means the acceptance
of reality. Reality is always there and always needs to be focused on.
But reality does not have the power to make one focus on it. The point can
be stated this way: there is always a potential reason to be in focus — namely:
to know reality and guide one’s actions by reference to one’s grasp of the facts.
But that potential reason is not actualized automatically; one has to make
that reason be operative, by a self-initiated choice.
Moreover, there can be “reasons” (i.e., motives) for not focusing. You can
feel at a given moment that full awareness would be a bad thing, requiring
you to face something you don’t want to face. But you cannot make a rational
assessment of the consequences of facing something until you face it — i.e.,
until you choose to focus.
 H OW W E K N OW • 1 0 : F ree W ill 354

When you are out of focus, focusing can feel like a very bad idea. But your
choice is whether to accept that feeling or to go by reason — i.e., to focus or not.
The fact that something — even focus itself — prompts a feeling of aversion
cannot dictate whether or not you go by feelings. And that is what the choice
to focus is: the choice to go by facts, not by feelings.
So it is absurd to say, as some determinists would, that one’s feelings deter-
mine whether or not one goes by feelings. Just as it is absurd to say, as other
determinists would, that the facts determine whether one goes by facts. If
that were true, no one would ever be irrational.
Focus is precisely the choice of which will govern how one proceeds:
facts or feelings. The inescapable issue is: which do you put in charge — your
best perception of the facts or your feelings? Neither facts nor feelings are
automatically in charge. The question is which side has authority, and you
have to decide that; it is up to you.
Neither reality nor your feelings is the equivalent of the Supreme Court
here. Neither reality nor your feelings can seize control of your mind.
You are the Supreme Court determining which factor, reason or feelings,
governs how you proceed.
Thus, the determinist cannot re-apply his argument about “the problem
of agency” to the issue of focus itself. He cannot claim that focusing would
be irrational unless the (potential) reasons to be in focus made one focus.
The choice to exercise reason is not irrational. The choice to accept reality is
not arbitrary. “Arbitrary” means: not based on the facts of reality. If the very
acceptance of reality is called “arbitrary,” then the concept of “arbitrary” loses
all meaning, becoming a stolen concept.
The acceptance of reality is not necessitated by anything prior; we are
neither forced to function rationally nor forced not to.
The determinist objection that free will would be a blind lurch is directed
against a straw man. The actual target of that objection is not free will but
the doctrine known as “indeterminism.” Indeterminism holds that human
action is not necessitated because some actions allegedly have no causes at
all. According to this notion, in certain cases it is a sheer, causeless acci-
dent — Ayer’s “chance” — which of two actions a man performs. But choice
is not chance or an Epicurean Swerve — i.e., free will is not indeterminism.
Nor would indeterminism do man any good. Whether a man’s life were
ruled by iron necessity, or by a necessity interrupted by freak accidents,
man would not be in control of himself. Free will is not non-causation but
a form of causation: self-determination.
 D eterminism : T he D enial of F ree W ill 355

Volition as Axiomatic
Volition is not only self-evident — directly introspectible — it is fundamental
to conceptual cognition. One’s volition controls the extent to which concep-
tual cognition exists, and the alternative of whether conceptual functioning
exists or does not exist is a fundamental alternative.
(A mind completely out of focus is still operating “conceptually” in the sense
that words go through such a mind, and some minimal conceptual content
will necessarily be involved, due to the triggering of automatized connections.
But out-of-focus, associational wandering, though beyond the capabilities
of an animal, is not cognition. Cognition is the pursuit of knowledge of reality;
striving to know, adopting the reality-orientation, setting the goal of full
awareness — these are precisely the things that require volitional activation.)
As both self-evident and fundamental to the conceptual level, volition is
an axiom of the conceptual level. Axioms cannot be proved to be true, since
they are the very basis of proof, but granted their truth, one can demonstrate
that they are axioms rather than derivative truths. As discussed in Chapter 1,
the demonstration that a truth has the status of an axiom consists in showing
that it is inescapable — that it is, in Rand’s words, “a statement necessarily
contained in all others, whether any particular speaker chooses to identify
it or not.” [AS, 1040] This means applying the test of “re-affirmation through
denial,” as I did in Chapter 1, for the axioms of “existence” and “conscious-
ness.” I will now show how this test applies in the case of volition: how even
those attempting to deny free will have to covertly assume it.

Determinism as Self-Refuting
One’s volitional control over one’s conceptual faculty indeed passes the test of
re-affirmation through denial. Suppose someone states, “I do not have voli-
tional control over my mind.” If so, then he cannot claim that that statement
is true or represents knowledge, only that whatever runs his mind forces him
to believe it. He has to assume he is free to consider the facts on the subject
of free will vs. determinism, in the very process of denying he is free to do so.
Freud, an arch-determinist, quoted approvingly the statement of Georg
Groddeck: “We are ‘lived’ by unknown and uncontrollable forces.” [Freud, 13]
That statement, if true, would itself be the product of unknown and uncon-
trollable forces — and thus not entitled to rational consideration.
As with any assertion of determinism, the claim commits the fallacy
of self-exclusion. To include itself would mean asserting: “I am ‘lived’ by
 H OW W E K N OW • 1 0 : F ree W ill 356

unknown and uncontrollable forces.” Anyone saying that, not as a joke but as
a serious report of his condition, would be recognized as psychotic (and pos-
sibly dangerous). All forms and varieties of determinism negate themselves,
because the very essence of determinism is to assert that one has no control
over the operations of his own mind.
A determinist cannot escape this self-invalidation by maintaining that
what runs his mind are facts and logic. How, on the determinist premise,
could he know that this was the case? After all, other men reach different
conclusions, opposite to his, on many subjects. He is claiming to be, in effect,
programmed to think in a certain way. How can he assess the validity of his
programming (vs. the invalidity of others’ programs)? Any attempt at validat-
ing his “programming” is doomed to failure. If he asserts that his program­
ming is validated by a given test, that will merely raise the question: how do
you know that? A deterministic mind could be programmed to announce
that it passed a test when it actually failed, or to accept an invalid test as valid.
And any further sentences he utters to defend against this objection would
simply raise the same question over again: what makes you say that that
is logical, if whatever you say is merely a reaction to the forces impinging on
you at the moment?
The determinist is asserting that he cannot be objective — that his mind is
in the grip of something that runs it. If so, then he also cannot be objective
about his meta-level beliefs — i.e., beliefs about what determines him and
what its nature is. By his own theory, he cannot judge either facts or his own
judgments. Instead of judgment, there is only stimulus and response.
A volitional, self-regulating mind can directly will itself to focus on reality,
to work honestly to integrate its conclusions into the full context of its knowl­
edge, without any factor censoring it. If a man knows the basic rules of logic,
he can choose directly to employ them and thus be objective. No feelings
or external factors can keep him from engaging in rational thought and
checking, without prejudice, the validity of his thinking.198
A volitional mind is never forced to be irrational, forced to deviate from
observations, forced into fantasy. It always retains the power to exclude
such invalidating conditions by a direct act of will. A deterministic mind,
in contrast, could not exclude invalidating conditions.

198 It is irrelevant that in extreme circumstances emotions can make rational thinking
temporarily impossible. It is easy for one to recognize that one is in that kind of state
(e.g., a panic) and to wait for a calmer moment in which one can resume rational thought.
 D eterminism : T he D enial of F ree W ill 357

Consider a determinist who says that what runs his mind is logic. Logic
is not an entity that can exert forces on a brain. In contrast, when one sees
an apple, light reflected from the apple physically stimulates the receptors
in one’s retina. But no such object-generated causation is available on the
conceptual level. There is no such thing as Logic Itself, Truth Itself, or Reason
Itself to act upon one.
Nor is logic some set of concrete operations that could be wired into the
brain. Logic is a process that adheres to reality by considering all and only
the evidence. But a deterministic mind would “output” whatever was neces-
sitated by the biased set of material, real or imaginary, that the determining
factor permitted entry into its consciousness. A deterministic mind would
be operating under constant censorship.
Thus a deterministic mind would be in the situation of a man forced at
gunpoint to read aloud what is written on a piece of paper shoved in front
of him. His utterances, in that case, would be cognitively valueless. Suppose
a bystander asks the victim, “How do you know that what you are forced
to say represents a valid, logical identification?” The gunman writes some-
thing on a piece of paper, shoves it over to his victim and orders him to read,
“I know it because . . .” It doesn’t matter what the next words he reads are:
it is still without cognitive content for the victim. That is the dilemma that
man would always be trapped in, under the assumption of determinism.
The concept of “logic” cannot be applied to a deterministic consciousness.
Logic requires independent judgment — independent of any prejudicial,
potentially censoring factor. The volitional ability to function rationally is
required to attain that independence. Free will means having the power to
choose to consider whatever is relevant, and hence to attain objectivity.
Determinism is actually a secularized version of Descartes’ “evil demon”:

I will now suppose . . . an evil spirit, as powerful as he is ­cunning

and deceitful, who has employed all his powers to deceive me;
I will suppose that the sky, the air, the earth, colors, shapes,
sounds, and all the other external things that we perceive, are only
illusions which he uses to snare my credulity. [Descartes, 62–63]

Once this fantasy is (arbitrarily) adopted, there is no way out. It is no

answer to say, “God would not allow such a being to radically deceive me”
(Descartes) or “The demon forces me to believe only those ideas that are in
fact true,” or “The demon forces me to be rational and logical.” In each case,
 H OW W E K N OW • 1 0 : F ree W ill 358

there is no answer to the reply: “That is just another false idea fed to you by
the demon.” Or, putting it another way, all such assertions are arbitrary.
Once one has asserted, “The things I think and say are fed to me without
my being able to reject them or independently test them,” there is no refuge.
At the deepest level, determinism implies not just that there is no such
thing as knowledge, but that there is no such thing as belief, assertion,
judgment — no such thing as even claiming truth for an idea. To believe,
assert, or judge that something is the case involves committing one’s
consciousness to the reality of that thing, and that commitment takes an
act of will. A computer-like consciousness would output whatever it had to
output, but it would be incapable of making a commitment, across time,
to anything being true: what it will output tomorrow will be necessitated by
the state tomorrow of the factors that run it. A free, self-controlled mind,
in contrast, can resolve to hold to a conclusion over a lifetime (provided that
the evidence continues to support it).
On every front and from every perspective, determinism is inconsistent
with its own assertion. The attempt to deny volitional control over one’s mind
has to covertly assume it. Thus, volition passes the test of reaffirmation
through denial and is indeed an axiom: a primary, self-evident fact, standing
at the base of knowledge.199 “Choice” is an axiomatic concept.
The debate between the advocates of free will and the advocates of deter-
minism comes down to the issue: which mental functions are automatic?
In that regard, the optical metaphor behind “focus” can be misleading:
optical focus is a physical state, specifiable in terms of the geometry of the
setup. Optical focus can be achieved automatically, as indeed it is for self-
focusing cameras. But mental focus is not some concretely specifiable state.
To be mentally in focus is to be committed to full awareness, and that means
being committed to performing whatever sequence of specific processes one
knows of and judges to be logically required. In one situation the process
required might be further perceptual observation, in another situation it
might be disengaging to do some creative mulling, in another seeking to
find an analogy, in another looking for a counter-example, or subsuming
something under a known rule, or trying new combinations, making new
generalizations, seeking examples — and so on, for all the various cognitive
strategies one knows.

199 Volition is an axiom of conceptual knowledge; perceptual knowledge, such as animals

possess, is automatic, not volitional.
 D eterminism : T he D enial of F ree W ill 359

Thus, a mind’s being “in focus” is like a violinist playing “with virtuosity”
— neither state consists of some concretely specifiable sequence of movements.
There is no mechanical formula or algorithm whose execution constitutes
rationality. Thus, one cannot ask, “Why couldn’t the brain be wired so as
to be in focus?” The brain could (in principle) be wired to proceed quickly,
or with a lot of circuits activated, but there’s no such thing as a wiring for
“proceeding rationally.” There are no physically specifiable neural conditions
for using one’s best judgment, for finding and adhering to the evidence,
nor for intellectual honesty and integrity. There is and can be no neural
switch automatically thrown to institute rationality.
The foregoing contradiction in the advocacy of determinism does not
constitute a proof of volition. (The existence of such a thing as proof pre-
supposes that one’s mind is free to form independent judgments.) Rather, it
demonstrates that volition, whose existence we already know from introspection,
is cognitively inescapable, and hence has the status of an axiom of the
conceptual level.

Free Will and Social Environment

Locating free will in the choice to be rational enables us to answer the follow-
ing common objection to free-will theories. If free will exists, the objection
goes, why do we observe a statistical correlation between people’s ideas and
their social environment? For instance, the vast majority of Catholics are
children of Catholics; most people come to hold the same political views as
their parents and/or social subculture; voting patterns tend to occur in blocs.
If choice reigns, why these correlations of beliefs with social environment?
Such correlations, however, are not evidence of social determinism but of
conformity. When one fails to think for oneself, one still has to know what
ideas to accept, what values to pursue, and how to act. The easiest and most
common “solution” is to conform to the beliefs and values of those around
one. Following the beliefs, values, customs, and taboos of others appears,
to many, to be an easy substitute for thinking things through on their own.
Thus, conformity is a result of non-thinking.
Further, note that the percentage of the population that seeks refuge in
conformity tends to vary directly with the difficulty of the thinking that
would be required in order to function independently: more people are
willing to think for themselves about which brand of car to buy than which
religion or philosophy to accept.
 H OW W E K N OW • 1 0 : F ree W ill 360

Rand observes:

. . . a social environment can neither force a man to think nor

prevent him from thinking. But a social environment can offer
incentives or impediments; it can make the exercise of one’s
rational faculty either easier or harder; it can encourage thinking
and penalize evasion or vice versa. [VOR, 102]

Educational innovator Maria Montessori, who worked extensively with

children, reaches the same conclusion:

Environment is undoubtedly a secondary factor in the phenom-

ena of life; it can modify in that it can help or hinder, but it can
never create. [Montessori 1912, 105, emphasis deleted]

Free Will and the Ego

Across a lifetime, from early childhood on, the proper use of free will consists
of the active approach eloquently concretized by Rand in this passage:

The process of concept-formation does not consist merely of

grasping a few simple abstractions, such as “chair,” “table,” “hot,”
“cold,” and of learning to speak. It consists of a method of using
one’s consciousness, best designated by the term ­“conceptualizing.”
It is not a passive state of registering random impressions.
It is an actively sustained process of identifying one’s impressions
in conceptual terms, of integrating every event and every obser-
vation into a conceptual context, of grasping relationships,
differences, similarities in one’s perceptual material and of
abstracting them into new concepts, of drawing inferences,
of making deductions, of reaching conclusions, of asking
new questions and discovering new answers and expanding
one’s knowledge into an ever-growing sum. [VOS, 21–22]

Actualizing the full potential of the conceptual faculty in that way is what
a life lived in full focus consists of, mentally.
 F ree W ill and the E go 361

The commitment to full awareness is the opposite of a duty imposed on

you: it is thoroughly self-interested. Exercising the will to know reality is
a corollary and expression of a metaphysical motive, a motive that transcends
immediate circumstances and the vicissitudes of concrete events: the desire
to live. From the first volitional act in early childhood until the end of your
life, the need and value of awareness are applicable and obvious. You turn on
your mind with exactly the same motive that impels you to open your eyes:
to be aware, so that you can live.
This means that, from another perspective, the choice to focus is directly
the choice to live. Here I am not referring to the fact that in order to maintain
and further your life, you have to choose to use your mind, because other-
wise you simply won’t know how to act. I am now making a deeper point:
the choice to focus is the choice to be fully alive in that moment. The choice
to use your mind is the choice to value yourself and fully to live.
In notes she made for Atlas Shrugged, Ayn Rand expresses this point:

Your self is your mind, and its constant choice is the act of self-
affirmation or self-denial, of perceiving or refusing to perceive,
the act of being or non-being by which your mind, like a pilot-light
within you, goes on or off. [JAR, 660]

It is by choosing to be fully aware that one chooses to be fully alive, alive

in the human sense.
Socrates said: “The unexamined life is not worth living.” I am adding:
The unfocused life is not truly lived.
H OW W E K N OW 362
 OVERVIEW
11
T his book upholds an Aristotelian, not a Platonic, approach
to knowledge, and to consciousness in general. The Aristotelian approach
recognizes that consciousness is action and is for action: consciousness
consists of mental processes, and its function is to guide existential action.
Consciousness is the faculty of awareness — of seeing, hear­ing, knowing,
remembering, etc. There is nothing supernatural or other-worldly about
a hawk’s ability to see its prey or a dog’s ability to follow a scent. The same is
true of man’s ability to form and use concepts. The Platonic/religious view
makes man’s conceptual faculty into a metaphysical outcast, alien to the natu-
ral world, and seeking absorption in another. But neither animal nor human
consciousness is an emigrant from a World of Forms or a Heaven. Homo
sapiens evolved from certain primates, and his consciousness is an enhance-
ment of, not a departure from, animal consciousness.
The human brain is a mammalian brain, but with a greatly expanded
capacity and with certain crucial add-ons, relating to language-usage and
manual dexterity. Man is the rational animal, the animal who conceptualizes
perceptual material. It is concepts that have enabled man to split the atom
and walk on the moon. And radically opposed theories about the nature
of concepts have created the battleground for the millennia-long clash
 H OW W E K N OW • 1 1 : Overview 364

between Platonism and Aristotelianism, between mysticism and science,

religion and secularism. Behind these polarities lie two opposed ideas about
concepts — about what concepts are and what they refer to.
For Plato, there is a metaphysical gap, a “divided line,” separating the con-
ceptual level from the perceptual level, human consciousness from animal
consciousness. Conceptual knowledge, Plato held, is directed to a separate
and higher reality, not the concretes of this world.
For Aristotle, the divide is epistemological not metaphysical: there is only
one reality, not two, and concepts refer to the things in this world, the world
given in sensory perception.
Aristotle got it right. Human consciousness is an enhancement of, not an
alternative to, animal consciousness. Animal consciousness is perceptual:
it provides an awareness of entities in the world. The world impinges on an
animal’s senses, and this sensory input is processed by its nervous system,
resulting in the animal’s perception of the array of entities in its environment.
And the animal is capable of learning: it forms associational expectations,
based on pleasurable and painful experiences. These expectations motivate
and guide the animal’s behavior in the world.
The sequence is a continuing cycle: the animal moves through the world,
acts on the world, perceives the effects of that action, and adjusts its next
actions on that basis. Schematically, the causal sequence for animals is:

perceptual
perception association emotion
sensory processing bodily action
world world
Human consciousness is the same, but with a crucial addition: concepts:

conceptual
perception integration emotion
sensory processing bodily action
world world

(“Perceptual association” here is meant to include any (non-conceptual)

proc­essing of perceptual material. “Emotion” is a stand-in for any affective
or motivational element. “Conceptual integration” includes every conceptual
process: concept-forma­tion, propositional judgment, and inference, since
each involves integration: grasping relationships and making connections.)
 C oncepts C ondense P ercepts 365

Concepts Condense Percepts

The conceptual faculty cannot originate its own content. There is no such
thing as “intellectual vision” or “intuition”: ultimately, the only source of
in­formation about the world is sensory perception. A concept represents
a condensation of perceptual data. When higher-level concepts are formed,
they, in turn, condense a number of earlier concepts — i.e., they condense
condensations of perceptual data.
“Conceptualization,” Rand observes, “is a method of expanding man’s con-
sciousness by reducing the number of its content’s units . . . concepts repre-
sent condensations of knowledge.” [ITOE, 64–65; emphasis deleted]
Concepts are formed not by “insight” but by interrelating — which means:
by perceiving similarities and differences. All the knowledge stored in any
concept was obtained, directly or indirectly, from perceptual observation.
In the case of higher-level concepts, the units are not perceptual concretes
but prior concepts; nonetheless, higher-level concepts must be reducible,
by a chain of intermediate concepts, to perceptual concretes. (If it cannot be
reduced to perception, it is invalid.)
Although concepts are integrations of percepts, concepts are not merely
words standing for a series of perceptual images, as the Nominalist theory
holds. “Table” is not something whose content is exhausted by the image of
a table or by several such images. A concept is integrative, not associational.
It integrates all the concretes of a given kind into a single mental unit.
The basis of concept-formation is the awareness of similarities — i.e., of
measurement-proximities. Tables, for instance, have shape-measurements
that are relatively close to each other when contrasted to the shapes of chairs,
sofas, and beds. In the case of first-level concepts, the measurement-proxim-
ities are perceived directly, in the form of similarity against difference.
Similarities that are abstract, and thus not perceptually available, are used
in forming higher-level concepts. But these concepts still depend on percep-
tion: they are graspable only on the basis of having first formed the required
lower-level concepts. The entire hierarchy is based on and reducible to simi-
larities that are given perceptually.
The cognitive value of concepts lies in the fact that, as condensations, they
provide unit-economy. Contra Plato, concepts do not have a different object
from that of percepts; rather, concepts are a way of dealing with perceptual
concretes wholesale. A shirt wholesaler still deals in shirts, but not one at
a time. Likewise, the concept “table” still deals with tables — not one at a time,
but wholesale.
 H OW W E K N OW • 1 1 : Overview 366

The shirt wholesaler, however, deals merely with groups of shirts bundled
together, say ten in a bundle. The bundle has the summed bulk and weight
of the ten shirts in it. The shirt-bundling achieves some condensation, but
very little compared to the astronomical condensation afforded by a concept,
which condenses an unlimited amount of information into a single word.
The concept “man,” for instance, integrates all the men who have ever existed
or will ever exist and stores all the knowledge applicable to them.
Concepts are tools for organizing and condensing perceptual data — for
the purpose of dealing more effectively with perceptual reality.
This perception-based, action-oriented understanding of cognition is
the polar opposite of the Platonic approach, which declares that concepts
are not based on percepts and are not directed toward the world that we
perceive. Instead, Platonists treat concepts as static revelations of another,
“higher” realm of pure abstractions, whose contemplation is an end in itself,
a view that underlies the theory-practice dichotomy, the elevation of “pure”
science above applied science, and the widespread contempt for philosophy
as “ivory-tower speculation.”

Two Primacies
The root of the clash between Aristotle and Plato lies in their opposed views
on a fundamental: the relation of concepts to perception. Aristotelians
uphold the primacy of perception: the recognition that perception is the
base of concepts, and concepts are abstracted from perceptual material.
(The raw data of introspection is understood as included in “perception,”
in that both are direct, pre-conceptual awareness.) Platonists, in stark
contrast, assert the primacy of concepts, claiming that some or all concepts
are not based upon perceptual observation. Concepts, Platonists hold, are
essentially independent of perceptual data and reflect an incommunicable
form of awareness of another reality. (Aristotelians recognize that there is
only one reality, the one we perceive by our senses.)
In Chapter 1, I explained the primacy of existence vs. the primacy of
consciousness. A philosopher’s stand on perception and concepts grows out
of his wider view of consciousness and its relation to existence. To make
these two primacies clear and lay out all the different positions, I offer the
following two tables.
 T wo P rimacies 367

The Primacy of Existence The Primacy of Consciousness

Metaphysics Existence is independent Consciousness is independent
of consciousness. of existence.
Consciousness is dependent Existence is dependent
on existence. on consciousness.
Epistemology Existence must be known Consciousness can be known
before consciousness is known. before existence is known.
Knowledge of existence Knowledge of existence
is gained by extrospection. is gained by introspection.

The second primacy has a similar structure:

The Primacy of Perception The Primacy of Concepts

Metaphysics Perception is independent Concepts are independent
of concepts. of perception.
Concepts are dependent Perception is dependent
on perception. on concepts.
Epistemology Perception is the given. Concepts are the given.
Perception is the standard for Concepts are the standard for
judging the conceptual. judging the perceptual.

Two points of clarification are needed in regard to the second table. First,
the primacy of concepts metaphysics is described as holding “Perception
is dependent on concepts.” Only extreme Platonists hold that the existence
of percepts depends on concepts; but it is commonly held that the identity
of percepts depends on concepts. E.g., Kant claims that perception is shaped
by “categories of the intuition,” and in contemporary jargon, perception is
“theory-laden.” The result in either case is viewing perception as distorted,
biased, “merely relative to us,” or not of “things as they are in themselves.”
It is true that after concepts have been formed and automatized, they get
automatically integrated with one’s perceptions, as Rand notes:

. . . you cannot perceive a table as an infant perceives it — as

a mysteri­ous object with four legs. You perceive it as a table, i.e.,
a man-made piece of furniture, serving a certain purpose belonging
 H OW W E K N OW • 1 1 : Overview 368

to a human habitation, etc.; you cannot separate these attributes

from your sight of the table, you experience it as a single, indivis-
ible percept — yet all you see is a four-legged object; the rest is an
auto­matized integration of a vast amount of conceptual know-
ledge which, at one time, you had to learn bit by bit. [RP, 55–56]

This overlay of conceptual content supplements perception rather than

distorting it — or even changing it, qua perception. Having the concept “table”
does not change any table’s seen color, shape, or number of legs. The theories
a scientist holds do not alter what dial-reading he sees on his instrument.
Thus, it remains true that perception is the base of and standard for judging
all further cognition.
In the lower table, under “Epistemology,” I contrast “Perception is the given”
with “Concepts are the given.” The given is that mental content which is not
shaped by our choices, so the issue here is: which mental activities involve
choice and which do not? The primacy of perception view holds that only
conceptual functioning is volitional. The primacy of concepts view denies
that conceptual functioning is volitional, and hold that some factor (God,
innate ideas, genes, “conditioning”) implants conceptual knowledge into
a passive mind. Thus, they are led to regard certain intellectual conclusions
as having the status of unquestionable givens.
(The primacy of concepts view implies also that perception is volitional.
This bizarre notion is expressed in the widespread belief that the senses
are capable of “deceiving” us, as if the sensory systems, like a journalist,
could choose what to present. Those attacking perception often refer to the
senses as “giving testimony” and being “lying witnesses” — terms that imply
volition.)
Finally, let me compare the two correct positions. The primacy of existence
primarily concerns metaphysics; the primacy of perception primarily con-
cerns epistemology. The primacy of existence is a metaphysical principle,
with epistemological implications; the primacy of perception is an epistemo-
logical principle, with metaphysical underpinnings.
The primacy of existence is essentially the fact that existence exists
independently of consciousness, with consciousness being dependent on
existence. The primacy of perception is essentially the fact that perception
exists independently of concepts, with concepts being dependent on percepts.
The primacy of perception is, in effect, an elaboration of a point included
in the primacy of existence: knowledge of existence is gained by extrospec-
tion. Extrospection means sensory perception — followed by the conceptual
 C oncept- F ormation in S cience 369

integration of perceptual material. Existence can be known only through

physical contact with it, the contact that sense organs provide. Perceptual
material is what results when physical existents interact with a consciousness’
physical means: the sense organs. Perception is thus automatically tied to
existence; but conceptual processes, being volitional, are fundamentally dif-
ferent: they can depart from or contradict existence — and will do so unless
one guides them in the required way.
The Platonic approach, in contrast, models concepts on percepts, claiming
that we “just see” abstract truth with “the mind’s eye.” People holding ideas
different from one’s own are regarded as intellectually blind.
(Here, Plato effectively demotes those who disagree with him to the level
of subhumans. Plato holds that those who persist in disagreeing with the
ideas of the “philosopher-king” are “incurably corrupt in mind” and should
be put to death [Republic, II, IX, 5]. The same attitude is shown by religious
fundamentalists who punish “blasphemy” and even murder “infidels.”)
The primacy of perception does not in any way deny the importance of
concepts — i.e., it is not the same as the “empiricism” of Hobbes, Hume,
or the Pragmatists. In fact, the opposite is the case: recognizing the primacy
of perception is what enables us to vindicate the conceptual level, to show
how it is based on facts and how it serves man’s life, as I have done in the
preceding chapters, based on Ayn Rand’s theory of concepts.

Concept-Formation in Science
The primacy of perception leads to the recognition that concepts are
“where the action is” in the advance of man’s knowledge. The history of
science is replete with cases of crucial breakthroughs coming from the
integration of observed data into new concepts. Such concepts as velocity
(vs. speed), inertia, gravity, energy, element, valence, supply, demand, evolu-
tion, natural selection, variable, derivative, integral, germ, and synapse, have
opened the door to previously unattainable discoveries. Taking three of these
concept-forming breakthroughs will illustrate the pivotal value of forming
the right concepts.200

200 The Logical Leap [Harriman, 2010] treats many more, and in considerable detail. A brief
summary on p. 179 of that work lists seven of the concepts behind the atomic theory.
 H OW W E K N OW • 1 1 : Overview 370

1. “Inertia”
Prior to Galileo’s work, the dominant conception (or misconception) was that
of “natural motion” toward a “natural place.” Bodies made of “earthy” stuff,
in this view, naturally fall toward the center of the Earth, their natural place.
Falling, it seemed, is just a given, nothing needing an explanation.
Galileo provided a total reconceptualization: place is relative to bodies (an
Aristotelian idea), motion is change in place, speed is the rate of that change,
velocity is speed in a given direction, acceleration is a change in velocity.
This new conceptual framework allowed Galileo to reach the concept of
“inertia,” expressed fully in Newton’s first law of motion: a change in a body’s
velocity requires the exertion of a force upon it. Because bodies have inertia,
their accelerating fall to Earth evidences the operation of a force: gravity. Since
the planets’ speed and direction change continuously, despite inertia, the
question arose as to whether gravity operates on them as well. Thus, forming
the concept of “inertia” opens the door to discovering universal gravitation.

2. “Natural selection”
Darwin’s concept of “natural selection” allowed biologists to achieve an
immensely clarifying new perspective on living organisms, and it led to
a new level of integration across the entire science of biology. As the great
biologist Theodosius Dobzhansky wrote, “. . . in biology nothing makes sense
except in the light of evolution.” [Dobzhansky 1970, 5–6] Natural selec-
tion — the differential survival rates of alternative genotypes — is what
makes sense of evolution and of the adaptedness of organisms’ structures and
actions to their survival.

3. “Germ”
The concept of “germ” as the agent in infectious diseases (later supplemented
by the concepts “virus” and “prion”) revolutionized the science of medi-
cine. No longer was “bad air” taken as a causative factor. The germ theory
of disease was spurred by the observations of John Snow on the cause of
the 1854 cholera outbreak in London. Using statistical analysis, Snow traced
the source of the outbreak to water obtained from the Broad Street pump,
refuting, in the process, the prevalent “miasma” theory.
Louis Pasteur’s experiments proving the role of microorganisms in the
fermentation process added evidence for the germ theory. Finally, Robert
 B ottom - up vs . T op- down T heories 371

Koch proved that a bacterium, Bacillus anthracis, was the cause of anthrax.
These discoveries, to which many of us owe our lives, allowed Joseph Lister to
prove that wound infections were caused not by the “bad air” of a “miasma,”
but by germs. That led to Lister’s introduction of sterile procedures, which
have saved millions of human lives — a dramatic illustration of the power
of forming the right concept.
Proper conceptualization is essential to the advance of knowledge. Thus,
epistemology must focus on concepts — how they are formed, what their
cognitive function is, and what makes a concept valid or invalid.

Bottom-up vs. Top-down Theories

The primacy of perception leads to a wider point: knowledge is essentially
“bottom-up,” not “top-down.” Conceptual knowledge is acquired by building
up from perceptual data.
Man is born tabula rasa: his consciousness is only a potential — until it is
actualized by input to his sensory organs. The fetus in the womb does not have
theories, thoughts, concepts, or even percepts. An infant born without any
senses would never be conscious. A boy may dream that he is driving a car;
a fetus cannot. Even to have a dream about cars, one must have perceived
them (or have been told about them by someone who has perceived them).
Directly or indirectly, perception of cars is necessary to remember cars,
imagine cars, form the concept “car,” make propositions about cars, draw
conclusions about cars, or develop the field of automotive engineering.
Who could deny this? Explicitly, Plato. Implicitly, every Rationalist
philosopher, every theologian, every believer in God, Allah, or Vishnu, every
advocate of some realm that is neither perceivable nor logically derivable
from perception. In other words, the bottom-up nature of knowledge has
been rejected, in theory and in practice, by 99 percent of mankind, including
the majority of the leading figures in the history of philosophy.
In contrast, I have stressed throughout this book that knowledge begins
with perception and builds up hierarchically from perception. The first
“upward” step is concept-formation, a process that starts with the perceived
similarity of two or more perceived objects. The next step upward comes
with the formation of higher-level concepts, in the required hierarchical
order. One cannot have a higher-level concept without the lower-level ones
on which it rests, and the first-level concepts are formed from perception.
 H OW W E K N OW • 1 1 : Overview 372

(The meaningless statements that result from concept-stealing show the con-
sequences of severing the chain back to perception.)
When we subsume something under a concept, that is a “downward” step,
one that cashes in on the power of previous upward steps. To say “Socrates
is a man” is to apply the concept “man” to him — which presupposes one
has formed the concept “man.” The very purpose of ascending the hierarchy
of knowledge is to use the knowledge, to apply it, which means applying it
back to perceptual reality. In cognition, too, what goes up must come down.
The purpose of concepts is conceptual identification. The purpose of theory
is practice. The purpose of consciousness is successful action in the world.
The same progression from the perceptual to the more abstract applies to
the hierarchy of propositions and to the hierarchy of inferences. The progres-
sion is always: from perception to the more abstract (for the sake of sub-
sequent downward application to concretes). The entire hierarchical struc-
ture of knowledge rests on perception as its base. New abstract knowledge
can, in some cases, be reached by deduction from still more abstract ideas,
but those more abstract ideas themselves have to be inductively based.
Induction from perceptual observation is the essential means of gaining
new knowledge. Though deductive steps often appear as intermediaries in
the process, the overall progression is inductive — bottom-up. To repeat the
suspension-bridge analogy to the hierarchy: whenever an item of know-
ledge, like a part of such a bridge, is suspended from one above it, that higher
part must still be supported by the ground; gravity is the ruling principle.
Deduction presupposes induction — the induction that supplied the univer-
sal premise(s) required to make the deduction valid (no valid deduction can
be made without using at least one universal premise).
Even mathematics, the arch-example of supposedly deduction-supplied
knowledge, is inductive — bottom-up — at its root. The basic concepts
of mathematics — “unit,” “number,” “equal,” “more than,” and the like —
are graspable only from perceptual experience. Infants do not have these
concepts; they must be learned by a process of generalizing from p ­ erceptual
observation. (Modern “symbolic logic” and meta-mathematics attempt
to get around this fact by arbitrarily introducing “undefined primitives” in an
“axiomatic system”; but these primitives are just meaningless shapes
or sounds unless based on conceptualizing perceived fact.201)

201 Gödel’s theorem is the reductio ad absurdum of the entire formalistic (Kantian) approach
to logic and mathematics — see Chapter 7.
 B ottom - up vs . T op- down T heories 373

The top-down approach rejects this whole perspective and scorns practical,
“materialistic” concerns. Communing with non-material, non-perceivable
Platonic Forms or “pure” theory is held to be an end in itself. Reason is thus
stripped of its biological function. But life is lived in the concrete, physical,
perceptual world; to reject perceptual reality is to reject life itself.
A revealing example of the top-down, Platonic, approach is contained,
unfortunately, in Euclid’s Elements. Euclid begins that great work by giving
basic definitions, starting with “point” and “line”:

1. A point is that which has no part.

2. A line is breadthless length.

The second definition contains the most apparent problem spot: what does
“breadthless length” mean? There is a Platonic interpretation and an Aristote-
lian one. The Platonic interpretation has dominated historically, whether or
not it is how Euclid himself understood “breadthless length.”
In the Platonic interpretation, a line is an abstract entity, like a Platonic Form,
and that entity has a width of zero. Perceptual reality contains no abstract
entities, only concrete ones — rocks and trees, people and buildings. Concrete
entities have both length and width (and thickness). Even the thinnest line
drawn on paper has some width. So much the worse for perceptual reality,
say the Platonists: “breadthless length” is an “idealization,” they claim, a term
that reflects the Platonic elevation of “ideal” abstractions over perceptual
concretes. Lines are held to inhabit a mathematical world of pure intellection.
We deduce geometrical principles by “pure logic,” then apply them, as best
we can, to the messy, approximate, ignoble perceptual concretes.
Applying the Aristotelian, bottom-up approach enables us to recognize
that the concept “line” is formed by a process of abstraction. We mentally
isolate one characteristic from the others that are co-present, without implying
that the characteristic can exist apart from those others. We can attend to
an entity’s length while ignoring its width; we cannot picture or entertain
any such thing as length-without-width. Lines do not inhabit some “ideal”
mathematical realm; they are aspects of concretes in this world. Specifically,
a line is the shape of an edge of a surface. (In ordinary usage, an ink trail on
paper counts as a “line.” But in geometry “line” does not refer to any such
entity, but to an attribute: edge-shape.)
Thus, it is wrong to say that lines have small, infinitesimal, or zero width.
The geometer’s line has neither breadth nor breadthlessness: width is not
a property that a geometrical line can have or be bereft of. (See Chapter 6,
 H OW W E K N OW • 1 1 : Overview 374

under “Negative Propositions.”) It is as wrong to say that lines lack width as

it is to say that they lack homes. Lines are neither breadthless nor homeless.
Consciousness is epistemologically active. A concept is the product of an
integrative process; it is not an internal reproduction of an external “uni-
versal.” Having a valid concept of X does not imply that X exists as such in
external reality. Just as no intrinsic “universal” is implied in conceptualizing
entities, so no separated characteristic, such as length, is implied in concep-
tualizing characteristics. In forming concepts of characteristics, we simply
abstract from the co-present characteristics, such as the width of a surface.
(This abstraction amounts to the omission of whole categories of measure-
ment.) Neither “line” nor any other concept asserts the absence of the char-
acteristics that are not included in that concept. Concepts do not pretend
that what is isolated by a process of abstraction exists separately in the world.
A long tradition of Platonism in mathematics treats not only lines but
also shapes, numbers, and mathematical functions as referring to “idealized”
mathematical objects — which severs mathematics from the real world. Thus,
Kline reports: “To thoughtful scientists it has been a constant source of wonder
that nature shows such a large measure of correlation with their mathematical
formulas.” He quotes Einstein asking: “How is it possible that mathe­matics,
a product of human thought that is independent of experience[!], fits so
ex­cel­lently the objects of physical reality?” [Kline 1985, 227 & 216]
Robert Knapp identifies the root error that generates this bewilderment:

The common fallacy that mathematics pertains specifically

to a mathematical universe; that mathematics applies to the
world, but is not about the world begins with the first page of
Euclid’s monumental work. . . . [In fact] mathematics applies to
the world because it is about the world. [Knapp, 1]

The “top-down” view also creates the problem Plato raises in the Meno:

. . . how will you look for something when you don’t in the least
know what it is? How on earth are you going to set up something
you don’t know as the object of your search? . . . even if you come
right up against it, how will you know that what you have found
is the thing you didn’t know? [Meno 80d]

But one doesn’t start a quest for “something I don’t know.” One looks for
more knowledge about something one does know. One may know the general
 B ottom - up vs . T op- down T heories 375

nature of something and perform further observation to get the informa-

tion required to identify it more specifically — e.g., “I see some a­ nimal
moving through the brush in the distance; I’ll get closer to see what kind
of animal it is.” Or, one may already have the prerequisites of reaching
a more abstract identification, and perform the required integrative process
— e.g., “These animals look similar when contrasted to those; I will try to
isolate the characteristics that make them look similar, in order to use them
as a basis of conceptualization.”
As discussed in Chapter 7, “the problem of the Meno” is often used by
skeptics to attack the rules of definition. The Rule of Scope says that a defi-
nition must have the same scope as the concept that it defines; it cannot
be either too broad or too narrow. But to test a proposed definition against
that rule, we must already know what things are units of the concept and
what things are not. Yet, the skeptics say, the definition is what tells us which
things are to count as units of the concept. For instance, a definition of “man”
as “a two-legged animal” is too broad, because that definition would include
birds. But how do we know, prior to defining “man,” that birds aren’t men?
The answer is that cognition is a bottom-up process. We do not begin
with a definition. We do not begin with a concept. We begin with percep-
tion, including perceived similarities and differences (for first-level concepts).
In the case of forming the concept “man,” a child begins by perceiving the
differences of these things (men) from other things with which they share
a Conceptual Common Denominator (a characteristic possessed by other
animals, including birds). The child attends to the characteristics (such as
shape) in which the similarity and difference appear. Next, he isolates the
human shape-range from that of other animals. Then he omits the measure-
ments within the human shape-range, integrates the men by using a word,
open-ends the “file” thus formed, and finally uses all this perception-based
knowledge to formulate a definition. The explicit formulation of a defini-
tion comes only after — years after, in this case — all the hierarchically prior
steps of the process have been taken. One already knows what things are and
aren’t men when one comes to the task of formulating a definition of “man.”
“The problem of the Meno” serves only as a refutation of the top-down,
Platonic-Rationalistic approach to knowledge.202

202 The Humean-empiricist approach could be styled “bottom-bottom.” In denying the

validity of abstraction and concepts, it implies that human cognition should not attempt
to advance beyond the animal level.
 H OW W E K N OW • 1 1 : Overview 376

The Prose Principle

To fully defend the bottom-up nature of knowledge, we need to recog-
nize what I call “the prose principle.” I take the name from a statement by
a character in a Molière play. Upon being told the difference between poetry
and prose, the character exclaimed: “By Jupiter! I’ve been speaking prose for
forty years without even knowing it.”203
The prose principle, then, is: one can apprehend and use information prior
to conceptualizing it.204
Knowledge is often implicit before it is made explicit. Molière’s character
knew how to speak in prose before he knew that what he was doing could
be described in that way.
The prose principle is inherent in the bottom-up approach to knowledge.
The fact that perception precedes conception means that pre-conceptual
awareness of information precedes the conceptualization of it. The implicit
precedes the explicit. To make something “explicit” means: to grasp it
conceptually, in words; prior to that act, one is aware of the thing, but
pre-conceptually. This process of explication is essential to the growth
of knowledge.
For example, long before one is able to form concepts of attributes, one is
aware of attributes perceptually and uses that awareness to form concepts of
entities. The child sees the shapes of tables, chairs, and beds, and uses that
awareness to form the concept “table,” but he will not for a long time be able
to form the concept “shape.”
Consider some cases in which it is quite difficult to name explicitly
the unconceptualized information that one plainly is aware of and uses.
For instance, it is fairly easy to recognize a Botticelli painting by its style.
But how many people have been able to identify what aspects of the paint-
ings they are using to do that — i.e., what Botticelli’s style consists of?
A similar difficulty is apparent in trying to identify what makes jazz recogniz-
able as jazz, or a person’s face recognizable as that person’s face. It is difficult,
but possible, to make explicit the means by which we recognize such things.

203 The character is M. Jourdain in The Bourgeois Gentleman.

204 This is my formulation, as a principle, of an observation that Ayn Rand makes in the
Workshops on Objectivist Epistemology; she agrees with the suggestion that the first time
a child abstracts, he does it without knowing that that is what he is doing: “That’s right.
He was talking prose and he didn’t know it. That joke [from Molière] really is very
important. In a certain sense, it names a great many psychological processes.” [ITOE, 151]
 T he P rose P rinciple 377

The prose principle also applies to knowledge of method: we can appre-

hend and use the right method of cognition before having conceptualized
and understood what we are doing. An infant is able to form concepts,
even though it has no theory of how concept-formation works. Men were
able to reason syllogistically long before Aristotle identified the syllogism
and its rules. Early scientists were able to induce long before any principles
of scientific induction had been conceptualized.
What then does the explication add? Conscious control. That control has
several payoffs: the ability to avoid sources of error, the ability to devise
new methods by extrapolating from old ones, and objectivity: the ability
to disting­uish between that which is valid, established, proven — vs. that
which merely “feels right.”
Principles drawn from attending to the clear, simple cases serve to guide
us in dealing with cases that are obscure and complex. As Rand wrote in
another connection, “That which is merely implicit is not in men’s conscious
control; they can lose it by means of other implications, without knowing
what it is that they are losing or when or why.” [FNI, 53]
The prose principle sweeps aside many paradoxes and claims to innate
ideas. Take Noam Chomsky’s arguments for regarding the principles of
grammar as being innate. Children learning language, at age two or three,
appear to be using rules that are much too complex for them to be able to
understand or apply. This is shown, Chomsky claims, by the mistakes they
make — e.g., using “mouses” as the plural of “mouse,” and “gooder” instead of
“better.” Aren’t they following such rules as: pluralize a noun by appending an
“s,” and: intensify an adjective by appending an “er”? The children are not old
enough to understand even the words used in these rules. Thus, Chomsky
claims, knowledge of the rules of grammar must be innate.
But by applying the prose principle we can unmask an equivocation
here — an equivocation on the phrase “following rules.” The children
certainly have implicit rules, but not explicit ones. From their experience
with spoken English, they have formed associations and expectations, with-
out having any explicit identification of the rules of grammar. They do learn
grammar, but by a wordless, implicit, unconceptualized process.
A child learning to ride a bicycle is, in one sense, “applying the laws of
physics.” But he has no explicit knowledge of those laws, so his “applica-
tion of the laws of physics” means only that he is aware perceptually of the
concrete pushes and pulls of the forces acting, and uses this perceptual
experience in learning to ride. Since a circus bear can be trained to ride
 H OW W E K N OW • 1 1 : Overview 378

a bike, it is abundantly clear that no conceptual knowledge is required.

What is required is perceptual awareness and perceptual-level association.
The prose principle saves us from being driven into the primitive swamp
of “innate ideas” — a supposed knowledge of reality prior to cognitive
contact with reality.
According to the bottom-up understanding of knowledge, the essential
step in the acquisition of new knowledge is inductive, not deductive. By mov-
ing up from perception, one, in effect, reaches a height from which one looks
down on the same perceptual reality, but with a wider field of view.
The top-down view of the Rationalists holds that deduction is the basic
source of knowledge. But, overall, knowledge cannot be derived top down, by
deduction from something still higher — with everything hanging, in effect,
from a skyhook.
In regard to generalizations formed by induction, Rationalists hold that
only the subsumption of a generalization under a still wider one provides
proof and certainty. For them, the starting point of the induction itself is
merely a guess or “a free invention of the intellect” (as in the notoriously
unsuccessful “model-building” approach of many economists).
This much is true: a generalization, even when proved with certainty by
a reduction to perception, does gain something from being subsumed under
a wider generalization. The gain, however, is not in certainty (unless the ear-
lier evidence was inconclusive), but in understanding.
For instance, it is a cognitive advance to subsume “Shoplifting is wrong”
under “Violating others’ rights is wrong,” or to subsume Kepler’s laws under
Newton’s. Such subsumptions add to our knowledge, but they are not the
primary means of acquiring and validating knowledge.
The cognitive value of subsuming a generalization under a wider one
does not represent shoring up an induction by a deduction. The subsump-
tion’s cognitive value comes from the fact that it integrates an induction with
other inductions. Such integration strengthens and clarifies a group of inde-
pendent inductions, but only because the entire structure rests on perception.
By analogy, in building a cabin, the walls support the roof, rather than
vice-versa; but the roof adds stability and strength to the whole structure,
including each wall — by more fully connecting the walls to each other.
Induction is the primary means of acquiring new conceptual knowledge.
To doubt or question the validity of induction itself is to doubt or question
conceptual knowledge itself — which is self-refuting (see Chapter 7). Thus,
it is a mistake to attempt to “rescue” induction by attempting to reduce it to
a disguised deduction or to shore it up by deduction.
 T he P rose P rinciple 379

Peikoff ’s theory of induction shows how induction works; it is not the mis-
begotten attempt to “prove induction is valid.” And the prose principle makes
a valuable additional point: we do not need a theory of how induction works
in order to induce. Just as, in the physical realm, man did not have to under-
stand the process of combustion in order to tame fire, or utilize a theory of
concept-formation in order to form concepts, so one does not need a theory
of induction in order to induce. In each case, discovering how the process
works is of great value; it brings the process under better control and extends
its possibilities, but one can use a physical or cognitive method before one
knows how it works, or even has conceptualized the process involved.
Induction is a means of learning from experience by forming the relevant
concepts, so that we can then apply them to observed facts; induction is not
a disguised deduction.
Induction proceeds by identifying causal connections, and the causal
connections grounding first-level generalizations are available perceptually.
When a toddler pushes a spoon, a pencil, or a ball, he sees and feels the caus­
ation. He feels his arm and hand’s effort, he sees his hand moving forward
and moving the object forward in front of it. He feels the object’s resistance
to his push and his overcoming of that resistance. He feels the greater effort
required to move larger objects, and feels the wall resist and thwart his push.
Such perceptual data are what the child uses for his first generalizations.
Just as a child learning to ride a bicycle does not make any explicit “appeal
to the laws of physics,” so an infant learning that pushing objects makes them
move does not make any explicit “appeal to the law of causality.” He is aware
of causal connections long before he can identify them as “causal connections.”
He is speaking prose, but he doesn’t know it.
Objections to induction stem from the hidden assumption that all learn-
ing is deductive, as if an infant needed syllogisms in order to learn or “prove”
that pushing things makes them move, that ice is cold, that cookies crumble.
That assumption reveals how the top-down, Rationalist approach completely
inverts the hierarchy of knowledge.
Beyond the perceptual level, all cognitive processes are hierarchical.
By the nature of a hierarchy, its base is a special case, something quite dif­
ferent from the elements that depend on that base. Everyone in the military
has to take orders from those of higher rank, but the Commander-in-Chief
does not; every organism is composed of cells, but one-celled organisms
(from which the others evolved) are not; every concept rests on preceding
concepts — except the base: first-level concepts, formed from perception.
Treating a hierarchy’s base as if it were equivalent to derivative elements
 H OW W E K N OW • 1 1 : Overview 380

is a prevalent form of hierarchy-inversion.205 In the attacks on induction,

that fallacy is committed by treating first-level generalizations as subject
to the same requirements as higher-level generalizations — specifically,
in requiring justification, proof, or validation in the case of first-level gen­er­
ali­za­tions. One must recognize the context: the facts giving rise to the need
for justification. Justification is needed for inductions that are fallible, but
first-level gener­aliza­tions are no more subject to error than are a child’s first
conceptualizations. Just as there is no such thing as a toddler getting the
concept “cookie” wrong, so there is no such thing as him getting “Cookies
crumble” wrong. “Wrong” has no meaning in either case. (Nor does “right.”)
In forming his first concepts, the child sees the similarities he sees, and in
forming first-level generalization, the child sees the causality he sees. (More-
over, there is no meaning to “right vs. wrong” or “valid vs. invalid” before
concepts of consciousness have been formed.)
For these basic conceptual acts, there is no such alternative as “justified
vs. unjustified,” “proven or unproven,” “hypothetical or certain.” Those issues
involve reducing the higher levels back to their base, but here it is the base
that we are considering.
By analogy, a lawyer might get into the habit of asking about every law,
“Is it constitutional?” But if he holds the context and remembers that the base
of a hierarchy is not just another element in it, he would not dream of asking:
“Is the Constitution constitutional?” But it is no better to ask: “Is the child
justified in concluding that pushing a ball moves it or that cookies
crumble?”206
Validation is reduction to perception. The first inductions need no reduc-
tion because they don’t climb a ladder of abstractions away from perception.
The prose principle’s negative corollary is: one does not have to know that
a process is logical in order for it to be logical. Understanding what justifies
a cognitive procedure requires the development of epistemology, and that
comes only long after people have been unselfconsciously using that procedure
to acquire knowledge.

205 Two other forms are: question-begging (petitio principii), which attempts to use a derivative
to establish its own base, and what could be termed “base-negation,” the attempt to use
a derivative to negate its own base. Hierarchy-inversion itself is a species within the wider
cate­gory of hierarchy-violation: any attempt to use an item apart from the prior items
that give it meaning and/or validity — e.g., floating abstractions and arbitrary assertions.
206 The wider point is: reduction to a base is a procedure that cannot be demanded of the
base itself. (In mathematical recursion, the process of deriving all subsequent elements
from the base cannot be required of that base. E.g., a “number” is 1 or any successor of
a number — which rules out asking: “Of what number is 1 the successor?”)
 S ome H istory 381

(Even after one has learned the need to check one’s conclusions to
determine whether they represent knowledge or mistaken belief, one does
not have to know how to justify an idea before being justified in using it.
For instance, centuries before Pythagoras came up with his proof of the theo-
rem that bears his name, Egyptian builders knew and applied the geometrical
facts that he later proved.)
One does not have to know that one knows before one can know. One can
apprehend and use information prior to the conceptual identification of that
information. That’s the prose principle.

Some History
In the broad history of thought, the major proponents of the bottom-up
approach are Aristotle and those who accept this aspect of his episte­mology
(notably, Aquinas, Bacon, and Locke). The major proponents of the top-
down approach are Plato and the Rationalists (notably, Descartes and Kant).
The clash between Aristotle and Plato reverberates throughout the realm
of ideas, both in individuals and in cultures. [See OPAR, 451–460.]
The battle between the two theories takes place on a highly abstract level,
but it is impossible to overstate the practical, real-world importance
of their impact upon a culture. Science and the Industrial Revolution arose
from and depended upon the Aristotelian recognition of the primacy of
existence and the primacy of perception. And it was the Platonic primacy
of consciousness and primacy of concepts philosophy that led to contempt
for this perceptually given world and the abandonment of the achievements
of Greco-Roman civilization, plunging mankind into the long, stagnant night
of the Dark Ages. In denigrating reality and perception, the Platonists reject
life on this earth.
Do I exaggerate? Plato calls the body “a prison” and “the tomb of the soul,”
[Cratylus, 400c]. He says that “true philosophers make dying their profes-
sion,” [Phaedo, 67e] and proclaims that a man should have “trained himself
throughout his life to live in a state as close as possible to death.” [Phaedo, 67e]
In holding that conceptual knowledge is gained by turning inward to align
the soul with a “higher” realm, Plato treats perception as illusion, and holds
that sensory observation is a distraction from “true knowledge,” which is of
a non-perceivable “higher dimension”:
 H OW W E K N OW • 1 1 : Overview 382

. . . if we are ever to have pure knowledge of anything, we must get

rid of the body and contemplate things by themselves with the
soul by itself. . . . No pure knowledge is possible in the company of
the body. . . . So long as we are alive, we shall continue closest to
knowledge if we avoid as much as we can all contact and associa-
tion with the body . . . and instead of allowing ourselves to become
infected with its nature, purify ourselves from it until God himself
gives us deliverance [i.e., death]. [Phaedo, 66e-67a]

For Plato, learning and discovery never actually occur: all knowledge
is innately present in the soul at birth, having been acquired before birth
when the soul dwelt in the world of Forms; what we take to be the discovery
of new knowledge, he says, is really the recollection of this innate knowledge
— a doctrine that is utterly incompatible with science, and with the techno-
logical-industrial progress science makes possible.
Descartes’ form of the primacy of concepts, while discarding Plato’s open
mysticism, nonetheless severs knowledge from sensory perception:

. . . no science is acquired except by mental intuition or deduc-

tion. . . . the first principles themselves are given by intuition alone. . . .
By intuition I understand, not the fluctuating testimony of the
senses . . . but the conception which an unclouded and attentive
mind gives us . . . [Descartes, Rules III & IV]

Descartes’ “intuition” is, of course, hopelessly subjective. To establish an

idea’s truth, we are to turn inward to inspect the qualities of the ideas as
such (their clarity and distinctness), not their relation to perceptual reality.
Cartesian “intuition” is the kind of thing that has given self-evidency a bad
name. But, in fact, the perceptually self-evident has nothing in common with
the anti-perceptual “intuition” Descartes is positing.
Kant synthesizes Plato with Descartes, adding a skeptical twist from Hume.
Following Plato, Kant holds that the world we perceive through our senses
is shadowy, not fully real; the “higher” realm is now called the “noumenal”
world, the innate ideas are now called “categories.” Following Descartes,
Kant’s philosophy starts from within consciousness and regards awareness
of the external world as problematic; Descartes’ “intuition” now becomes the
voice of “noumenal” reality heard through (of all things) our alleged sense
of “moral duty.” Following Hume, Kant regards the external world as utterly
 S ome H istory 383

unknowable (except for this alleged sense of “duty”); Hume’s “habits of asso-
ciation” become Kant’s innate “synthesizing categories.”
Beneath Kant’s notoriously convoluted terminology is an all-out separa-
tion of the mind from reality. Both perception and conception are cut off
from “things as they are in themselves,” which remain forever unknowable.
Science deals only with “appearances” (“phenomena”), and is, in effect, held
to be a shared delusion.
The contemporary attack on science by Popper, Kuhn, and Feyerabend
is grounded in Kantianism. In their post-modern version of Kant, science
deals not in truth but in public relations, and is the embodiment not of reason
but of faith.207 Thomas Kuhn writes that the scientist must “have faith that
the new paradigm will succeed with the many large problems that confront
it, knowing only that the older paradigm has failed with a few. A decision of
that kind can only be made on faith.” [Kuhn 1970, 158]
Among more recent philosophers, there are many who regard percep-
tion as uncertain, impoverished, concept-laden, non-cognitive. In one form
or another, their position is that perception cannot serve as the base of the
conceptual level. If so, there are only two ways one can defend concepts:
by claiming that concepts are based on and justified by something non-per-
ceptual (e.g., innate ideas or mystical revelations) — or by claiming that con-
cepts do not need any base — the position known as “anti-foundationalism.”
The founder of anti-foundationalism was Richard Rorty. In 1986, Rorty
wrote: “Nothing grounds our [cognitive] practices, nothing legitimizes
them, nothing shows them to be in touch with the way things really are.”
[Rorty, 753]
This is utter bankruptcy. The very existence of epistemology depends upon
the recognition that beliefs need justification. Such justification must consist
in identifying the foundation — the earlier knowledge — on which a con-
clusion rests. The alternative to reducing an idea to its base in perception
is Platonism unhinged, Platonism gone native, Platonism without a World
of Forms, without any metaphysical base, championing instead “networks”
of interrelated beliefs supported by nothing. Is it supported by “coherence”
(consistency)? No, because coherence itself has no support, according to anti-
foundationalism. And, remember, even a paranoid schizophrenic’s beliefs
have that same apparent consistency.

207 See Kuhn, 158 and Lakatos & Musgrave, 228–229. On the Kantian roots of this phenom-
enon, see Lennox, 1981.
 H OW W E K N OW • 1 1 : Overview 384

The actual “legitimizer” of beliefs, according to anti-foundationalism, is

the beliefs of other people. In a confession of profound psychological depen-
dence, of the inability to conceive of objective fact or independent judgment,
the anti-foundationalists blithely assert that conformity to others’ opinions
is the only conceivable justification for an idea. Undeterred by the 20th
century’s spectacle of millions surrendering their minds and their lives to
vicious notions promulgated by Stalin, Hitler, and fatwa-issuing ayatollahs,
these pseudo-philosophers nonetheless expect to find safety in justification
by “consensus” or “social practices” or . . . whatever their latest jargon is for:
what other people say.
Consider Rorty’s response to John McDowell, when McDowell raised
objections to Rorty’s social view of justification. McDowell used the example
of the bogus claims that two chemists made to having achieved “cold fusion.”
McDowell points out that

. . . whether or not cold fusion has occurred is not the same as

whether or not saying it has occurred will pass muster in the
current practice [i.e., is in line with the beliefs of one’s peers]. . . .
Without this difference, there would be no ground for conceiv-
ing one’s activity as making claims about, say, whether or not
cold fusion has occurred, as opposed to achieving unison with
one’s fellows in some perhaps purely decorative activity on a level
with a kind of dancing. [Brandom, 125]

Here is Rorty’s response:

What, I still want to ask, is so “mere” about getting together with

your fellow inquirers and agreeing on what to say and believe?
. . . How do I tell a world constituted by linguistic practices from
a world constituted by facts . . . ? I have no idea.208 [Brandom, 125–
126, emphasis added]

208 Rorty “has no idea” what a fact is because, following Wilfrid Sellars, he rejects perception
as the base of knowledge. The anti-foundationalist motto is: “Only a belief can justify
another belief.” The source of this perverse position is the failure to understand how
concepts can be objectively formed from perceptual data. The solution is the Objectivist
understanding of concept-formation (plus the recognition that perception is a form
of awareness). [See Bayer, 2011]
 S ome H istory 385

In all this, Rorty admits that he is taking Kant straight, without a “sugar
coating” and as “replacing objectivity with solidarity.” 209 [ibid.] In that regard,
Rorty is correct: the destruction of objectivity is the ultimate result and goal
of Kant’s entire system. “I have therefore found it necessary,” Kant wrote,
“to deny knowledge in order to make room for faith.” [Kant 1958, B xxxi]
To understand contemporary philosophy, one must understand what Kant
did to “deny knowledge.”

The Kantian Reversal

Kant reversed a crucial distinction, the distinction between the what and the
how — between what one knows and how one knows it. Kant turns the means
of awareness into the only object of awareness. We cannot know extra-mental
reality, Kant says, because our means of awareness stand in the way; we are
incapable of knowing “things as they are in themselves”; we can know only
the appearances of things:

. . . we can therefore have no knowledge of any object as a thing in

itself, but only in so far as it is . . . an appearance. [Kant 1958, B xvi]

But, contrary to Kant, an “appearance” is how we perceive and know the

object. Objects are grasped in a certain form, and what is thereby known is
the object, the thing in reality.
By illicitly making the appearance into the object of awareness, Kant is
able to claim that what we are aware of is the wrong thing — not tables and
rivers but only the appearances of tables and rivers. Because we see a table by
a certain means and in a certain form, we cannot see the table; we see
instead our sight of the table.210 Likewise, because we conceive tables by a
certain means and in a certain form, what we are conceiving is our concept,
not tables. The cognitive how turns into a what that blocks our way to know-
ing reality, locking us up inside our own minds. The existence of a means
of cognition is held to make cognition impossible. Thus, Rand’s immortal
demolition of Kant:

209 “Solidarity” is a euphemism for cognitive surrender to the smiles and frowns of others.
210 As H. J. Paton, a premier Kant scholar, puts it: “. . . the world we know is a world of
appearance.” [Paton, 62] But “the world we know” means the object of awareness.
E.g., when we look at or touch an apple, what we see and feel, Kant claims, is not the apple,
but appearances internal to us. Here, “appearances” and “internal” are stolen concepts.
 H OW W E K N OW • 1 1 : Overview 386

His argument, in essence, ran as follows: man is limited to a con-

sciousness of a specific nature, which perceives by specific means
and no others, therefore, his consciousness is not valid; man is
blind, because he has eyes — deaf, because he has ears — deluded,
because he has a mind — and the things he perceives do not exist,
because he perceives them. [FNI, 32]

Kant does not shy away from but embraces the consequence: a thorough-
going, blatant primacy of consciousness:

Hitherto it has been assumed that all our knowledge must con-
form to objects. . . . [We may] have more success in the tasks of
metaphysics, if we suppose that objects must conform to our
knowledge. [Kant 1958, B xvi]

When the how is turned into the what, process becomes everything, and
substantive content becomes irrelevant or impossible. This formalism, which
infects every aspect of modern culture, is another consequence of Kant’s
destructive legacy.
Consider how widely formalism reigns. It rules foreign policy (which
consists in diplomacy and “talks” while ignoring the nature of the regimes
involved), in painting (where the object has disappeared, leaving only formal
relationships among color patches), in ethics (where “sincere intentions” and
“compassionate” attitudes count, but results do not), in economics (where the
leading journals appear as though they were journals of mathematics, filled
with equations relating to the economists’ own constructs, not to actual
production and trade), in education (where “progressive” educators hold
that “students need to learn not a body of knowledge but ‘how-to’ skills” 211),
and in constitutional law (where “due process” has become “substantive,”
while what is actually substantive — individual rights — is ignored).
Kant’s conversion of the means of cognition into a barrier to cognition
has corrupted even the field of logic, turning it into something purely for-
mal, into the avowedly arbitrary manipulation of avowedly meaningless

211 E. D. Hirsch, Jr., in an article deploring the formalism: “Students would be better off
gaining knowledge by studying real subject matters in a sensible, cumulative sequence.
Instead, elementary schools are dominated by content-indifferent exercises . . . on the
erroneous assumption that reading comprehension is a formal skill akin to typing.”
[Hirsch, 2012]
 S ome H istory 387

“inscriptions and utterances,” as it is put, rather than being the method

of proceeding from truth to truth. Kant was well aware of this implication:
“[L]ogic . . . is justified in abstracting . . . from all objects of knowledge and
their differences, leaving the understanding nothing to deal with save itself
and its form.” [Kant, 1958, B ix]
Or, as Wittgenstein more succinctly put the same doctrine, “all proposi-
tions of logic say the same thing, to wit nothing.” [Wittgenstein, 2011, 53]
In fact, logic is the means of staying in contact with perceptual reality
and of using the unit-economy provided by concepts to extend our grasp of
the world ever further. Logic is not severed from content; it is the means of
expanding that content, in accordance with the metaphysically given identity
of man’s consciousness.
Rand identified the root of all Kantian systems: their “attempts to regard
identity as the disqualifying element of consciousness.” [ITOE, 80] Objectivism
proceeds from exactly the opposite premise: the identity of consciousness is
its means of operation, which enables it to grasp reality.
Of those means, I would select three facts as the essentials:

1. Perception provides data that are metaphysically given and inerrant.

2. A concept integrates that data by measurement-omission.
3. Conceptual integration provides unit-economy (“the crow epistemology”).

Thus, Rand was able to show how (to modify her wording) identity is the
qualifying element of consciousness — i.e., its enabler.
In addition, Rand identified the fact that man’s volition, his free will, is not
a blind, disruptive factor independent of reason but is precisely one’s power
to use reason — i.e., to initiate a process of focused thought, oversee it, subject
it to the canons of logic, and thus attain objectivity.
This radically new understanding of volition allowed her to integrate voli-
tion with the identity of both the knower and the known: consciousness has
a determinate means of knowing a determinate world; man’s choice is to use
those means or not.

. . . man exists and his mind exists. Both are part of nature, both
possess a specific identity. The attribute of volition does not con-
tradict the fact of identity, just as the existence of living organ-
isms does not contradict the existence of inanimate matter. . . .
[M]an is able to initiate and direct his mental action only in
accordance with the nature (the identity) of his consciousness.
 H OW W E K N OW • 1 1 : Overview 388

His volition is limited to his cognitive processes; he has the power

to identify (and to conceive of rearranging) the elements of
reality, but not the power to alter them. He has the power to use
his cognitive faculty as its nature requires, but not the power to
alter it nor to escape the consequences of its misuse. [PWNI, 26]

Both the world and consciousness have identity; the facts of reality are
firm and absolute; the methods of cognition a given man adopts are a matter
of his choice, but there is a right choice: the one that accords with the require-
ments of cognition that are set by the identity of his cognitive equipment.
Conceptual knowledge is not automatic. Nor is it infallible. Even when
a man chooses the right method and takes into account all the facts he knows,
he can yet be mistaken. Logic is a necessary condition of knowing, but not
a sufficient one. It is not illogical to conclude that the stick half-submerged
in water is bent, but the fact remains that the stick is not bent. The senses
didn’t err, but the intellect did. The fact that such occasions are rare is not the
point. The point is whether or not one can accept the fact that even a fully
logical, rational process can reach a mistaken conclusion — a fact that does
not negate or call into question the knowledge that one does possess.
The majority of philosophers have not been willing or able to accept the
fact that there is no immunity from error. They have assumed that only
the infallible can be certain, believing that without a metaphysical guarantee
against error, there can be no certainty.
Some men, upon discovering that they can be wrong, pretend that they
can gain infallibility by relying on supposed revelations or following rote
rules (e.g., Rationalistic deduction). This is intrinsicism. Others respond by
retreating into a world of “appearance” or “seeming” or “true for you, not
for me.” This is subjectivism, which is best captured by an epistemological
paraphrase of Dostoyevsky: Since God is dead, everything is uncertain.
Both sides assume that only an infallible, divinely guaranteed certainty
would be valid. They differ only as to whether or not that infallibility can
be obtained. Both sides long for the epistemic safety they felt as children,
the sense that parents and teachers were infallible authorities. Though h ­ aving
matured enough to realize the contrary, they have not matured enough
to take a first-hand look at the epistemic situation, define rational standards
for themselves, and take responsibility for their own thought and judgment.
Instead, they continue to long for an automatic guarantee — a guarantee from
another consciousness. They do not welcome the responsibility of choice,
but rebel against it, feeling “a stranger and afraid in a world I never made,”
 S ome H istory 389

as Housman’s poem puts it.212 In this regard, Rand’s description of the logical
positivist applies to all those who long for automatic certainty:

Like a spoiled, disillusioned child, who had expected predigested

capsules of automatic knowledge, a logical positivist stamps his
foot at reality and cries that context, integration, mental effort and
first-hand inquiry are too much to expect of him, that he rejects
so demanding a method of cognition, and that he will manufac-
ture his own “constructs” from now on. (This amounts, in effect,
to the declaration: “Since the intrinsic has failed us, the subjective
is our only alternative.”) [ITOE, 67–68]

Choice and the responsibility it brings are not to be denied. Nor do they
threaten rationality. In fact, our fundamental and all-pervading choice is
precisely whether to be rational or irrational. Choice is our power to free
ourselves from merely reactive response. Our volition gives us the ability to
make an unbiased assessment of the facts, to be self-critical and objective.
Man’s ability to focus his mind and set it to the task of knowing and living in
reality is what makes conceptual knowledge and rational certainty possible.
Rational certainty, as opposed to the mystics’ desire for automatic immu-
nity to error, is not only possible, in regard to fundamentals it is normal — at
least for a rational man. But the standard of certainty is not intrinsic. Nor is
it subjective. Epistemic standards cannot demand omniscience, but they do
demand doing what is possible: non-contradictory integration into the sum
of one’s knowledge. As I noted in discussing the need for logic:

The issue is not: “What would one conclude if one were omni-
scient?” but: “what is the proper conclusion to draw given all
the facts available now?” Epistemic standards are prospective,
not retrospective. [Ch 6, 204. ]

The alternative is not commandments vs. whims, dogmatism vs. skepti-

cism, or standards without choice (intrinsicism) vs. choice without standards
(subjectivism). The proper alternative to all such dichotomies is: the choice
to apply rational standards.

212 Surprisingly, Housman’s poem opens with a vibrant assertion of proper independence.
[A. E. Housman, “The Laws of God”]
 H OW W E K N OW • 1 1 : Overview 390

Summary
Knowledge, Rand writes, is “a mental grasp of a fact(s) of reality, reached
either by perceptual observation or by a process of reason based on percep-
tual observation.” [ITOE, 35] The method of reason is logic. Thus, conceptual
knowledge is acquired by applying logic to perception, deriving ideas from
perception and reducing them back to perception, integrating them into
the full context.
I close by revisiting the inspiring passage in which Rand brings to life
the meaning of a commitment to reason.

It consists of a method of using one’s consciousness, best des­ig­­-

na­ted by the term “conceptualizing.” It is not a passive state of
registering random impressions. It is an actively sustained
proc­ess of identifying one’s impressions in conceptual terms,
of inte­grating every event and every observation into a concep-
tual context, of grasping relationships, differences, similarities
in one’s perceptual material and of abstracting them into new
concepts, of drawing inferences, of making deductions,
of reaching conclusions, of asking new questions and discovering
new answers and expanding one’s knowledge into an ever-growing
sum. The faculty that directs this process, the faculty that works
by means of concepts, is: reason. [VOS, 21–22]

And that’s . . . how we know.

 BIBLIOGRAPHY

Aphrodisas, alexander of, Quaestiones (1994), R. W. Sharples

(trans.), Cornell University Press.
Aristotle, The Complete Works of Aristotle (1984), J. Barnes (ed.),
Bollingen Library Series, LXXI 2.
Aristotle, The Basic Works of Aristotle (1941), R. McKeon (ed.),
Random House.
Ayer, A. J. (1954), Philosophical Essays, Macmillan.
Bach-y-Rita, P., et al. (1969), Vision Substitution by Tactile Image
Projection, Nature, 221.
Bayer, Benjamin (2011), A Role for Abstractionism in a Direct Realist
Foundationalism, Synthese, vol. 180, no. 3, 357–389.
Bergson, Henri (1946), The Creative Mind, The Philosophical Library.
Binswanger, Harry (1981), The Possible Dream, The Objectivist Forum,
vol. 2, nos. 1 & 2.
Binswanger, Harry (1986), The Goal-Directedness of Living Action,
The Objectivist Forum, vol. 7, no. 4.
Binswanger, Harry (1989), Consciousness as Identification ( audio),
https://estore.aynrand.org
Binswanger, Harry (1990), The Biological Basis of Teleological Concepts,
The Ayn Rand Institute Press.
 H OW W E K N OW 392

Binswanger, Harry (1991), Volition as Cognitive Self-Regulation,

Organizational Behavior and Human Decision Processing,
vol. 50, no. 2, 154–178.
Binswanger, Harry (1999), Free Will (audio),
https://estore.aynrand.org
Blanshard, Brand (1939), The Nature of Thought,
George Allen & Unwin Ltd.
Boyer, Carl (1968), A History of Mathematics, John Wiley & Sons.
Brandom, Robert (2000), Response, in Rorty and His Critics, Blackwell.
Chalmers, David (1997), The Conscious Mind, Oxford University Press.
De Wulf, Maurice (1959), The System of Thomas Aquinas, Dover.
Descartes, René, Rules for the Direction of the Mind, in Haldane & Ross
(eds.) Philosophic Works of Descartes (1955), Dover.
Descartes, René, Meditations on First Philosophy, in Essential Works of
Descartes (1961), L. Blair (trans.), Bantam Books.
Dewey, John (2008), The Logic of Inquiry, Searching Press.
Dobzhansky, Theodosius (1967), The Road Traversed and the Road
Ahead, in I. W. Knobloch (ed.), Readings in Biological Science,
pp. 441–466, Meredith.
Dobzhansky, Theodosius (1970), Genetics of the Evolutionary Process,
Columbia University Press.
Efron, Robert (1966), What is Perception? Boston Studies in the
Philosophy of Science, vol. 4, 1969, 137–173, Springer.
Freud, Sigmund (1960), The Ego and the Id, W. W. Norton.
Gibson, J. J. (1950), The Perception of the Visual World, Houghton Mifflin.
Gibson, J. J. (1986), The Ecological Approach to Visual Perception,
Lawrence Erlbaum Associates.
Gödel, Kurt (1992), On Formally Undecidable Propositions of Principia
Mathematica and Related Systems, Dover.
Goodman, Nelson (1955), Fact, Fiction, and Forecast, Bobbs-Merrill.
Harriman, David (2010), The Logical Leap, New American Library.
Harris, Sam (2012), Free Will, The National Post (Toronto),
July 31, 2012, A15.
Hawkins, J. & Blakeslee, S. (2005), On Intelligence, St. Martin’s Griffin.
Heidegger, Martin (1929), What Is Metaphysics? [Inaugural lecture
to the combined faculties], University of Freiburg.
 B ibliograph y 393

Held, R. & Hein, A. (1963), Movement-Produced Stimulation in the

Development of Visually Guided Behavior, Journal of Comparative
and Physiological Psychology, vol. 56, no. 5, 872.
Hirsch, E. D. Jr. (2012), Vocabulary Declines with Unspeakable Results,
The Wall Street Journal, Dec. 12, 2012, A 15.
Hitler, Adolf (1928),Volkische Beobachter, in C. Cohen (ed.),
Communism, Fascism, and Democracy (1962),
University of Michigan Press.
Hume, David (1961), A Treatise of Human Nature, Dolphin Books.
Innocent III (1966), On the Misery of Man, in B. Murchland,
Two Views of Man, Frederick Ungar.
James, William (1890), Principles of Psychology, Henry Holt & Co.
Jonas, Hans (1966), The Phenomenon of Life, Harper & Row.
Kalish, Donald (1966), quoted in: What (If Anything) to Expect
from Today’s Philosophers, Time Magazine, Jan. 7, 1966.
Kant, Immanuel (1958), Critique of Pure Reason,The Modern Library.
Kant, Immanuel (1949), On a Supposed Right to Lie from Altruistic
Motives, L. W. Beck (transl.) in Immanuel Kant: Critique of Practical
Reason and Other Writings in Moral Philosophy, University
of Chicago Press.
Kline, Morris (1980), Mathematics: The Loss of Certainty,
Oxford University Press.
Knapp, Robert (forthcoming), Mathematics is About the World.
Kuhn, Thomas (1970), The Structure of Scientific Revolutions,
2nd ed., University of Chicago Press.
Kunze, A. & Moroney, J. (eds.) (1999), Glossary of Objectivist
Definitions, Second Renaissance Books.
Lakatos, I. & Musgrave, A. (eds.) (1965), Criticism and the Growth
of Knowledge, Cambridge University Press.
Lennox, James G. (ed.), (2000), Aristotle’s Philosophy of Biology,
Cambridge University Press.
Lennox, James G. (1981), The Anti-Philosophy of Science,
The Objectivist Forum, vol. 2, no. 3.
Leroy, Charles G. (1870), The Intelligence and Perfectibility of Animals
from a Philosophic Point of View, Chapman and Hall.
Locke, John (1959), An Essay Concerning Human Understanding, Dover.
McDermott, Terry (2011), 101 Theory Drive, Pantheon Books.
 H OW W E K N OW 394

Mill, J. S. (1950), A System of Logic, Hafner.

Mill, J. S. (1957), Utilitarianism, Bobbs-Merrill.
Miller, George A. (1956), The Magical Number of Seven,
Plus or Minus Two: Some Limits on Our Capacity for Processing
Information, The Psychological Review, vol. 63, 81–97.
Montessori, Maria (1912), The Montessori Method, Transaction.
Nagel, Ernest (1944), Logic without Ontology, in Y. H. Krikorian (ed.),
Naturalism and the Human Spirit, Columbia University Press.
Noë, Alva (2004), Action in Perception, MIT Press.
Parker, F. H. (1953), Realistic Epistemology, in John Wild (ed.),
The Return to Reason, pp. 153–159, Henry Regnery Co.
Paton, H. J. (1970), Kant’s Metaphysics of Experience, vol. I,
George Allen and Unwin.
Peikoff, Leonard (1976), The Philosophy of Objectivism, Lecture series,
https://estore.aynrand.org
Peikoff, Leonard (1988), Fact and Value. The Intellectual Activist,
www.aynrand.org
Peikoff, Leonard (1991), Objectivism: The Philosophy of Ayn Rand,
Meridian.
Peikoff, Leonard (2013), Objective Communication, NAL Trade.
Pierson, L. & Trout, M. (2005), What Is Consciousness For?
Cogprints (preprint), www.cogprints.org/4482.
Plato, Meno (Guthrie, transl.), in Hamilton & Cairns (eds.),
The Collected Dialogues of Plato (1961), Bollingen Series LXXI,
Pantheon Books.
Plato, Phaedo (Tredennick, transl.), in Hamilton & Cairns (eds.),
The Collected Dialogues of Plato (1961), Bollingen Series LXXI,
Pantheon Books.
Plato, Republic, F. M. Cornford (transl.), The Republic of Plato (1945),
Oxford University Press.
Quine, W. V. O. (1953a), On What There Is, From a Logical Point of View,
Harper Torchbooks.
Quine, W. V. O. (1953b), Two Dogmas of Empiricism,
From a Logical Point of View, Harper Torchbooks.
Rand, Ayn (1957), Atlas Shrugged, New American Library.
Rand, Ayn (1961), For the New Intellectual, Signet.
Rand, Ayn (1964), The Virtue of Selfishness, Signet.
 B ibliograph y 395

Rand, Ayn (1966), Our Cultural Value-Deprivation,

in The Voice of Reason (1990) L. Peikoff (ed.), Meridian.
Rand, Ayn (1967), Capitalism: The Unknown Ideal, Signet.
Rand, Ayn (1971), The Romantic Manifesto, Signet.
Rand, Ayn (1982), Philosophy: Who Needs It, Signet.
Rand, Ayn (1990), Introduction to Objectivist Epistemology,
expanded 2nd edition, H. Binswanger (ed.), Meridian.
Rand, Ayn (1990), The Voice of Reason, L. Peikoff (ed.), Meridian.
Rand, Ayn (1997), Journals of Ayn Rand, D. Harriman (ed.), Dutton.
Rand, Ayn (1999), The Comprachicos, in Peter Schwartz (ed.),
Return of the Primitive, Meridian.
Randall, John H., Jr. (1940), The Making of the Modern Mind,
Riverside Press.
Riggs, L. A. & Ratliff, F., The effects of counteracting the normal
movements of the eye, Journal of the Optical Society of America,
vol. 42, 872–873.
Rorty, Richard (1986), From Logic to Language to Play, Proceedings
and Addresses of the American Philosophical Association, vol. 59, no. 5.
Ruby, Lionel (2000), Logic: An Introduction, Paper Tiger.
Salmieri, Gregory (2006), Objectivist Epistemology in Outline,
Ayn Rand Institute Summer Conference (audio)
https://estore.aynrand.org
Salmieri, Gregory (2013), Conceptualization and Justification, in
A. Gotthelf & J. G. Lennox (eds.), Concepts and Their Role in Knowledge:
Reflections on Objectivist Epistemology, Ayn Rand Society Philosophic
Studies Series, University of Pittsburgh Press.
Sarnat, H. B. (1985), The brain of the planarian as the ancestor
of the human brain, The Canadian Journal of Neurological Sciences,
vol. 12, no. 4.
Sarnat, H. B. (2002), When does a ganglion become a brain? Seminars in
Pediatric Neurology, vol. 9, no. 4.
Schwartz, J. & Begley, S. (2003), The Mind and the Brain:
Neuroplasticity and the Power of Mental Force, Harper Perennial.
Simpson, George G. (1964), This View of Life, Harcourt, Brace & World.
Spinoza, Baruch (1949), Ethics, Hafner.
Stace, W. T. (1955), The Philosophy of Hegel, Dover.
 H OW W E K N OW 396

Stalin, Josef (1924), The Foundations of Leninism, in C. Cohen (1962),

Communism, Fascism, and Democracy (1962), University of Michigan.
Tallis, Raymond (2011), Aping Mankind, Acumen.
Watson, John B. (1970), Behaviorism, W. W. Norton.
Windelband, Wilhelm (1958), A History of Western Philosophy,
Harper Torchbooks.
Wittgenstein, Ludwig (1958), The Blue and Brown Books,
Harper Torchbooks.
Wittgenstein, Ludwig (2011), Tractatus Logico-Philosophicus,
Empire Books.
 INDEX

Prepared by Allison T. Kunze

Abstraction from abstractions 137–70 Ad Ignorantiam, fallacy of 285, 288

axiomatic concepts 167–70 Adjectives 155–57, 160–62, 177, 178,
concepts of characteristics 152–62 238–39
concepts of consciousness 162–67 Affirming the consequent,
cross-classifications 142–43, fallacy of 273
149, 150 Agency
subdivisions (narrowings) and free will 344–45
142–43, 149 “problem” of 352–54 index
wider integrations 139–41 Agnosticism 289
see also Hierarchy (of concepts) Alexander of Aphrodisias 322 n173
Abstraction (process of) 115, 157 Analogies 309 n163
Abstractions see Concepts Analytic-synthetic dichotomy
Abstractness 149 133 n60, 181–83
Accident, fallacy of 201–02 Anti-foundationalism 383–85
Action Appearance vs. reality 93–95
concepts of 154, 162 Aquinas, Thomas 58, 79 n33, 381
as irreducible primary 348 Arbitrary, the 278–92
types of 331 dismissal of 284–85
and volition 331–38, 352–54 in history of philosophy 291–92
 H OW W E K N OW 398

Arbitrary, the (continued) Axiomatic concepts

and imagination 280–82 as base of knowledge 28–30
and Law of Rationality 278–85 examples of 167
and logic 279–81 formation of 167–70
as neither true nor false 283 as foundation of objectivity 170,
and objectivity 298 191–92
vs. proof or validation 261–92 irreducibility of 42
see also Burden of proof principle as lacking CCDs 169–70
Ad Ignorantiam, fallacy of Axioms
Argument from Delusion 85 as base of knowledge 22, 26–27
Argument from Illusion 73–76, 85 characteristics of 24–25, 26–28
Argument from the Relativity of consciousness 26–28, 32
of Perception 79–80, 85 of existence 22–26, 32
Aristotle 15, 23, 37, 54, 58, 87, 175, and free will 355–59
187, 193 n96, 196 n100, 201–02, of identity 30 n6, 193–94
225, 227 n118, 256, 257 n133, implicit vs. explicit 167–68
341, 341 n186, 347, 364, 377 and proof 22–23
logic, contribution to 192–95, as propositions or concepts 28–30
215–16, 255, 293–94, as self-evidencies 22–24
and Moderate Realism 102, Ayer, A. J. 352, 354
106–07, 297 Bach-y-Rita, P. 83
vs. Plato re concepts 102, 106–07, “Back to the Future” 341 n187
366–69, 373–75 Bacon, Francis 58, 381
vs. Plato re knowledge 381–82 Bayer, Benjamin 384 n208
on proof 23 Begging the Question, fallacy of
Artificial intelligence see Computers 207, 255, 380 n205
Atheism 289 Begley, S. 52
Attributes Behaviorism 27
see Characteristics, concepts of Bergson, Henri 244–45
Automatization 223–24, 303–04, Berliner, Judith 311 n167
338–39, 367–68 Binswanger, Harry 37, 59 n20,
Awareness 116 n47, 204, 267, 328 n176,
commitment to 345 349 n192, 351
and focus 324–26 Biological perspective
levels of 327–28 on concepts 128–31
states of 26–28, 166 on consciousness 37–41, 49–55,
see also Consciousness 57–58, 86, 332
on knowledge 40, 308–09
 inde x 399

on principles 306–07, 320 Chalmers, David 45

on reason 97–98 Chamberlain, Neville 202–03
Blakeslee, S. 70 n Character see under Free will/volition
Blanshard, Brand 24 n3 Characteristics
Borderline cases (concept-formation) see Characteristics, concepts of;
116, 222, 238–39 Conceptual Common
Bottom-up view of knowledge Denominator; Distinguish-
371–78, 381 ing characteristic
Bourgeois Gentleman, The 376 n203 Characteristics, concepts of 152–62
Boyer, Carl 251 and causal knowledge 152–53
Brain see Mind-brain relation concepts of action 154, 162
Brandom, Robert 384 concepts of relationships 154–55
Burden of proof principle 286–90 formation of 153–55
see also Arbitrary, the; Proof as isolating from entities 156–62
Causality and nominalized adjectives
and concepts of characteristics 155–56
152–53 scope of term as used 152–53
and free will 347–51 Choice
and fundamentality 300–03 in action and decision-making
Greek vs. modern view of 347 331–38, 353–54
and induction 256–58, 379 as axiomatic concept 358
Law of 152 n73, 348–49 and character 338–44
perceptually given cases of vs. determinism 323–24, 345–46
257–58, 379 353–54
and principles 305, 307, 312 as epistemological 321-24, 326-28
reason as a causal primary 351 focus as fundamental choice
CCD see Conceptual Common 324–26
De­nom­inator and man’s consciousness 349–51,
Certainty 271–78 387–89
attacks on 276, 290 see also Causality (and free will);
as contextual 272–73 Focus; Free will/volition
and error 272, 290, 388–89 Chomsky, Noam 377–78
and evidence 271–73, 275-78, 288 Circularity (in definitions) 220
and evidentiary continuum 271 Color blindness 78
and knowledge 271–72 Color concepts 110, 111–14, 140,
standard of 272, 334, 389 153–54, 158–60
see also Arbitrary, the; Burden of Color perception 66
proof principle; Evidence Communism, concept of 236
 H OW W E K N OW 400

Computers 40, 45–47, 129 n56 Realist, Nominalist, and

Concepts Objectivist theories of 101–36
of action 154, 162 of relationships 154–55
biological perspective on 97–98, and similarity 110–13
128–31 unit-economy and 129–36,
cognitive role of 129–36 365–66
and context 198–205 and “universals” 101–07, 127–28,
and “the crow epistemology” 296–97
130–32, 135–36, 148–49, valid and invalid 227–32
230–31, 234 and volition 191–92, 321–24,
definition of 127 326–28
and definitions 134, 214–27 and words 124–27
and differentiation 110–15, see also Abstraction from ab­­strac-­
158–62, 215–17 tions; Axiomatic concepts;
false division 235–36 Characteristics, concepts of;
false integration 236–38, 242–43 Consciousness, concepts of;
file-folder analogy to 124–25, Hierarchy; Unit-economy
134–35, 140–41, 183–84 Concepts of characteristics
first-level 137–38, 144, 153 n74 see Characteristics, concepts of
foils in concept-formation Concepts of consciousness
111–17 see Consciousness, concepts of
formation of 110–28 Conceptual Common Denominator
and integration 123–28 112–17, 216–17, 235–37
and intrinsic-subjective-objective Conceptual relativism see Nominalism
trichotomy 296–98 Conceptualizing, policy of 360, 390
and logic 214 Consciousness
meaning of 134–35 as active process 35–36
objectivity of 296–98 in animals 37–40, 59–63, 364
as open-ended 132–33, 185 as axiomatic 26–28
and optical illusions 73–78 biological role of 37–41 49–50,
order of generality in 210 57, 86, 332
vs. perception 97–98, 365–66 casual efficacy of 49–54
primacy of as false 366–69, as difference-detector 36, 68–69
381–89 and form-object distinction
and propositions 171–73, 177, 78–84, 94–95
177 n90, 183–90 as identification 86–93
purpose of 234 identity of 26, 30–41, 54–55, 196,
“Rand’s Razor” 232–36 319, 363
 inde x 401

as irreducible 28, 42–45 Context-dropping, fallacy of 202–03,

vs. materialism 42–54, 60, 350 310, 311–12, 337
and mind-brain relation 43–45, Contradictions
51–52, 349 n193 attempt to maintain one 314–15
mystical view of 48, 92 Law of Non-Contradiction
Naïve Realism on 86–88, 91–93 192–95, 310–11
object and subject of 30–32 vs. logic 194–95, 259–60
primacy of, as error 32–35, Cross-classifications
107, 249, 328, 385–86 see under Abstraction from
and the primacy of existence abstractions
32–35, 50–51, 366–69 “Crow epistemology”
and the primacy of perception and cognitive role of concepts
196–97, 366–71 130–32, 135–36, 319
reification of (by Descartes) 53 and fundamentals 303
Representationalist view of 88-93 and hierarchy 148–49, 196-97, 206
and values 329–30 and higher-level concepts 144
see also Axiomatic concepts; and hypotheses 278
Axioms; Consciousness, and the identity of consciousness
concepts of; Introspection 196–97, 319
Consciousness, concepts of and logic 196–97, 206, 278
formation of 162–66 and means of cognition 387
intensity, role of 164–66 and “Rand’s Razor” 235
and teleological measurement and sources of error 268 n138
166–67 and unit-economy 130–32, 135,
Context 198–205 230–31, 365–66
and absolutes 315–17 see also Unit-economy
and certainty 272–73 Culture, modern 386
context-dropping 202–03, 337 De Wulf, Maurice 103
and definitions 225–27 Decision-making/deliberation
and generalizations 258 an extended example 334–38
implications for epistemology and volition 321, 332–34 Deduc-
204 tion 194–95, 254–56, 262–63,
and integration 198, 200–01 378
and knowledge 266–67 see also Induction; Inference
and logic 199–200 Definition, rules of 215–25
and principles 311–12 fundamentality 222–24
and propositions 252–53 genus and differentia 216–19
relation to hierarchy 206–07 reference 219–20
 H OW W E K N OW 402

Definition, rules of (continued) Dewey, John 225, 318

scope 220–22 Differentia
unit-economy 224–25 see under Definition, rules of
Definitions, theory of 134, 214–27 Differentiation
and analytic-synthetic dichotomy in concept-formation 110–15
181–83 consciousness as difference-
as contextual 225–27 de­tector 110–11
and defective propositions 242–43 in defining concepts 215–17
defined 214 and integration 139–43
function of 214–15, 225 in isolating characteristics
label on file folder, analogy to 215 158–62
by non-essentials, fallacy of second-order 69, 159–60, 162
236–38 Direct realism (re perception) 68–71
revisions of 226–27 Distinguishing characteristic
see also Definition, rules of 115–17, 121–23, 127–28,
Deliberation 139–40, 142-43, 216-17, 222-24
see Decision-making/ Division of labor 301
deliberation Dobzhansky, Theodosius 98, 138, 370
Delusions 84–85, 207–08 Dostoyevsky, Fyodor 237, 306 n161,
Demand, Law of (concept of) 264 388
Derivation 262–64 Doubt 290–91
Descartes, René Dreams 57 n19, 85, 167, 282–83
Cartesian dualism 53 Drift (mental) 324–25, 327–28,
and evil demon 357–58 332–33, 338–41
and primacy of consciousness Dualism, Cartesian 53
33–34 Effort
and Rationalism 58 vs. the automatic 327–28, 345
and reification 53 and free will 322, 324–25, 342,
and Representationalism 90–91 344–45, 355
top-down approach of 382 Efron, Robert 67 n24
Determinism Ego 331, 360–61
and causality 347–51 Einstein, Albert 266, 374
as denial of free will 323, 345–46 Emotions
as self-refuting 355–59 as automatic response 345
social and genetic theories of and focus 329–30, 342–44
359–60 harmony with reason 337–38
see also Agency (“problem” of); Empiricism 14–16, 369, 375 n202
Indeterminism Entities 60–63, 167, 178, 183, 348
 inde x 403

Epiphenomenalism 49–50 begging the question

Equivocation 207, 255, 380 n205
examples of 54, 79, 155–56, 195, context-dropping 202–03, 310,
243–44, 290, 377 311, 337
fallacy of 195, 255 definition by non-essentials 236
Error 267–71, 290 equivocation 195, 255
see also under Certainty; false division 235–36
Perception false integration 236–38, 242–43
Essentials 203, 222–23 hierarchy-inversion 379–80
Ethics 311, 338–44 package-dealing 236–38
Euclid 373–74 pure self-reference 248–51
Evasion 325, 332–33, 339–41, reification 66
345 n189 retroactive self-evidency 146, 157
Evidence re-writing reality 80
vs. Ad Ignorantiam 285–86 self-exclusion 253, 356
and certainty 271, 272–73, 334 stolen concept 34–35, 46–47, 48,
and hypotheses 275–78 54, 94, 148, 208, 214, 247–48
see also Arbitrary, the; Burden Fallibility, cognitive 192, 197, 290–91,
of proof principle; Certainty; 388–89
Rationality, Law of False division 235–36
Evolution and evolutionary biology False integration 236–38, 242–43
49–50, 59, 98 Falsehood
see also Natural selection and the arbitrary 283–85
Excluded Middle, Law of 193–94, of premises 268–69
207–08, 259–60 as not implying truth 268 n139
Existence of propositions 239–41
as axiomatic 22, 24–25, 29–30 see also Propositions; Truth
as irreducible primary 24–26 Fascism, concept of 236
primacy of 32–33, 50–51, Feelings see Emotions
366–69, 381 Feyerabend, Paul 383
as self-evident 22–24 File-folder analogy (to concepts)
“Extremism,” as invalid concept 124–25, 133, 134–35, 140–41,
237–38, 240 n120, 242 183–84, 187–89, 195, 215, 341
Extrospection 33, 249, 367, 368–69 First-level concepts 137–38, 139–41,
Fallacies 255 144, 153 n74, 167–68, 173
accident 201–02 First-level generalizations 257–59,
ad ignorantiam 285, 288 378–80
affirming the consequent 273 Floating abstractions 151, 214, 380 n205
 H OW W E K N OW 404

Focus (mental) 324–38 338–44

and action 331–38 and the choice to live 361
and automatization 339 vs. determinism 323–24, 345–60
and character-development epistemological nature of 322-24,
338–44 326–28
and choice to live 361 vs. indeterminism 354–54
as commitment to full awareness locus of 323–24
324 ff., 361 as presupposition of objectivity
defined 324 292, 294–95
different aspects of 328–29 and problem of agency 352–54
vs. drift and evasion 325–26, and psycho-epistemology 341
332–33 and social environment 359–60
fundamentality of 329 n177 validation of 344–46
introspective awareness of 344 see also Focus (mental)
as managing one’s mind 325, Freedom, concept of 236–37, 242
334–38 Freud, Sigmund 229, 355
monitoring 325–29 Fundamentals/fundamentality
and moral values 338 defined 300, 301
and motives 329–30 as domain-relative 301–02
as precondition of thinking 324 examples of 300–02
and epistemology 341 and philosophic principles
as reality-orientation 329–30 317–18
see also Causality (and free will); of principles 305–07
Free will/volition Rule of 222–24
Foil (in concept-formation) 111–17, as source of cognitive power
118, 153 303–04
Ford, Henry 264 and unit-economy 303–04
Form-object distinction 78–84, 94–95 Galileo 58, 148, 276, 347
Formalism 386–87 Ganzfeld experiments 68
Forms (Platonic) 98–99, 101-02, 104, Generalizations
229, 311–12, 363–64; first-level 257–59, 378–80
see also Plato/Platonism fundamental 305
Free will/volition 321–62 high-level (principles) 307–08
and action/decision-making see also Induction
331–38 Genus 112–13, 216–19
as axiomatic 355–59 Gibson, J. J. 62, 63, 64 n22, 68–71,
and causality 347–51 76 n32, 89–90, 95
and character-development God 25–26, 180, 207, 228, 243–44, 289
 inde x 405

Gödel’s theorem 251, 372 n201 in induction 256–58

Gold, concept of 226–27 inversion of 247–48, 379–80
Goodman, Nelson 207 n112 of knowledge 205–06, 371 ff.
Gotthelf, Allan 351 n197 and logic 205, 207, 214
Grammar/grammatical concepts metaphysical basis of 205–06
175, 377–78 opposition to, in contemporary
adjectival nouns 161 philosophy 207–08
adjectives 155–57, 160–62, 177, and propositions 246–48
178, 238–39 and reduction 262
and defective propositions see also Funda­mentals/
243–46 fundamentality; Reduction;
gerunds 155 Spiral process of knowledge;
nouns 155–56, 161–62, 178 Stolen concept, fallacy of
philosophical grammar 244–46 Hierarchy-inversion, fallacy of
prepositions 155 379–80
proper names 101, 215, 242 n122 Higher-level concepts
role in propositions 178–79 see Abstraction from abstractions
verbs 154, 178 Hirsch, E. D., Jr. 386 n211
Gravity 148, 266–67, 276, 309, 369–70 Hitler, Adolf 202–03, 238, 246 n124,
Guilt 330, 332–33, 339 384
Hallucinations 84–85, 207–08 Honesty, principle of 305-06, 308, 310
Harriman, David 256 n131, 259, Housman, A. E. 389
276 n145, 369 n200 Hume, David 14–15, 31, 198 n101,
Harris, Sam 352 332, 365, 369, 375 n202, 382
Hawkins, J. 70 n30 Hypotheses 272–78, 284–86
Hearing, sense of 61, 67, 73, 75–76, Hypothetico-deductive method
84, 129 274–75
Hegel, G. W. F. 119, 141, 229 Identification, conceptual 171–73,
Heidegger, Martin 245 174, 177
Hein, Alan 63 Identity 169, 215, 302, 307
Held, Richard 63 and causality 152 n73, 347–48
Heraclitus 262 and consciousness 57, 86, 92,
Hierarchy 205–08 129, 131, 319
of composition 300 and existence 30 n6, 72, 105–06,
of concepts 143–49, 149–51 129, 134
and context 206–07 Law of 30 n6, 72, 193–94, 259,
and derivation 262–64 292, 298, 319
of generality 210 Illusions 73–78, 85
 H OW W E K N OW 406

Imagination 31, 57, 228, 280–84, 330 Intrinsic-subjective-objective

Implicit concepts 29–30 trichot­omy 296–98
Implicit knowledge 376–78 Intrinsic theory of concepts 296–97
Impossibility 277–78 Intrinsicism 296–97, 388–89
Indeterminism 354–54 Introspection 31–32, 51
Individual rights 315–17 and concepts of consciousness
Induction 255–59 162–66
and concepts 256–59 and free will 344–46
and generalizations 256–59 validity of 54
and hierarchy 256–58 Invalid concepts
and logic 259 and defective propositions
and open-endedness of concepts 242–51
258–59 in history of philosophy 229
Peikoff ’s theory of 256–59 intellectual package-deals 236-38
and principles 307–08 in the sciences 229
Inertia 369-70 “stripes” and “solids” example 230
Inference 22–24, 194–95, 254–60 types of 227–38
see also Deduction; Induction see also False division; False
Inflation, concept of 218–19 integration; “Rand’s Razor”
Innate ideas 378, 382 James, William 36, 318, 322 n173
Innocent III 317 n172 Jonas, Hans 39
Integration Joyce, James 13–14
in concept-formation 123–28 Judgment
and context 198–200 of mental processes 191-92, 333-34
in definitions 215–17 vs. perception 73, 75–78, 93–94
and differentiation 215–17 propositional 239–41
and knowledge 201–05 see also Justice, concept of
and objectivity 298 Justice, concept of 231–32
and principles 307 Kalish, Donald 15 n2
and proof 264–65 Kant, Immanuel 58, 229, 310, 385 ff.
and understanding 209 formalism 386
valid vs. invalid 236–39, 242–43 influence 14–15, 381, 382–87
wider 139–41 Rand vs. 291–92, 385–86
see also Concepts view of consciousness 90, 382,
Intellectual package-deal 385–87
see Package-dealing, fallacy of re the what and the how 385
Intensity (of a psychological process) Kitten experiment (perception) 63
164–66 Kline, Morris 248–49, 374
 inde x 407

Knapp, Robert 374 Kuhn, Thomas 383

Knowledge Kunze, Allison T. 215 n114
Aristotle vs. Plato on 381–82 Lakatos, I. 383 n207
and action 331–38 Language 126–27, 223–24
and automatization 303–04 see also Word
axioms as base of 22, 26–27, 28–30 Law of causality see under Causality
biological perspective on 13, 40, Law of identity see under Identity
97–98, 308–09, 332 Law of rationality
bottom-up vs. top-down see under Rationality
theories of 371–75, 376–81 Leibniz, Gottfried 229
and burden of proof principle Lennox, James G. 102, 383 n207
286–90 Likelihood 271–72, 275, 288
causal 286–87 Lister, Joseph 371
as contextual 198–205, 266–67 Locke, John 15, 58, 102–07, 118, 293,
defined 21 297, 315, 381
derivation of 262–64 see also Moderate Realism
and evidentiary continuum 271 Logic
expansion of vs. error 269–71 vs. the arbitrary 278–92
as hierarchical 143–46, 205–08 basic laws of 192–94
implicit and explicit 376–77 and concepts 214
importance of perception to and context/integration 198–205
57–59, 371 ff. deduction 194–95, 254–56
induction’s role in 255–59 and defective propositions 239–51
integration and 201–05 defined 194
major theories of 296–98 and definitions 214–27
as mental product 21, 195–96 and error 268–71
new as not contradicting old and free will 323–24
265–67 and hierarchy 205–08, 214
objective 294–95, 298 and the identity of consciousness
Rand on 21, passim 195–97
requirements of 196–97 as impossible under determinism
spiral process of 208–11 356–57
storage of 199 induction 255–59
undercutting of 314–15 and inference 254–59
validation of 261–65 and Kantian formalism 386–87
see also Axioms; Certainty; nature and purpose of 192,
Con­cepts; Perception 259–60
Koch, Robert 370–71 and objectivity 292–95
 H OW W E K N OW 408

Logic (continued) standard of 119, 122

as perceptually based 192 teleological 166–67
and propositions 239–51 Meno, problem of 221, 374–75
Rand’s contribution to 195–97 Mental focus see Focus (mental)
reduction 262–64 Metaphor 160–61, 220
and rules of 295 Metaphysical categories 166, 244
scope of 196–97 Metaphysically given vs. man-made
and truth 251–54 72–73
see also Fallacies; Mill, John Stuart 158 n76, 245–46
Invalid concepts; see also Mill’s Methods
Proof; “Rand’s Razor” Miller, George A. 131
Logical Leap, The 256 n131, Mill’s Methods
257 n132, 276 n145, 369 n200 Agreement 159–62
Logical positivism 314 n169, 389 Difference 158
Love, concept of 243 Mind-body dichotomy 98, 319–20
Man, nature of 13, 99–100 Mind-brain relation
Marx/Marxism 229, 253 materialism on 43–48, 49–54, 60
Materialism 42–54, 60, 350 mental action 50–52, 349 n193
Mathematics 34–35, 120, 135–36, see also Consciousness; Dualism,
186–87, 203–04, 248–49, 372 Cartesian; Neural processes
see also Euclid; Gödel’s theorem Moderate Realism (re concepts)
Matter 44–45, 64–65, 349–50 n6 102–04, 106–07, 297
see also Materialism see also Realism
McDermott, Terry 174 n86 Molière 376
McDowell, John 384 Monitoring (introspective) 325–29
Measurement 115–23 Montessori, Maria 360
measurement-exclusion 179 Moral values 338
measurement-inclusion 174, Moroney, Jean 200 n103, 215 n114
177, 179 Motives 329–32, 361
measurement-integra- Müller-Lyer illusion 77
tion 116–17 Munch, Edvard 13
measurement-omission 115–17, Musgrave, A. 383 n207
118–19, 121–23, 174 Mysticism 15–16, 58, 92
measurement-proximity 112 see also God; Invalid concepts;
and propositions 173–78, 179 Plato/Platonism
and similarity 112, 113–17 Nagel, Ernest 314 n169
“some but any” principle Naïve Realism 86–88, 91–93
118, 121–22 Narrowings
 inde x 409

see under Abstraction from 292–95

abstractions and the object of cognition 293
Natural selection 39, 40, 98, vs. Platonism and Pragmatism
146 n69, 301–02, 370 319–20
Negative propositions/negation Omniscience 99–100, 204–05, 209,
179–81, 193, 287 n152 269–71, 389
Neural processes 28, 43–44, 70, 83, Open-ended nature of concepts
86, 90, 129 n56, 133 n59, 162 n78, 132–33, 182, 185, 234, 240, 258
199, 325 n174, 359 Open-ended vs. closed-ended prop-
Newton, Isaac 284 ositions 185–86
gravity, concept of 369–70 Optical illusions see under Perception
law of gravitational attraction Optional concepts 238–39
148, 266 “Package-dealing,” fallacy of 236–38,
“Newton’s Razor” 284 276–77
Nixon, Richard 299 Paradoxes 248–51
Noë, Alva 44 n16 Parker, Francis H. 87 n38
Nominalism 107–09, 124, 207, 239 Parts of speech
97, 365 see Grammar/grammatical
Non-Contradiction 192–95, 310–11 concepts
Non-existence 24–25, 119, 168, Pasteur, Louis 370
179–81, 245, 288 Paton, H. J. 385 n210
Nordby, Klaus 4, 68 n27 Peikoff, Leonard 17, 25, 33, 36 n14,
Nouns 155–56, 161–62, 178 44, 92 n41, 93 n42, 149,181, 182,
Objective status of concepts 296–98 197, 198, 202-03, 206-07, 228,
Objectivism passim, 86, 133–34, 238 n119, 256-59, 262, 263, 264,
294, 296, 387 266–67, 270, 281 n148, 283,
Objectivism: The Philosophy of Ayn 294–95, 296, 306–07, 324, 328,
Rand (OPAR) 17, 296 329-30, 338, 379
see also Peikoff, Leonard Perception 57–95
Objectivist theory of concepts as axiomatic 58–59
110–36, 163–67, 239 as base of all knowledge 58–59,
Objectivity 371 ff.
of concepts 296–98 and concepts 365–66
defined 295 as contextual 198, 200 n102
vs. intrinsicism and subjectivism as continuous process 63
296–98 defined 63, 85–86
of knowledge 298 direct realism 68–71
of mental process or product form-object distinction 78–84
 H OW W E K N OW 410

Perception (continued) see also Realism; Forms

vs. hallucination 84–85 Popper, Karl 383
as inerrant 72–78 Possibility
J. J. Gibson on 68–71 and the arbitrary 279, 282, 288–89
and logic 192, 196–97 re atheism vs. agnosticism 289
vs. materialism 60 “can” vs. “might be” 276-78, 290-92
optical illusions 73–78 vs.“likely” and “certain” 271–72
primacy of 366–71 metaphysical vs. epistemic 276–78,
and propositions 172–73, 190 290–92
and reduction 262 “possible for” vs. “possible that”
relativity of 79–84 276–78, 290–92
and self-produced movement 63 as requiring evidence 275–76,
vs. sensation 59–63 288–89
sensationalism 64–68, 76 Pragmatism
and spatial array of entities 60–62 vs. Platonism/dogmatism 318–20
see also Color perception vs. principles 41, 299, 318–20
Philosophical grammar 244–46 Primacy of existence vs. primacy of
Philosophy consciousness 366–69, 381
Aristotelian vs. Platonic 366–69, Primacy of perception vs. primacy
381–82 of concepts 366–71, 381
defined 317 Principles 299–320
as framework for living 317–18 as absolute 312–14, 315–17
Physics 229, 256 n131, 307, 317–18 biological perspective on 306–07,
Pierson, Lee 64 n22, 192 n95, 351 n197 320
Plato/Platonism 15, 35, 221, 309–10, as cognitive bridges 307–10
319–20, 331, 363 cognitive function of 306–09,
as anti-body 37, 98, 381–82 319–20
as anti-learning 381–82 as contextual 311–12, 315–17
vs. Aristotle 37, 101–02, 296–98, defined 305
366–69, 373–74, 381–83 as fundamental generalizations
on concepts 101–04, 106–07, 305–07
365–67, 369 and long-range action 306–07
and determinism 347 moral 306, 311
and geometry 373–74 philosophic 317–19
problem of the Meno 374–75 vs. Pragmatism 41, 299, 318–20
on reason 98, 296–98 rights as 315–17
top-down approach 371–75 as simplifiers 306, 310–11
on “universals” 127–28 Probability, mathematical 285–86
 inde x 411

Progressive education 318, 386 vs. sentences 171–72

Proof 261–71 as trans-temporal 240–41
vs. the arbitrary 278–85, 288–89 and truth 251–54
Aristotle on 23 see also Pseudo-propositions
burden of proof principle 286–90 Prose principle 376–81
and certainty 271–72 Pseudo-propositions 241–42,
vs. derivation 262–64 246, 284
disproof 289 see also Pure self-reference,
and error 267–71 fallacy of
and inference 22–24 Psycho-epistemology 135, 341,
of a negative 288 325 n174
as reduction 262–64 Pure self-reference, fallacy of 248–51
and the self-evident 22–24 Pythagoras 167, 203,315, 381
and validation 22-23, 261, 264-65 Quine, Willard van Orman 207–08
Proper names 101, 215, 242 n122 Rand, Ayn (where quoted)
Propositions abstraction from abstractions
and analytic-synthetic dichotomy 138–42, 145, 147
181–83 abstraction, process of 115
classificatory vs. descriptive 173-78 analytic-synthetic dichotomy
closed- vs. open-ended 185–86 133 n60, 181–82
cognitive function of 171–72, 176, automatization 233–34, 304,
183, 186–90 367–68
and context 252–53 axiomatic concepts 28-30, 169-70
defective 241–51 axioms 22, 24
defined 179 causality and consciousness 350
and definitions 219 character 339, 343 n188
and evidence 242 characteristics 156–57
formation of 172, 173–74, 175-78 concepts/concept-formation
and hierarchy 246–48 122–23, 124, 127–28, 132,
and inference 254–59 134–36, 183
and logic 239–51 concepts of consciousness
meaning/content/implications of 162, 164, 166–67
183–90 Conceptual Common Denom­-
negative 179–81 inator (CCD) 112
as organizations of concepts conceptual package-deals 236–38
171–73, 183–86, 186–90 conceptualization 365
relation to perception 172–73 consciousness 31, 32, 35, 36,
role of grammar 178–79, 243–46 86, 92
 H OW W E K N OW 412

Rand, Ayn (where quoted, continued) principles 320

context-dropping 202 propositions 171 n83, 186
“crow epistemology” 130–31 prose principle, the 376 n204
definitions 215–17, 223, 227 psycho-epistemology 135
disparate vs. different 118 “Rand’s Razor” 232–33
distinguishing characteristic re-writing reality, fallacy of 80
114–15 vs. Realist theory of concepts 107
facts and values/is-ought 332 reason 16
floating abstractions 151, 214 rights 315 n170
free will/volition 322–25, 328–29, self-exclusion, fallacy of 253
352, 355, 360–61 sensations 64 n22
grammar 179 similarity 110
hypothetico-deductive method stolen concept, fallacy of 34
274–75 teleological measurement 166–67
implicit knowledge 377 truth 252
induction 259 unit 117
integration 201, 209, 241 unit-economy/“crow epistemology”
invalid concepts 227–28 130–31
justice, concept of 231–32 “universals,” problem of 105,
vs. Kant 386, 387 127–28
knowledge 21, 22, 389 words 125–26
language 126–27 Randall, John H., Jr. 195
logic (definition) 194 “Rand’s Razor” 232–36, 257, 287
logic, contribution to 195 false division 235–36
logical positivism 389 false integration 236–38, 242–43
measurement 119–20 statement of 232
measurement-omission 115–16 see also “Package-dealing,”
metaphysical vs. man- fallacy of
made 72–73 Rangnekar, Shrikant 334 n180
on modern intellectuals 14 Rationalism 14–16, 58, 309–10, 371,
vs. Nominalism 109 378, 381, 388
objective vs. intrinsic/subjective Rationality
296–98 Law of 278–85, 288–89
objectivity 292 reality-orientation 328–29
objectivity of concepts 296–98 and volition 326–28
optional concepts 238 see also Awareness (levels of);
philosophy 317, 318 Focus (mental); Reason
primacy of existence, the 32 Ratliff, F. 68
 inde x 413

Re-affirmation through denial Science 369-71, 381–85

24–25, 27, 59, 355 Self-awareness/self-consciousness
Realism 27, 33, 54, 63, 249
direct 68–71 see also Introspection;
failure of Platonic/Moderate Representationalism
104–07, 128, 133, 141, Self-evidency
163–64, 239 and axioms 22, 24, 26–27, 355
Moderate 102–04, 106–07 defined 22, 23
Naïve 86–88, 91–93 opposition to 23–24
Platonic 101–02, 103, 296–98 and proof 22–24
Reality-orientation retroactive 146, 157
see under Rationality Self-exclusion, fallacy of 253, 356
Reason Self-produced movement and
vs. the arbitrary 282–84 percep­tion 63
biological function of 97–98 Self-reference see Pure self-reference,
Plato on 296–98 fallacy of
as volitional 99–100, 322, 326–28 Sellars, Wilfrid 208, 384 n208
see also Concepts; Principles Sensation
(as absolute); Rationality in lower animals 60
Reduction 228-29, 257, 262–64 vs. perception 59–63, 64–68
Reification, fallacy of 66 vs. physical events 64
Religion see God; Mysticism sensory inputs 60–61, 64 n22, 67
Representationalism 88–93 stimulus 60, 65
Retroactive self-evidency, fallacy of usage of term 60
146, 157 see also Sensationalism; Sensory
Re-writing reality, fallacy of 80 qualities
Riggs, L. A. 68 Sensationalism 64–68, 76, 198 n101
Rorty, Richard 383–84 J. J. Gibson vs. 68–71
Ruby, Lionel 255 n130 “zing-ping” error 65
Russell, Bertrand 236–37, 242, 248, Sense-perception see Perception
250 Sensory qualities 65–67, 79, 92–93,
Russell’s paradox 250 198 n101
Salmieri, Gregory 57 n19, 75 n31, Similarity 110–23
116 n48, 176 n89, 257 n134 and Conceptual Common
Sarnat, H.B. 59, n20 Denominator 112–17
Schwartz, Jeffrey 52 defined 110
Schwartz, Peter 252 n127, 330 n178 and definitions 221, 375
 H OW W E K N OW 414

Similarity (continued) and propositions 240–41

and difference 110–13 Top-down view of knowledge
and distinguishing characteristic 371–75, 378, 379, 381
115–17 Touch, sense of 61, 64, 67, 75
and measurement-omission Trans-temporal nature
115–17 of the conceptual see Time
objective basis of 118 Trichotomy (intrinsic-subjective-
as requiring a foil 111–17 objec­tive) 296–98
Simpson, George G. 146 n69 Trout, Monroe 192 n95
Simpson, O. J. 279, 291 Truth
Skepticism 15–16, 90 and logic 254, 268–71, 295
Skinner, B. F. 27 of propositions 239–41, 251–54
Smell, sense of 61, 67 vs. validity 241–42
Snow, John 370 Ulysses 13–14
“Some but any” principle 121–22 Understanding 209–11
127–28, 135–36, 140 Unit 117, 119–20
Spinoza, Baruch 229, 291 Unit-economy 129-36, 303-04, 365-66
Spiral process of knowledge 208–11 see also “Crow epistemology”;
Stace, W. T. 119 see also under Definition, rules of
Stalin, Josef 246 n124, 384 “Universals” 101–07, 128, 296–98
Standards 204–05, 269–71, 334, 389 “Universals,” problem of 105, 127–28
“Standing order” 201, 204 Validation 261, 264–65
Stimulus see under Sensation of a concept 231–32
Stolen concept fallacy 34–35, 46–47, as wider than proof 22–23
48, 54, 94, 148, 208, 214, 247–48 see also Invalid concepts; Proof
Subdivisions see under Values see Moral values
Abstraction from abstractions Volition see Free will/volition
Subjectivism/subjectivity 106–09, 295, Washington, George 238, 242
296–98, 388–89 Watergate 299 n156
Syllogism 194–95, 254–55 Watson, John B. 27
Tallis, Raymond 47 n17 White lie 313
Taste, sense of 61–62, 67 Widenings see under Abstraction
Teleological measurement 166–67 from abstractions
“Tiger-lily” process 160–62 Windelband, Wilhelm 347
Time Wittgenstein, Ludwig 107–09, 113,
and concepts 240–41 124, 127, 387
and generalizations 258 Words 124–27, 133, 136
and perception 63, 75, 240–41 “Zing-ping” error 65
How We Know
Epistemology
on an Objectivist
Foundation

Harry
Binswanger
 416

How We Know
H OW W E K N OW

W hat is knowledge? How is it acquired? How are claims

to knowledge to be validated? Can man achieve rational certainty,
or is he doomed to perpetual doubt?
How We Know presents an integrated set of answers to these and
related questions, based on Ayn Rand’s Objectivist philosophy, including
her unique theory of concepts. Rejecting the false alternative of mysticism
vs. skepticism, Harry Binswanger provides an uncompromising defense
of reason, logic, and objectivity.
Using vivid examples, he traces the hierarchical development of knowl­
edge, from its base in sensory perception, to concept­formation, to logical
inference, to its culmination in the principles of science and philosophy.
How We Know explains how following methods of cognition based
on the facts of reality and on the nature of our cognitive equipment makes
it possible to achieve rational certainty, no matter how abstract the issue.

H arry Binswanger earned his Ph.D. in philosophy from

Columbia University in 1973. He has taught philosophy at several
universities, most recently at the University
of Texas at Austin. Dr. Binswanger was an
associate and friend of Ayn Rand in her
final years, and since 1986 he has served on
the Board of the Ayn Rand Institute.

Photo: Mark da Cunha

He is the author of The Biological Basis
of Teleological Concepts and the editor of
The Ayn Rand Lexicon, a mini­encyclopedia
of Objectivism. He co­edited the expanded
second edition of Ayn Rand’s Introduction
to Objectivist Epistemology.
Currently, Dr. Binswanger runs HBLetter, a subscription­based website
on Objectivism, and he is a Senior Contributor at RealClearMarkets.

Jacket design and illustration: Klaus Nordby