Katsafanas P Agency Foundations of Ethics Nietzschean Constitutivism 13 PDF

Agency and the Foundations of Ethics
This page intentionally left blank

Agency and the
Foundations of Ethics
Nietzschean Constitutivism
Paul Katsafanas
1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
# Paul Katsafanas 2013
The moral rights of the author have been asserted
First Edition published in 2013
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
ISBN 978–0–19–964507–7
Printed in Great Britain by
MPG Books Group, Bodmin and King’s Lynn
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
For Danielle
Acknowledgments
During the years that I have spent writing this book, I have discussed it with almost
everyone I know and received help in ways too numerous to count. Here, I will record
some of my largest debts.
I owe special thanks to R. Lanier Anderson, Jacob Beck, Selim Berker, Maudemarie
Clark, David Dudrick, David Enoch, Ken Gemes, Charles Griswold, Louis-Philippe
Hodgson, Walter Hopp, Nadeem Hussain, Christopher Janaway, Scott Jenkins, Christine
Korsgaard, Douglas Lavin, Brian Leiter, David Liebesman, Paul Loeb, Simon May,
Richard Moran, Russell Powell, Ryan Priddle, Bernard Reginster, John Richardson,
Mathias Risse, Amelie Rorty, Carlos Ruiz, Danielle Slevens, Sanem Soyarlsan, Susanne
Sreedhar, Daniel Star, and David Velleman. Each of these individuals read and com-
mented upon substantial portions of the material that eventually became this book, and
the many insights that they offered impacted every aspect of the argument.
I am deeply grateful to Mathias Risse and two anonymous reviewers for their
exceptionally helpful commentary on the complete manuscript. Their incisive com-
ments improved the book immeasurably.
In addition, I presented portions of this manuscript at the Second Annual Rocky
Mountain Ethics Congress at the University of Colorado at Boulder, the Sixth Annual
Metaethics Conference at UW Madison, Texas Tech University, Temple University,
Vassar College, Harvard University, Boston University, and several meetings of the
APA. I thank the participants for the productive conversations that ensued.
Finally, I’d like to thank Peter Momtchiloff for his helpful guidance and advice
throughout the publishing process.
An earlier version of Chapter 5 appeared as “Activity and Passivity in Reflective

Agency,” Oxford Studies in Metaethics 6 (2011): 219–54. Parts of Chapters 6 and 7
appeared, in an earlier form, as “Deriving Ethics from Action: A Nietzschean Version
of Constitutivism,” Philosophy and Phenomenological Research 83(3) (2011): 620–60.
I gratefully acknowledge the publishers’ permission to use this material.
For a purpose—Of all actions, those performed for a purpose have been least
understood, no doubt because they have always been counted the most under-
standable and are to our consciousness the most commonplace. The great problems
are encountered in the street.
Nietzsche, Daybreak
Man is a frivolous and incongruous creature, and perhaps, like a chess player, loves
the process of the game, not the end of it. And who knows (there is no saying with
certainty), perhaps the only goal on earth to which mankind is striving lies in this
incessant process of attaining, in other words, in life itself, and not in the thing to be
attained.
Dostoevsky, Notes From Underground
Contents
Reference to Nietzsche’s Works x
Introduction 1
1. Three Challenges for Ethical Theory 6
2. Normativity as Inescapability 47
3. Constitutivism and Self-Knowledge 68
4. Constitutivism and Self-Constitution 86
5. Action’s First Constitutive Aim: Agential Activity 109
6. Action’s Second Constitutive Aim: Power 145
7. The Structure of Nietzschean Constitutivism 183
8. The Normative Results Generated by Nietzschean Constitutivism 211
9. Activity, Power, and the Foundations of Ethics 238
Appendix: Is Nietzsche Really a Constitutivist? 243

References 254
Index 265
Reference to Nietzsche’s Works
Reference edition of Nietzsche’s works

Friedrich Nietzsche: Sämtliche Werke, Kritische Studienausgabe in 15 Bänden, herausgegeben von
G. Colli und M. Montinari (Berlin: Walter de Gruyter, 1967–1977)
List of abbreviations of Nietzsche’s works

A The Antichrist, trans. W. Kaufmann (Viking, 1954)
BGE Beyond Good and Evil, trans. W. Kaufmann (Modern Library, 1968)
CW The Case of Wagner, trans. W. Kaufmann (Random House, 1967)
D Daybreak, trans. R. J. Hollingdale (Cambridge University Press, 1982)
EH Ecce Homo, trans. W. Kaufmann (Modern Library, 1968)
GM On the Genealogy of Morality, trans. W. Kaufmann (Modern Library, 1968)
GS The Gay Science, trans. W. Kaufmann (Vintage, 1974)
HC “Homer’s Contest,” trans. W. Kaufmann (Viking, 1954)
HH Human, All Too Human, trans. R. J. Hollingdale (Cambridge, 1986)
KSA Kritische Studienausgabe
NCW Nietzsche Contra Wagner, trans W. Kaufmann (Viking, 1954)
TI Twilight of the Idols, trans. W. Kaufmann (Viking, 1954)
UM Untimely Meditations, trans. R. J. Hollingdale (Cambridge, 1997)
WLN Writings from the Late Notebooks, trans. Kate Sturge (Cambridge, 2003)
WP The Will to Power, trans. W. Kaufmann and R. J. Hollingdale (Vintage, 1967)
Z Thus Spoke Zarathustra, trans. W. Kaufmann (Viking, 1954)
Note on citations
I cite Nietzsche’s works in the standard fashion. Works with continuous section
numbering are cited by section number (for example, GS 354). Works that contain
multiple parts, such as GM and TI, are cited by part and section number. For example,
GM III.1 denotes the first section of the third essay of the Genealogy. Passages from
Nietzsche’s Nachlass are cited by volume, notebook, and entry number from the
Kritische Studienausgabe. For example, KSA 9:11[293] refers to volume 9, notebook
11, note 293.
R E F E R E N C E T O N I E T Z S C H E ’ S WO R K S xi
Note on translations
Translations from KSA are my own, though I have consulted Kaufmann’s and Sturge’s
translations when available. When quoting from Nietzsche’s published works, I use the
translations listed above. In many cases, I have made minor modifications for the sake
of clarity. When I do so, I typically include the German in brackets.
Introduction
Our experience of the world is pervaded by norms. Having promised to meet a friend
for dinner, I feel obligated to do so. Upon entering a café on the heels of another
customer, I think that he should have held the door for me. Rousing myself from a
relaxing nap, I tell myself that I have reason to go to the gym. Watching the evening
news, I judge that the criminal was wrong to murder his victim. As these examples
illustrate, normative claims are ubiquitous. They inform our most unexceptional as
well as our most dire activities.
Despite their pervasiveness, however, normative claims are rather mysterious. They
purport to have a certain authority over us: they tell us how we ought to live, or which
actions we should perform, or which ends to pursue. But what justifies this authority?
What makes it the case that we should keep promises, hold doors, or go to the gym?
More momentously—what makes it the case that murder is wrong?
In its general form, this is the foundational question in ethics: how is the authority of
normative claims to be justified? Recently, a great deal of attention has been directed at
the idea that we might answer this foundational question by turning to the philosophy
of action. According to a view that I will call constitutivism, action has a certain structural
feature—a constitutive aim—that both constitutes events as actions and generates a
standard of assessment for action. We can use this standard of assessment to derive
normative claims. In short, the authority of certain normative claims arises from the
bare fact that we are agents.
Thus, the great hope of constitutivism is that an investigation of the structure of
agency will enable us to answer the foundational question in ethics. It will reveal why
certain normative claims are justified. To see how this might work, consider an
example. If you understand the nature of a game, such as chess, you thereby understand
a host of normative claims that regulate chess players. For example, part of what it is to
play chess is to aim at checkmating your opponent. This aim simply must be present in
order for a series of movements to count as an episode of chess-playing. This aim
therefore seems to generate a standard of assessment for chess players: a player is
successful to the extent that she fulfills the aim. Moreover, this aim generates reasons
for action: if the player sees that she can achieve checkmate by moving her rook to a
certain space, then she has a reason to do so.
2 I N T RO D U C T I O N
Constitutivists call aims of this sort constitutive aims. A constitutive aim is present in
every instance of the activity-type that it regulates. It is present precisely because its
presence is part of what constitutes the activity as an instance of its kind. Or, in plainer
language: if there is an activity, and participation in this activity requires having a
certain aim, then all participants in that activity are going to have the aim. They are
going to have the aim because otherwise, they wouldn’t be participants in the activity
at all. For example, chess players have the aim of checkmate because, absent this aim,
they wouldn’t be chess players.
Constitutivists hope to show that action itself has a constitutive aim. If we could
manage this—so the hope goes—then we would be able to derive normative conclu-
sions from that aim. Just as we can move from the fact that chess players aim at
checkmate to the claim that (for example) they have reason to capture their opponents’
pieces, so too constitutivists hope that we can move from the fact that action has a
constitutive aim to normative claims about what agents have reason to do. Indeed, the
most ambitious versions of constitutivism attempt to show that all normative claims are
ultimately derived from the constitutive aim of action.
In the following chapters, I will show that the attractions of constitutivism are
considerable. If constitutivism works, then it will provide a way of justifying normative
claims without positing irreducible normative truths or grounding norms merely in
subjective, variable elements of human psychology. It will thereby avoid some central
and longstanding problems in ethics.
Unfortunately, constitutivism is not well understood. Explicit defenses of the theory
are relatively new, having arisen only in the past two decades.1 Accordingly, it has been
difficult to determine what the essential elements of the constitutivist framework are.
The first two chapters address this point by examining constitutivism in isolation from
any particular view about the nature of action. I explain what is essential to the
constitutivist approach, and I show what would be necessary in order to derive
normative claims from facts about the nature of action. In addition, I address a number
of objections that have recently been leveled at the very possibility of a constitutivist
theory. I show that these objections depend on misunderstandings of the constitutivist
project, and can therefore be rebutted.
Thus, Chapters 1 and 2 explain what constitutivism is and show that a constitutivist
theory is possible. Of course, it is one thing to show that constitutivism could succeed,
and quite another to show that it actually does succeed. The hardest part of any
constitutivist theory is developing a conception of action that is minimal enough to
be independently plausible, but substantial enough to yield a constitutive aim. Chapters
3 and 4 examine David Velleman’s and Christine Korsgaard’s attempts to do so.
While Velleman’s and Korsgaard’s theories are extremely insightful, I argue that they
1
While explicit defenses of constitutivism are new, some philosophers contend that aspects of the
theory—described in different terminology—are present in the work of Plato, Aristotle, and Kant (see, for
example, Korsgaard 2009). I will argue that aspects of constitutivism are also present in Nietzsche’s work.
I N T RO D U C T I O N 3
succumb to a common problem: each theory can generate substantive normative

results only by alternating between an excessively strong conception of agency and a
much weaker conception of agency. The weaker conception of agency is all that
Velleman’s and Korsgaard’s arguments establish, while the stronger conception is
necessary for their normative conclusions to follow. Absent an argument for this
stronger conception of agency, then, these theories are unsuccessful.
If constitutivism is to succeed, then we need a new conception of agency. Thus far,
constitutivism has been strongly associated with Kantian theories of agency. As
I explain in the following chapters, Korsgaard’s version of constitutivism is avowedly
Kantian, and Velleman describes his theory as “Kinda Kantian” (Velleman 2009). But
despite a resurgence of interest in Kant, I think it is fair to say that many philosophers
believe that Kant’s moral theory fails. The Kantian theory is subject to a staggering
number of objections that strike many of us as decisive: the conception of agency upon
which Kant relies seems empirically implausible; the moral theory depends on the
assessment of maxims, and yet the idea of a maxim is terribly obscure and imprecise; the
arguments attempting to show that we are committed to the Categorical Imperative
are dubious; and even if we could solve those problems, the Categorical Imperative
itself seems to generate no substantive results.2 If constitutivism were bound up with
the Kantian enterprise, it would inherit all of these difficulties.
But I will show that it is not. Kant was right about this: certain rules of practical
reason are constitutive of agency. We can hold on to that idea while developing it in a
non-Kantian manner and grounding it in a more plausible theory of action. In pursuit
of this goal, Chapters 5 and 6 articulate a conception of agency that is indebted both to
contemporary empirical work on human psychology and to Nietzsche’s philosophical
arguments. In Chapter 5, I show that in order to account for certain empirical facts
about the nature of human agency, we must reject elements of the dominant philo-
sophical conception of reflective agency. The dominant account, which I trace to
Locke and Kant, distinguishes activity and passivity in agency and treats reflective or
deliberative acts as paradigmatic cases of agential activity. I argue that although we need
a distinction between the active and the passive in action, philosophical and empirical
considerations show that this distinction has nothing to do with whether the action was
brought about in a reflective or deliberative manner. I defend a new account of agential
activity, according to which an agent is active in the production of her action iff two
conditions are met: (i) the agent approves of her action, and (ii) further knowledge of
the motives figuring in the etiology of this action would not undermine her approval of
the action. By drawing on a psychologically realistic account of motivation and agency,
we can show that agents constitutively aim at this form of agential activity.
2
I will explore the first objection in Chapter 5. There is a vast literature on the other three objections. For
helpful introductions to these disputes, see for example Brewer (2002) on the idea of maxims, Williams
(1986) on our alleged commitment to the Categorical Imperative, and Wood (1990) on Hegel’s argument
that the Categorical Imperative generates no content.
4 I N T RO D U C T I O N
Thus, Chapter 5 argues that action has a constitutive aim. But this aim, on its own,
generates very little normative content. Its importance becomes apparent only when
we link it to another aspect of agency, which I discuss in Chapter 6.
To bring out this second aspect of agency, Chapter 6 turns to a largely untapped
source of ideas about the relationship between agency and value: the work of
Nietzsche. Nietzsche might seem an unpromising source for ideas conducive to the
defense of constitutivism; after all, he is famously skeptical of ethical theorizing, and he
flatly denies that there are any objective facts about what is valuable. However, as I will
argue in Chapter 6, Nietzsche does offer ethical ideals of his own, and his critiques of
traditional morality rely on the idea that a certain value—power, in particular—has a
privileged normative status. I will suggest a novel way of interpreting Nietzsche’s
claims: power has a privileged normative status precisely because we are committed
to this value merely in virtue of acting. Nietzsche’s obscure claim that all actions
manifest, and are to be evaluated in terms of, “will to power” can be read as an attempt
to move from a claim about the essential nature of action to a claim about value. Thus,
surprising as it may seem, I will argue that we can use a Nietzschean claim about the
constitutive features of action to derive a standard of success for action.3
In short, we can use Nietzschean considerations to show that action has a second
constitutive aim: power (this term is given a special technical sense, as I will explain
below). In defending this idea, I begin by considering Nietzsche’s baffling claims about
revaluation. Nietzsche famously argues that we must “revalue” our values, critiquing
them and in some cases replacing them with new values. I argue that Nietzsche’s
revaluations are based upon the idea that power has a privileged normative status:
power is the one value in terms of which all others values are to be assessed. If this is the
correct interpretation of Nietzsche’s ethical theory, though, it raises a question: how
could power have a privileged status, given that Nietzsche denies that there are any
objective facts about what is valuable? I argue that Nietzsche’s account of agency
provides the answer: he grounds power’s privileged status in facts about the nature of
human motivation. In particular, Nietzsche’s account of drives entails that human
beings are ineluctably committed to valuing power. So Nietzsche’s ethical theory
follows from his account of the nature of agency.
3
Even the most casual readers of Nietzsche recognize that his texts are enormously complex and
ambiguous. Decisively establishing that Nietzsche held any particular ethical view is no easy task; it requires
sustained textual analysis, reconciliation of apparently conflicting passages, reconstruction of often fragmen-
tary arguments, and so on. My primary goal in this volume is to defend a version of constitutivism, rather than
an interpretation of Nietzsche’s texts. Accordingly, I will bracket many interpretive issues in the following
chapters. I will argue that some pervasive and central ideas in Nietzsche’s texts seem deeply and obviously
inconsistent, but make perfectly good sense if we interpret Nietzsche as a constitutivist. This strongly suggests
that Nietzsche intends his ethical theory to be interpreted along constitutivist lines (though, obviously, he
does not use this terminology). However, providing sufficient textual analysis to establish that this is
Nietzsche’s actual view would take us too far afield. Accordingly, skeptical readers can treat my reading as
a way of developing some of Nietzsche’s central ideas, rather than as an explication of Nietzsche’s actual view.
I return to these matters in Chapter 6 and in the Appendix.
I N T RO D U C T I O N 5
Thus, if we accept a Nietzschean account of agency—as I argue that we should—

then power turns out to be a constitutive aim of action. But what exactly is “power”?
We might suppose that valuing power denotes valuing conquest, mastery of others,
and so forth. But this is not what Nietzsche means. Power is a term of art, for
Nietzsche; he gives it a special sense. To will power is to aim at encountering and
overcoming resistance in the course of pursuing other, more determinate ends. In other
words, to say that we will power is to say that whenever we will an end, we aim not
merely to achieve the end, but also to encounter and overcome resistances that arise in
the pursuit of the end. For example, to say that I will power in the pursuit of writing
this book is to say that I will not only to complete the book, but also to encounter and
overcome challenges or resistances in the course of doing so.
The Nietzschean account of agency entails that all actions have this structure:
whenever we will any determinate end at all, we also will to encounter and overcome
resistance in the course of pursuing that end. Although this claim is paradoxical, I hope
to show that it is supported both by compelling philosophical arguments and by recent
empirical work on human psychology.
Thus, I will use a Nietzschean theory of action to argue for a bipartite constitutivist
theory: our actions aim both at agential activity and at power. In Chapters 7 and 8,
I argue that this theory generates a range of substantive normative claims. I show that
while some of these claims conform to our ordinary thoughts about what there is
reason to do, others are quite surprising. In particular, this constitutivist theory requires
a reassessment of some of our most cherished values, such as the positive valuation that
we place on certain forms of egalitarianism and the negative value that we place on
certain forms of pain. I close, in Chapter 9, by discussing the advantages that this
Nietzschean version of constitutivism enjoys over competing ethical theories.
The hope, then, is that by drawing on a roughly Nietzschean theory of agency, we
can answer the foundational question in ethics, showing how normative claims are
justified. In particular, we can justify normative claims by showing that every agent
aims jointly at activity and power.
1
Three Challenges for
Ethical Theory
The most gripping and persistent philosophical problems arise when we have strong
and unshakeable convictions that certain claims must be true, and yet, upon reflection,
we cannot see how they could so much as aspire to truth. So, freedom is a philosophical
problem because our practices and the experiences of deliberation firmly wed us to the
idea that we must have a distinct form of control over our own actions, and yet a
realistic view of the world seems to commit us to seeing all of our actions as determined
by events not under our control. Consciousness is a philosophical problem because our
understanding of the ways in which our brains process stimuli is ever increasing, and yet
we want to say that these results still leave it utterly mysterious why some of these
processes are accompanied by conscious awareness.1 And, turning now to our topic,
morality is a philosophical problem because we want to say that there are universally
valid normative facts, and yet we cannot see how such facts could be woven into the
fabric of the universe. We want to say that needlessly inflicting suffering on innocents is
wrong, universally wrong, for all people and all times; and yet even the most appalling
torments do not have inscribed on them “you shall avoid me.”
At the close of the eighteenth century, it was still possible for Kant to claim in all
earnestness that “two things fill the mind with ever new increasing wonder and awe,
the oftener and more steadily we reflect on them: the starry heavens above me and the
moral law within me” (Critique of Practical Reason, 5:161). Today, many find that steady
reflection on morality inspires less wonder and awe than skepticism and detachment.
A host of studies purport to show that moral beliefs are relics of affects and dispositions
instilled deep in our evolutionary past; others argue that morality is a vestige of religion;
and philosophers such as Nietzsche predict that we will unlearn our awe of morality,
just as past ages unlearnt their awe of astrology and alchemy (HH 4, BGE 32 and 188).2
1
Huxley and Youmans wrote: “what consciousness is, we know not; and how it is that any thing so
remarkable as a state of consciousness comes about as the result of irritating nervous tissue, is just as
unaccountable as the appearance of the Djin when Aladdin rubbed his lamp in the story” (Huxley and
Youmans 1868, 178).
2
For an excellent overview of the evolutionary arguments, see Joyce (2006, Chapter 6). For an example
of the religious arguments, see Anscombe (1958), Weber (2002), or virtually any of Nietzsche’s works.
T H R E E C H A L L E N G E S F O R E T H I C A L T H E O RY 7
What sparks this skepticism about morality? Why have many reflective individuals
gone from seeing in morality a source of awe to finding it dubious and archaic?
Doubtless the reasons are manifold, but in the following pages I will trace one answer.
An adequate account of morality would have to overcome three challenges: it would
have to show why we should have confidence in our moral beliefs, why these moral
beliefs don’t rely on outmoded or outlandish metaphysical claims, and why we should
take morality to be prescriptive. I call these the epistemological, metaphysical, and
practical challenges. Versions of these challenges are familiar in the literature on ethics,
but I will argue that Nietzsche presents especially powerful forms of each challenge.
Moreover, I will argue that the chief ethical theories encounter serious difficulties in
trying to overcome these challenges. It is the perceived inability to meet these
challenges that leads many individuals to moral skepticism. But I will also argue that
this result is not inevitable: we can avoid skepticism by pursuing a different strategy in
ethics—constitutivism.
1. Three challenges for ethical theory

Before beginning, a word on what morality is. Moral claims are normative. They
purport to have a certain authority over us. They purport to be claims according to
which we should regulate or guide our actions. But not all normative claims are
classified as moral claims. The claim “you should drink coffee” is not typically taken
as moral; the claim “you should not murder innocent children” is. So what distin-
guishes moral claims from non-moral normative claims?
The answer to this question is not obvious. However, many philosophers believe
that there is at least one necessary condition for a normative claim’s being a moral
claim: universality. Moral claims are universal in the sense that they apply to all agents.
We typically take only some agents to have reason to drink coffee, but we hold that all
agents should refrain from murdering innocent children. The claim about drinking
coffee has a suppressed premise: if you enjoy the taste of coffee (or want to wake up,
or . . . ), then you should drink coffee. The claim about murder does not; it purports to
apply to all agents, regardless of facts about their preferences, goals, and characters. After
all, we do not take the fact that someone has a desire to murder as undermining the
authority of the claim that murder is wrong.3
3
For this reason, some philosophers take moral claims to be categorical. A normative claim is categorical if it
applies to agents regardless of their preferences, goals, or aims. Notice that a normative claim can be universal
without being categorical: for example, if a normative claim applies to agents in light of their having a
particular aim, and if this particular aim is present in all agents, the normative claim will be universal but not
categorical. I won’t be assuming that moral claims are categorical, though I discuss the possibility both below
and in Chapter 4. In addition, it is worth noting that some philosophers argue that moral claims have yet a
third distinguishing feature: overridingness. That is, moral claims either always or typically trump competing
normative claims: if I am faced with a conflict between fulfilling a moral demand and some other demand,
I am obligated to fulfill the moral demand. This feature is controversial; many philosophers offer powerful
arguments against it (Sidgwick 1981 is a classic example). I will not assume that moral claims are overriding.
8 T H R E E C H A L L E N G E S F O R E T H I C A L T H E O RY
Universality alone may not be enough to distinguish moral claims from other
normative claims. Consider the claim “you ought to regulate your beliefs in accord-
ance with the relevant evidence.” That’s a plausible candidate for a universal normative
claim, but it would be unusual to classify it as a moral claim. So we might need to say
something about the content of universal normative claims. Here, there are several
options. We might think that moral claims govern our interactions with other agents;
paradigmatic moral claims might then be prohibitions on harming others, requirements
to aid others, and so forth. Or we might think that moral claims specify what it is to live
well or to flourish; paradigmatic moral claims might then tell us to seek happiness, or
achievement, or to avoid squandering our capacities.
I think that any recognizably moral claim will have some such content; it will either
govern our relationships with other agents or specify what it is to live well. However,
I do not want to assume, in advance, that moral claims have any specific content. I will
not assume, for example, that we can justify a universal normative claim requiring us to
help others, or to be compassionate, or to cultivate our talents. So I’ll start with a rather
loose definition of morality: moral claims are universal normative claims that either
specify appropriate behavior toward other agents or specify what it is to live well. We’ll
refine this notion as we progress.
Below, I will consider three challenges for morality. That is, I will consider three
challenges to the very idea that there can be universal normative claims specifying
appropriate behavior toward others or specifying what it is to live well. I will then ask
whether the dominant ethical theories give us a way of answering these challenges.
1.1 The epistemological challenge
The first challenge for morality arises from a simple fact: morality has a history.4 To
illustrate the relevance of this point, let’s consider Nietzsche’s On the Genealogy of
Morality. As is well known, the Genealogy sets out to demonstrate that many of our
most basic moral beliefs arose approximately two thousand years ago as the product of a
resentment-inspired revolt carried out by a lackluster, vengeful underclass. Put briefly,
Nietzsche’s argument is as follows. In the ancient world, the dominant moral code was
organized around a good/bad dichotomy, where the traits labeled “good” were those
associated with the nobility, and those labeled “bad” were those associated with the
commoners. Strength, self-assertion, power, desire to rule, competition, wealth,
4
It is important to notice that the term “morality” can be used in either a descriptive or a normative sense.
Descriptively, the term refers to the moral code that is generally accepted within a particular society or social
group at a particular time. Thus, we might speak of the morality of the fifth-century Athenians, the morality
of the antebellum South, and so on. Normatively, the term “morality” refers to a putatively correct moral
code. These can come apart: while the fifth-century Athenians accepted moral claims such as “slavery is
permissible,” one hopes that this was an error; one hopes that they ought to have accepted the claim that
slavery is impermissible. When I say, above, that morality has a history, I am starting with the descriptive sense
of morality.
health, and beauty were taken as good; weakness, humility, lack of power, servility,
inability to compete, ordinariness, and ugliness were taken as bad. Nietzsche argues
that this system of evaluations was, for a time, accepted by the bulk of society: slaves,
commoners, and nobles all embraced these values. About two thousand years ago,
Nietzsche claims, things began to shift. A new set of values emerged. These new
values—which Nietzsche calls the “good/evil” or “slave” morality—invert many of
the earlier values. Thus, manifestations of strength, self-assertion, power, desire to rule,
competition, wealth, and certain forms of beauty are taken as bad; weakness, meekness,
humility, lack of power, servility, poverty, compassion, concern with suffering, and an
idea of equality are taken as good.5 These values are most clearly present in early
versions of Christianity, but many of them remain today.6
To see how Nietzsche’s story constitutes a challenge to moral philosophy, we must
focus on three crucial claims. First, ancient and modern moralities endorse distinct and
conflicting sets of values. Second, an examination of morality’s history reveals that
these changes in the moral code cannot be construed as mere refinements of earlier
values, but must instead be seen as discontinuous breaks. Third, these breaks are best
explained by psychological and social factors rather than by appreciation of rational
considerations. I will explain each of these claims below.
I take it that the first claim is beyond dispute, but let me illustrate it with a
non-Nietzschean example. To this end, consider Aristotle. Aristotle claims to be
systematizing certain culturally pervasive intuitions (endoxa), rather than developing
revisionary moral claims, so his writings are a particularly helpful guide to views
prevalent in the ancient world. Aristotle heaps praise on a character trait that he calls
megalopsychia. Perhaps the best translation for this Greek word is “greatness of soul.”
Megalopsychia, Aristotle tells us, is “a sort of crown of the virtues; for it makes them
greater, and is not found without them” (Nicomachean Ethics 1124a1). What distin-
guishes this character trait? Aristotle remarks that the individual with megalopsychia
distinguishes between persons of high and low rank (1124b15–25); he disdains the
honor and praise emanating from persons of low rank, and accepts the praise of his
high-ranked peers (1124a5–11); and he is motivated by honor “again, it is characteristic
of the [man of megalopsychia] . . . to be sluggish and to hold back except where great
honor as a great result is at stake, and to be a man of few deeds, but of great and notable
ones” (1124b22–26). So, the individual with megalopsychia is acutely sensitive to
social hierarchy and motivated by the desire to be perceived as honorable by those of
5
I am eliding a complication that won’t be relevant for our purposes: Nietzsche claims that slave morality
replaces the concept of bad with the concept of evil. For helpful analyses of this point, see Reginster (1997)
and Leiter (2002).
6
Nietzsche’s arguments for these claims are given in GM I. For helpful discussions, see for example
Reginster (1997), Ridley (1998), Leiter (2002), Janaway (2007), Owen (2007), and Wallace (2007). I address
these claims in more detail in Katsafanas (2011d).
high social rank. We might put the point less kindly: honor-seeking elitism is the
“crown of the virtues.”7
Clearly, modern individuals do not take these character traits to be virtues, much less
the highest virtues. Indeed, our modern moral code would be more likely to deem this
type of pride as a vice. The Bible is illustrative: we are told that “God opposes the
proud but gives grace to the humble” (1 Peter 5:5).
Speaking of humility—Aristotle labels it a vice:
for the unduly humble man, being worthy of good things, robs himself of what he deserves, and
seems to have something bad about him from the fact that he does not think himself worthy of
good things, and seems also not to know himself; else he would have desired the things he was
worthy of, since these were good . . . Such a reputation, however, seems actually to make them
worse; for each class of people aims at what corresponds to its worth, and these people stand back
even from noble actions and undertakings, deeming themselves unworthy, and from external
goods no less. (1125a19–27)
The shifting valuations attached to humility and greatness of soul are evidence in favor
of Nietzsche’s first claim: ancient and modern moralities endorse conflicting values.8
Examples could be multiplied.9
So the first claim is uncontroversial. But it might also seem insignificant. After all, the
claim that beliefs about morality have changed over time is not surprising. The same is
true of every field of human inquiry. Our current beliefs about physics, chemistry,
biology, and so on also grew out of earlier forms.
But this brings us to the second point: Nietzsche argues that history reveals not a
smooth process of rational development, but a series of discontinuities between the
moral beliefs embraced at different times. The later moral systems cannot be under-
stood as rational developments of the former, but must be seen as distinct. To see what
Nietzsche has in mind, notice that certain valuations can be understood as natural
extensions or developments of earlier valuations. When a government moves from
claiming that all property-holding white males deserve equal treatment to claiming that
all persons, regardless of wealth, race, or gender, deserve equal treatment, this can be
seen as a development of the internal logic of valuing equality: an ideal that is implicit
or imperfectly realized in the earlier moral code is developed, rendered more consist-
ent, and made fully explicit. Cases of this sort give us no reason for skepticism about our
7
There is a tradition of interpreting megalopsychia in ways that make it seem less objectionable to modern
sensibilities: see, for example, Crisp (2006) and Sarch (2008). With Cordner (1994), I find it preferable simply
to accept that Aristotle’s virtues differ from those that we would endorse. For additional reflections on the
striving for something like megalopsychia in the ancient world, see Nietzsche, Homer’s Contest.
8
Of course, I am not claiming that the ancient and modern worlds were univocal in their respective
valuations. For example, by the early modern period we can find a number of philosophers arguing that
humility is not a genuine virtue, and today our attitude toward humility is decidedly mixed.
9
Consider a few obvious examples: the valuations attached to slavery, torture, public execution,
cannibalism, imperialism, monogamy, sexual promiscuity, masturbation, and homosexuality have undergone
dramatic shifts over time.
values. But other values are not like this. We moved from considering elitism as good
to considering it evil; we moved from considering humility bad to considering it a
central virtue. It is difficult to see this as anything less than a complete inversion of these
values. Nietzsche’s claim is that many of the changes from classical to modern moral
codes represent just such profound breaks, rather than a continuous process of rational
development.10
Of course, it is compatible with this story that our current moral beliefs are better
aligned with the truth. After all, there are also discontinuities in science: it has been
clear since Kuhn (1970) that the move from Newtonian to relativistic physics cannot
be understood as a process of smooth continuous development, but instead reveals
discontinuities and breaks. Nevertheless, we take relativistic physics to provide a more
accurate representation of the physical world than does Newtonian physics. The same
could be true of our modern moral code.11 For this reason, the presence of breaks and
discontinuities does not by itself imply that the system in question is problematic.
Nietzsche is aware of this. Indeed, he claims that “there is no more important
principle” for the study of history than this: “the cause of the origin of a thing and
its eventual utility, its actual use and arrangement in a system of purposes, lie worlds
apart; whatever exists, having somehow come into being, is again and again reinter-
preted to new ends, taken over, transformed, and redirected” (GM II.12). In other
words, the history of all systems of beliefs will display discontinuity, alogical leaps, and
so forth. “The entire history of a ‘thing,’ an organ, a custom can in this way be a
continuous chain of signs of ever new interpretations and adjustments, whose causes do
not even have to be related to one another but, on the contrary, in some cases succeed
and alternate with one another in a purely chance fashion” (GM II.12). Given the
ubiquity of alogical developments, Nietzsche warns us against committing the genetic
fallacy:
The inquiry into the origin of our evaluations and tables of the good is in absolutely no way the
same as a critique of them, as is so often believed: even though the insight into some pudendo origo
certainly brings with it a feeling of diminution in the value of the thing that originated in that
way and prepares the way to a critical mood and attitude toward it. (KSA 12:2[189]/WLN 95)12
As Nietzsche notes here, the discontinuities alone don’t undermine our moral code.
10
Note that this is consistent with other values being held constant. Nietzsche’s claim is not that every
valuation present in the ancient world was inverted or altered in the modern world; his claim is simply that
many of the central valuations were inverted or altered.
11
Indeed, we can put Nietzsche’s point in terms familiar from the philosophy of science: changes in moral
codes resemble paradigm shifts rather than normal science.
12
Compare GS 345: “Even if a morality had grown out of error, this would not so much as touch on the
problem of its value. . . . The mistake made by the more refined among them [historians of morality] is that
they uncover and criticize the perhaps foolish opinions of a people about their morality, or of humanity about
all human morality—opinions about its origin, religious sanction, the superstition of free will, and things of
that sort—and then suppose that they have criticized the morality itself.”
The discontinuities revealed by history do, however, prompt us to ask why these
shifts occurred. As Nietzsche puts it, history makes us feel that the value of the thing in
question has diminished, and thereby “prepares the way to a critical mood and attitude
toward it.” In short, these histories make us wonder why we hold the moral beliefs that
we do.
With shifts in the sciences, we at least have some conception of what motivated
changes in theories and what would serve as a check on an incorrect theory: we can
appeal to explanatory adequacy, coherence, simplicity, and so forth. There are well-
known problems with these standards, and they clearly won’t be sufficient conditions
for deciding between competing theories, but we at least have a sense of what we are
after.13 With morality, though, what are we after? Conformity to moral intuitions?14
Conformity to emotions such as compassion or sympathy? The promotion of some
end, such as long-term self-interest or happiness? It is not obvious: each of these
purported standards of success is highly controversial—far more controversial than in
the scientific case. Indeed, each of these standards would be rejected by proponents of
certain moral codes.15 Below, we’ll see that Nietzsche calls all of these grounds into
question.16,17
If we had a theory-neutral criterion of success, then discontinuities in moral codes
would not be troubling. For we could use the criterion to determine whether, despite
the alogical leaps, the successive moral codes were getting closer to or further from the
truth. But we don’t have that—unless one of the moral theories that I will discuss
below can provide it.
13
For some classic discussions of the problems with these standards, see, for example, Popper (1959),
Kuhn (1970), and Laudan (1977).
14
Some philosophers do try to draw analogies between the role of data in science and in morality. But, as
Peter Singer notes, this analogy is at best highly strained: “The analogy between the role of a normative moral
theory and a scientific theory is fundamentally misconceived. A scientific theory seeks to explain the
existence of data that are about a world ‘out there’ that we are trying to explain. Granted, the data may
have been affected by errors in measurement or interpretation, but unless we can give some account of what
the errors might have been, it is not up to us to choose or reject the observations. A normative ethical theory,
however, is not trying to explain our common moral intuitions. It might reject all of them, and still be
superior to other normative theories that better matched our moral judgments . . . ” (Singer 2005, 345).
15
We might think that we could avoid this problem by claiming that morality aims at human flourishing.
But as soon as the notion of flourishing is given any substantive content, this claim becomes controversial.
I address this point below and, in a rather different way, in Katsafanas (2011d).
16
Thus, in Daybreak 106 Nietzsche remarks that morality is presented as enhancing or preserving
mankind. But “preservation of what? Is the question one immediately has to ask. Advancement to what? is
the essential thing—the answer to this of what? and to what? not precisely what is left out of the formula?” He
goes on to note that there are many potentially conflicting goals here.
17
Philosophers sometimes appeal to convergence in moral beliefs in order to support the idea that we’re
moving increasingly closer to a correct moral theory. Mere convergence won’t be convincing, though, unless
we have some reason for thinking that we aren’t converging toward errors. After all, from late antiquity to the
Middle Ages, we can see the moral codes of various European nations as converging toward Judeo-Christian
values, and yet we can give an obvious explanation for this convergence: the spread and increasing political
and cultural influence of Christianity.
Nietzsche’s third point becomes relevant here. If these changes in moral codes didn’t
occur in response to rational reflection, what prompted them? Nietzsche argues that
attention to history reveals that moral systems arose and changed for psychological and
social reasons. The ancient nobility, Nietzsche tells us, affirmed their own way of life
and hence deemed it good.18 Their positions of authority enabled them to promulgate
these values. The fundamentals of our modern moral code arose when a vengeful
underclass rejected the earlier moral system; consequently, it rests on a psychological
state that Nietzsche calls ressentiment.19 It took hold for social reasons: a set of priestly
figures preached values that would appeal to the oppressed, downtrodden servant
classes, who constituted the bulk of society.20 Thus, rather than arising in response to
appreciation of rational considerations, moral shifts occurred for contingent psycho-
logical and social reasons. If this is correct, it should undermine our confidence in them.
Aspects of Nietzsche’s story may be fanciful, and at the very least we would need
more evidence for these claims than Nietzsche himself provides. But we can set that
aside; the details are irrelevant for our purposes. What matters is the truth of a general
claim: there are discontinuities in the development of moral codes, and these discon-
tinuities are best explained by psychological and social considerations. Do we have
reason to believe this is true?
I think we do. Even if Nietzsche’s evidence for this claim is rather spare, it does not
stand alone. We have Marx’s arguments that moral shifts are best explained by
economic factors.21 We have Weber’s arguments linking some of our central moral
beliefs to religious assumptions.22 We have evidence from anthropology and evolu-
tionary biology that certain moral beliefs arose in response to highly contingent and
now vanished circumstances, such as conditions of low population density or the
authority of various religions.23 And there are simpler ways of making the same
point. Gilbert Harman gives a very nice example. Most cultures seem to accept a
judgment of the following form: harming someone is worse than failing to help someone. For
example, if I kill someone, this is terrible; but if I know that an individual will starve to
death unless I donate a negligible amount of money that I would otherwise waste on
frivolous entertainments, and nonetheless fail to donate the money, this is widely
regarded as perfectly acceptable. This is rather odd: the consequences are the same,
after all. There is a voluminous literature investigating potential justifications for the
claim. But Harman points out that we can give an exceedingly simple and elegant
18
“The noble type of man experiences itself as determining values . . . Everything it knows as part of itself it
honors: such a morality is self-glorification” (BGE 260).
19
See GM I.7–11. “The slave revolt in morality begins when ressentiment becomes creative and gives birth
to values” (GM I.10). Ressentiment is Nietzsche’s term for a vengeful hatred born of impotence; see Reginster
(1997), May (1999), and Wallace (2007) for discussions.
20
For discussions of this point, see the Genealogy and the Antichrist.
21
For introductions to these ideas, see Cohen (2001) and Wolff (2002).
22
See, for example, Weber (2002).
23
See Prinz (2007, 220–87) for a helpful overview. See also Joyce (2006) and Sinnott-Armstrong (2007).
explanation for why this valuation became widespread: it aids the wealthy and
powerful.
Whereas everyone would benefit equally from a conventional practice of trying not to harm
each other, some people would benefit considerably more than others from a convention to help
those who needed help. The rich and powerful do not need much help and are often in the best
position to give it; so, if a strong principle of mutual aid were adopted, they would gain little and
lose a great deal, because they would end up doing most of the helping and would receive little in
return. On the other hand, the poor and the weak might refuse to agree to a principle of non-
interference or noninjury unless they also reached some agreement on mutual aid. We would
therefore expect a compromise . . . the expected compromise would involve a strong principle of
noninjury and a much weaker principle of mutual aid—which is just what we now have.
(Harman 1977, 111)
In this passage, Harman provides a mini genealogy of one moral principle. Just as with
Nietzsche, Harman’s story doesn’t show that the moral principle is false. But it does
make us wonder whether any good reasons can be given for its acceptance. In general,
if we have a powerful social or psychological explanation for why we hold a value, and
we have difficulty seeing what independent grounds can be given for the value’s
acceptance, then our confidence in the value should be undermined.
To sharpen this Nietzschean argument, it will be helpful to contrast it with the traditi-
onal argument from disagreement. John Mackie gives a classic statement of that argument:
M1. There is moral disagreement: different cultures exhibit different moral beliefs.
M2. The best explanation for this disagreement is that there are no objective facts
about morality.24
M3. Therefore, there are no objective facts about morality.25
This argument from disagreement has come under criticism.
Some critics have objected that there really isn’t so much disagreement about
morality. In particular, apparent disagreements about values often turn out to be
based on factual disagreements or ignorance. For example, consider the claim that
slavery is permissible. Some individuals who endorsed this judgment supported it
with erroneous factual claims. Aristotle claimed that certain individuals had physical
and psychological aspects that rendered them “natural slaves”;26 analogously, some
24
Mackie writes, “the argument from relativity has some force simply because the actual variations in the
moral codes are more readily explained by the hypothesis that they reflect ways of life than by the hypothesis
that they express perceptions, most of them seriously inadequate and badly distorted, of objective values”
(Mackie 1977, 37).
25
Loeb (1998) offers a different version: if moral realism is correct, then moral questions must be in
principle resolvable; but a number of moral questions are not even in principle resolvable; so moral realism is
false. See Leiter (forthcoming b) for a version of this argument that focuses on putatively irresolvable
disagreement about moral theories rather than moral beliefs.
26
Aristotle asks “is there any one thus intended by nature to be a slave, and for whom such a condition is
expedient and right”? He answers, “There is no difficulty in answering this question, on grounds both of
American slaveholders claimed that slaves were cognitively inferior and therefore were
best served by slavery. It is possible that if these false factual beliefs had been corrected,
then Aristotle and some of the American slaveholders would have abandoned their
belief that slavery was justified. Generalizing this point, some philosophers respond to
Mackie’s argument by rejecting M2: these philosophers claim that if we get all of our
non-moral facts straight, moral disagreement will vanish.
Another explanation for moral disagreement is simply that morality is hard. After all,
there has also been a great deal of disagreement about physics, biology, economics, and
so forth. We don’t readily conclude, from the fact that different cultures or different
times have disagreed about the nature of physical reality, that there are no objective
facts about physics. We don’t readily conclude, from the fact that economists disagree
about which tax policy would maximize GDP, that there is no fact of the matter about
which tax policy would maximize GDP. Morality might have an analogous explan-
ation: it is complex and difficult.
Moreover, in morality as in economics there are clear pressures toward self-decep-
tion. When we consider a question such as “do low tax rates on the wealthy maximize
GDP?,” the wealthy have a clear stake that might bias their answers. Just so, when we
consider certain moral questions, agents have clear stakes that might bias their answers.
Consider one example: the ancient Greeks and the early American slaveholders were
under pressure to regard slavery as permissible because it was economically advanta-
geous for those involved in certain modes of agriculture. These pressures could explain
why disagreement arises.
In short, critics allege that the best explanation for moral disagreement is a combin-
ation of factual disagreement, difficulty of the subject matter, and pressures toward self-
deception. These criticisms have made the traditional argument from disagreement
seem less persuasive. But notice that Nietzsche’s argument is very different. We can
start with Mackie’s M1:
M1. There is moral disagreement: different cultures exhibit different moral beliefs.
From there, Nietzsche’s argument takes a different turn:
N2. These differences in moral beliefs were caused by social and psychological
factors.
N3. In general, the kinds of social and psychological factors that shift moral beliefs
do not track the truth.
N4. If we recognize that a belief was caused by factors that do not track the truth,
then we need justification for continuing to hold this belief.
N5. Therefore, we need justification for continuing to hold our moral beliefs.27
reason and of fact. For that some should rule others and not be ruled is a thing not only necessary, but
expedient; from the hour of birth, some are marked out for subjection, others for rule” (Politics 1254a).
27
For a related reading of Nietzsche’s argument, see Sinhababu (2007, 276–9). Sinhababu provides a
helpful discussion of the ways in which the presence of unreliable psychological processes in belief-formation
If Nietzsche is right about N2, then claims about the difficulty of morality will be
irrelevant. And indeed, the responses to Mackie’s argument are committed to some
version of N2: the respondents want to show that certain moral beliefs, which they
regard as disagreeable (such as Aristotle on slavery) were explained by social and
psychological factors, whereas others (such as all of ours) were not. Once we’ve
admitted the possibility of influence, though, we need some reason for thinking that
it does not affect our own moral beliefs.
So M1 and N2 seem well supported. We can hardly deny N3. No one is going to
argue that resentment, desire for power, desire for economic advantage, and so forth
are psychological mechanisms that track the truth. N4 also seems uncontroversial. For
these reasons, Nietzsche’s argument won’t be defused by the traditional objections to
Mackie’s argument from disagreement.28 Nietzsche’s argument poses a genuine chal-
lenge for moral philosophy.29
With these points in mind, we can summarize the epistemological challenge as
follows: attention to the way in which morality developed undermines our confidence
in and justification for our current evaluations. If our commitment to all of our values
can be explained in the above fashion, then the worry is that we will not be able to
sustain our commitments. After all, morality is demanding. It tells us how to live. It tells
us how to structure our interpersonal relationships. Kant claims that it aspires to
overrule any competing inclinations: the thought of moral duty “strikes down all
arrogance as well as vain self-love” (Kant, Critique of Practical Reason, 5:86). All of this is
right. The question is whether morality can continue to occupy these roles once we
appreciate its origins. Nietzsche thinks not: he tells us that
If you had thought more subtly, observed better, and learned more, you certainly would not go
on calling this ‘duty’ of yours and this ‘conscience’ of yours duty and conscience. Your
knowledge of the way in which moral judgments have originated would spoil these grand
words for you, just as other grand words, like ‘sin’ and ‘salvation of the soul’ and ‘redemption’
have been spoiled for you. (GS 335)
Below, we will ask whether Nietzsche is right.
should undermine our faith in the belief. See also Leiter, who writes, “we should be suspicious of the
epistemic status of beliefs that have the wrong causal etiology” (2006, 104).
28
Enoch (2009) attempts to rebut all forms of the argument from disagreement. He considers ten versions
of the argument and offers responses to each. Despite Enoch’s aspiration to comprehensiveness, however, he
does not consider anything like the Nietzschean argument from disagreement mentioned above.
29
Nietzsche’s argumentative strategy is clearly presented in the following passage: “Historical refutation as
the definitive refutation.—In former times, one sought to prove that there is no God—today one indicates how
the belief that there is a God could arise and how this belief acquired its weight and importance: a counter-
proof that there is no God thereby becomes superfluous.—When in former times one had refuted the ‘proofs
of the existence of God’ put forward, there always remained the doubt whether better proofs might not be
adduced than those just refuted: in those days, atheists did not know how to make a clean sweep” (D 95). In
this passage, Nietzsche indicates that seeing how the belief in God originated undermines the belief—not
because it shows that the belief is false, but because it shows that the belief arose for dubious reasons.
1.2 The metaphysical challenge

So far, we have one criterion of adequacy for an ethical theory: it must explain why,
despite the discontinuities in moral beliefs and psychological explanations for these
discontinuities, we should have confidence in our current moral beliefs. This brings us
to the second challenge. To introduce the challenge, it is helpful to begin with an
objection. When speaking to educated, non-religious individuals outside of philoso-
phy departments, there is a very common reaction to claims about morality: the idea of
universal values is antediluvian, a relic of discredited religious or outmoded scientific
accounts of the world. The idea that there are any objective facts about what we should
do, or what is valuable, is just one last form of anthropocentrism lurking among
the scientifically ignorant. A realistic, empirically informed account of morality
shows it to be nothing more than a series of conventions and customs, devoid of any
deeper justification.
Why might the idea of universal values seem outmoded? The problem is one of
naturalism. We want a theory that is compatible with our best account of the natural
world, and morality seems to face two problems on this score.
First, some attempts to justify conventional morality appeal to properties that seem
fanciful. Again, John Mackie gives a classic formulation of this objection: “if there were
objective values then they would be entities or relations of a very strange sort, utterly
different than anything else in the universe” (1977, 38). For if such values existed, then
it would be possible for a certain state of affairs to have “a demand for such-and-such an
action somehow built into it” (1977, 40). And this, Mackie concludes, would be a
decidedly “queer” property.
Of course, there are controversies regarding which qualities should count as queer.
Contemporary physics posits a number of bizarre properties. But consider just how odd
moral facts would be: they would be facts with intrinsic prescriptivity or imperative-
ness. A moral fact would be something that directs us or commands us to act in a certain
way.30 This is what makes purported moral facts queer: they demand that we do
something, independently of any facts about our motives or goals. And it’s hard to
see what kind of facts or properties could have this intrinsic demandingness built into
them. Richard Garner puts the point well: “it is hard to believe in objective pre-
scriptivity because it is hard to make sense of a demand without a demander, and hard
to find a place for demands or demanders apart from human interests and conventions.
We know what it is for our friends, our job, and our projects to make demands on us,
but we do not know what it is for reality to do so” (Garner 1990, 143). In short, what’s
queer is the idea of intrinsic prescriptivity lodged in the world.
30
Philosophers sometimes express this point in terms of motivation: appreciation of a moral fact is
supposed to be capable of motivating the agent. For now, I want to set aside questions about whether
morality necessarily motivates, and focus instead on the fact that a demand for motivation is somehow built
into the moral fact. I’ll address the question about whether moral facts are necessarily motivating in the next
section.
This brings us to the second way in which moral philosophy can seem to run up
against naturalistic troubles: certain attempts to justify conventional morality make
presuppositions about human beings that are either demonstrably false or otherwise
problematic. For example, Aristotle’s moral theory relies on an outmoded natural
teleology that implies that human beings have a function. Kant is committed both to
the idea that we can draw a clean distinction between reason and passion, and that all
actions are performed on maxims. Contemporary biology and psychology give us
reason to doubt each of these claims.31 Insofar as the justifications of moral theories
require or presuppose indefensible claims about human beings, they are unacceptable.
In sum, then, we have two requirements on an adequate moral theory. First, the
theory must be metaphysically respectable: the account of reasons and values must not
appeal to any non-natural qualities. Second, the theory must be psychologically
realistic: the account of reasons and values must not presuppose a model of agency or
human psychology that is ruled out by our best philosophical and scientific accounts.
The psychological requirement seems to me the more difficult one. We will see
below that most of the dominant ethical theories can avoid appeal to non-natural
qualities, and thereby meet the metaphysical constraint. However, the psychological
constraint is a substantial impediment.
Nietzsche is again relevant here, for he is focused mainly on the psychological
point.32 He takes it for granted that we should avoid metaphysically extravagant
properties, dismissing views that posit “intercourse between imaginary beings” and
31
For criticisms of Aristotle along these lines, see Williams (1986, Chapter 3); for Kant, see Blackburn
(2001, Chapter 8), Leiter (2002), and Risse (2007).
32
Nietzsche endorses some version of naturalism, but it is not obvious which version. Brian Leiter has
argued that Nietzsche is a “methodological naturalist”; that is, Nietzsche thinks “philosophical inquiry . . .
should be continuous with empirical inquiry in the sciences” (Leiter 2002, 5). He adds that Nietzsche is best
viewed as a “speculative methodological naturalist” (2002, 5, emphasis added). Speculative naturalists do not
merely take the current scientific discourse to be correct, but aim to go further; they “construct theories that
are ‘modeled’ on the sciences . . . in that they take over from science the idea that natural phenomena have
deterministic causes” (Leiter 2002, 5). This seems to me entirely correct, but it does leave open the difficult
question of what counts as “modeling” a philosophical theory on the sciences.
There is no denying that Nietzsche was fascinated with the sciences of his day; he writes, in Ecce Homo, “A
truly burning thirst took hold of me: henceforth I really pursued nothing more than physiology, medicine and
natural sciences” (EH III: HH-3). But, at the same time, he criticizes much of the science that was current in
his time, dismissing the “clumsy materialists” who “can hardly touch on the soul without immediately losing
it” and inveighing against “materialistic atomism” (BGE 12). One important aspect of this critique is that
unlike some naturalists, Nietzsche has no aspirations to eliminate all evaluative discourse. He thinks that
affective and purposive orientations toward the world already include evaluations; indeed, he writes that even
“sense-perceptions are permeated with values” (KSA 12:2[95]). He seems untroubled by the idea that certain
psychological descriptions must be posed in evaluative terms. In short, Nietzsche’s accounts of the natural are
complex. As Janaway puts it,
If Nietzsche’s causal explanations of our moral values are naturalistic, they are so in a sense which includes
within the “natural” not merely the psychophysical constitution of the individual whose values are up for
explanation, but also many complex cultural phenomena and the psychophysical states of past individuals and
projected types of individual. (Janaway 2007, 53)
But, as Leiter has argued, this is consistent with the idea that Nietzsche is a naturalist (Leiter forthcoming a).
Nietzsche wants “to complete our de-deification of nature . . . [and] to ‘naturalize’ humanity in terms of a
rely on “an imaginary natural science (anthropocentric; no trace of any concept of

natural causes)” (A 15; italics removed). But he seems to think the real work is done by
the psychological constraint: the models of agency, consciousness, deliberation, and
knowledge employed by the traditional ethical theories are problematic. Thus, a very
common form of objection in Nietzsche’s works is this: Plato, Kant, Mill, or some
other philosopher has a defective, unrealistic account of agency; recognizing this fact
vitiates the philosopher’s moral theory.33
The psychological constraint seems simple enough: as Nietzsche puts it, all that is
required is that we “translate man back into nature” (BGE 230). The idea that we need
to translate back implies that our current conception of human beings has somehow
gone astray. For example, Nietzsche writes, “We no longer derive man from ‘the spirit’
or ‘the deity’; we have placed him back among the animals . . . Descartes was the first to
have dared, with admirable boldness, to understand the animal as machine. The whole
of our physiology endeavors to prove this claim. And we are consistent enough not to
except man, as Descartes still did . . . ” (A 14). Nietzsche’s idea, then, is that our concept
of human being and human agent must be freed from the accretions of defunct religious
and philosophical interpretations. But these accretions and errors aren’t obvious.
Almost everyone agrees that we should avoid appeal to psychologically unrealistic
accounts of agency; almost everyone disagrees about what counts as psychologically
unrealistic. As a result, uncovering the psychological errors inherent in certain moral
theories is a substantial task—and one that will occupy much of Chapters 3 through 6.
1.3 The practical challenge
So far, we have two challenges for ethical theory. First, there is an epistemological
challenge: an adequate ethical theory must explain why we should have confidence in
our moral beliefs. Second, there is a metaphysical challenge: the theory must be
naturalistically respectable, both in its treatment of normative properties and its analysis
of agency. This brings us to a third challenge: as mentioned above, morality is
prescriptive. It not only tells us what to do, but purports to outweigh many competing
claims about what to do. An adequate ethical theory must explain why and how
morality has this grip on us.
A common way of making this point is by appealing to Motivational Judgment
Internalism (hereafter MJI): if an agent judges that she ought to ç, then insofar as she is
rational she is motivated to ç.34 The idea behind MJI is quite simple: if you are rational,
pure, newly discovered, newly redeemed nature” (GS 256). In the following chapters, I will explore how
these considerations inform Nietzsche’s account of motivation and agency.
33
For a few examples, see BGE 32, GM I.13, GM II.2, and TI VI. We will examine these claims in depth
in the following chapters. For discussions of these claims, see Leiter (2002), Leiter and Sinhababu (2007),
Risse (2007), and Gemes and May (2009).
34
Scanlon has a clear statement of this claim: “If a person judges that she has conclusive reason to do X at t,
then two things follow. First, insofar as she does not abandon or forget this judgment, she is irrational if she
does not intend to do X at t. Second, the fact that she holds this judgment about reasons can explain her
then the answers to the questions “What ought I to do?” and “What will I do?” are the
same. For, as Ralph Wedgwood puts it, “if you are rational, your question ‘What
ought I to do?’ is a deliberative question about what to do” (2007, 25). So, if I am
rational, when I judge that I ought to brush my teeth, I will acquire some motivation to
do so. We can see this as a constraint on the content of normative claims: in order for it
to be true that I ought to ç, the thought of ç-ing must be capable of motivating me.
Although appeals to MJI were at one point quite common in the literature, a
number of objections have emerged. Some philosophers have argued that once we
include the caveat that MJI applies only to rational agents, MJI becomes stipulative or
merely definitional. Others have argued that agents can be amoralists, who make moral
judgments perfectly well but are not motivated to conform to them.35 If this is correct,
then the connection between moral judgment and motivation drawn by MJI may be
too tight.
However, we needn’t resolve these disputes, for there is a second way of putting the
point. Our moral beliefs have a grip on us. Morality tells us what to do. We may not do
what it tells us to do; if amoralists are a real possibility, we may not even be motivated in
the slightest to do what it tells us to do. But if morality could be completely severed
from motivation—if my judgments about what is valuable, what is wrong, what
I ought to do were utterly disconnected from what I actually do—then it is hard to
see what the point of making these judgments would be. Morality serves a purpose
only if it is possible for moral judgments to have some grip on us.
It’s easiest to explain this point with an analogy. Let’s consider a (partially) hypo-
thetical story about the rise and fall of the norms of etiquette. Suppose there were a
group that conformed to a rigid set of rules of etiquette for dinner parties. These rules
proscribed eating one’s salad with the entrée fork, placing one’s glass on the right side
of the place setting, and so forth. Everyone in the community recognizes these rules as
valid. But over time agents begin to bother less about them; conformity to them begins
to drop off, people talk about them less, and they come to play an altogether less
pervasive role in the community’s dinners. When asked, everyone in this community
can cite the relevant rules perfectly well; it’s just that, as we might put it, they care less
about the rules.
Then along comes a philosopher of etiquette, who is terribly concerned to show
these agents that they should return to their earlier concern with these norms. How
would the philosopher motivate these agents to preserve these norms? Well, it
wouldn’t be enough to state that the rules of etiquette are such and such. After all,
these agents already know what the rules are; they just aren’t concerned to conform to
intending to do X at t, and her so acting” (Scanlon 2003, 12). See also Wedgwood (2007, 32), who offers the
following version of motivational judgment internalism: “Necessarily, if one is rational, then, if one judges
‘I ought to ç’, one also intends to ç.”
35
See Brink (1986) and Svavarsdottir (1999 and 2006). For responses to these kinds of objection, see
Wedgwood (2007).
them. So what is the proponent of etiquette to do? The answer seems clear: he needs to
connect the rules to something that these agents care about, to some of their goals or
motives or aspirations. Perhaps they take pleasure in the pomp of formal dinners;
perhaps the norms promote a desired social cohesion; perhaps they aspire to preserve
tradition; and so on. These kinds of considerations might restore etiquette to its former
role.
I suggest that morality is analogous. If agents develop a skeptical attitude toward their
moral code, we need to be able to say something to them. It won’t do simply to insist
that these just are the rules of morality, any more than it would do to insist that these
just are the rules of etiquette; the agents are perfectly well aware of what the rules are.
We need to offer some explanation of why these rules should have a grip on the agents.
Absent such an explanation, it’s hard to see why the rules shouldn’t just wither away, in
much the way that the more recherché rules of etiquette have, to a considerable extent,
withered away. (Notice that I am not claiming that our moral code actually would
wither away. We have many powerful motives to conform to morality: sentiments of
compassion, a desire to do what is in our long-term self-interest, a desire to be an
accepted member of the community, and so forth. I am claiming that absent such a
connection between motives and morality, morality might die out.)
Regardless of whether particular normative judgments are necessarily motivating,
the etiquette story shows that it is possible for entire systems of norms to lose their
connection to motivation. And we want to know whether moral norms—i.e., pur-
portedly universal norms—might meet a similar fate. To avoid that fate, it looks
like morality needs to have some connection to our motives. As Harry Frankfurt
puts it, it looks like “what we should care about depends upon what we do care about”
(2006, 24).
These considerations suggest that moral norms have to be grounded in our mo-
tives.36 But there is a tension: we also want morality to provide a check on our motives.
Let me explain.
Nietzsche phrases a version of this objection in terms that are somewhat unfamiliar
to contemporary ethicists: he calls it the problem of nihilism. I think Nietzsche’s
remarks on nihilism constitute a very powerful challenge for ethical theory, but to
appreciate its strength the point must be put carefully. The groundwork for Nietzsche’s
views on nihilism emerged from earlier nineteenth-century discussions of value, so it
helps to start there. Consider a passage from Hegel’s Philosophy of Right, in which Hegel
considers the view that all value is grounded merely in arbitrary choices of the
individual:
36
Williams (1981) famously argues that an agent has a reason to ç only if A would be motivated to ç if he
deliberated in a procedurally rational way from his existing motives. The idea, here, is that a consideration can
be a reason only if it can motivate me; moreover, a consideration can motivate me only if it bears an
instrumental connection to my existing motives.
This implies that objective goodness is merely something constructed by my conviction,

sustained by me alone, and that I, as lord and master, can make it come and go. As soon as
I relate myself to something objective, it ceases to exist for me, and so I am poised above an
immense void, conjuring up shapes and destroying them. (Hegel 1991, }140A)
Hegel here argues that if the authority of my values arose merely from my arbitrary,
unconstrained acts of will, then these values would not appear as objective constraints.
Rather, they would appear as empty, ephemeral shapes—for the agent could rescind
the value’s authority as easily as she could bestow it. Kierkegaard makes the same point
in The Sickness unto Death. In this work, Kierkegaard mocks the idea that the authority
of values could be grounded in the agent’s own doings. He claims that if “the self exerts
the loosening as well as the binding power”—that is, if the authority of values consists
in the agent’s binding herself to these values, and if the self can loosen anything it binds,
then:
The self is its own master, absolutely its own master . . . On closer examination, however, it is easy
to see that this absolute ruler is a king without a country, actually ruling over nothing; his
position, his sovereignty, is subordinate to the dialectic that rebellion is legitimate at any
moment. Ultimately it is arbitrarily based upon the self itself. Consequently, this despairing self
is forever building only castles in the air . . . just when it seems on the point of having the building
finished, at a whim it can dissolve the whole thing into nothing. (Kierkegaard 1983, 69–70)
Here, Kierkegaard claims that a value whose authority is grounded merely in uncon-
strained choices is no value at all: if “rebellion is legitimate at any moment”—in other
words, if I can reject the value as soon as I feel like doing so—then the value does not
constrain me, and amounts to nothing more than a whim.37
I think we can fairly summarize Hegel and Kierkegaard’s point as follows: in order
for us genuinely to will something, in order for our goals to inspire real allegiance, we
need to see these goals as having more authority than mere whims. We need to see
something as non-arbitrarily structuring and constraining our choices. (What counts as
arbitrary structuring is going to vary: the voice of tradition and authority used to be
enough for us, but, Nietzsche thinks, no longer is.)
Nietzsche’s discussions of nihilism build on this point. He offers the following
definition of nihilism: “What does nihilism mean? That the highest values devalue
themselves. The goal is lacking; ‘why?’ finds no answer” (KSA 12:9[35]/WLN 146). In
other words, the values that were formerly regarded as highest or most central are
experienced as unsupportable. Nietzsche explains that “this realization is a conse-
quence of the cultivation of truthfulness—thus itself a consequence of the faith in
morality” (KSA 12:10[192]/WLN 205). That is, values devalue themselves in the
37
Hegel and Kierkegaard actually have a deeper target: the Kantian account of normativity. They argue
that the Kantian account, according to which normative authority issues from volition, provides no
substantive constraints; to put the point briefly, Kant’s categorical imperative is an “empty formalism.”
Accordingly, the complaint above is directed at Kant: his theory makes it impossible to distinguish norm and
whim. I discuss this issue in more detail in Katsafanas, “The Problem of Normative Authority in Kant, Hegel,
and Nietzsche.”
sense that our moral code prizes truthfulness, and when faithfully pursued the com-
mitment to truth leads us to doubt whether any of our valuations—truthfulness
included—can be justified. (Mere appeals to tradition, for example, are no longer
accepted.) Nihilism is the belief that no values can be justified. So, a nihilist could be
someone who accepts the epistemological and metaphysical challenges discussed
above, and thinks that no ethical theory can answer them.
The consequences of nihilism are far-reaching. Nietzsche describes the nihilist as
holding that “life is no longer worthwhile, all is the same, all is in vain” (Z IV.11). We
might put the point more clearly: because “all is the same”—because no values enjoy
any support—life is no longer worthwhile and all is in vain. Projects, commitments,
and ways of life appear unsupported, arbitrary: any way of life, any choice, any action is
as good as any other.
For this reason, Nietzsche is not interested in the typical bogeymen from contem-
porary ethical theory, the egoist, the amoralist, and their ilk. The egoist can’t see any
reason to conform to morality unless doing so is in his self-interest (where self-interest
is assumed to be something that can be unproblematically specified). The amoralist is
fully cognizant of the accepted moral code, but isn’t moved by it. These characters
express the following point of view: universal normative demands would be so hard, so
constraining; we need to show why we’d be motivated to live under them, instead of
throwing off their yoke and enjoying freedom from them.38
But Nietzsche isn’t worried about whether morality is hard. In fact, he’s interested in
something like the reverse of this position.39 He argues that in order for us to view our
actions, projects, and indeed our lives as meaningful—in order for our goals to inspire
real allegiance, real sacrifice, real direction for the will—we must take certain values as
authoritative.40 From this perspective, a coherent egoist or amoralist wouldn’t experi-
ence relief and freedom; the ‘freedom’ from universal normative demands would
instead bring despair and senselessness. As Nietzsche puts it in a passage that is worth
quoting at length:
38
For an insightful discussion of the problems with this point of view, see Bergmann (1994). Note also
that certain accounts of rational egoism treat the rational egoist as subject to universal normative demands.
See, for example, Sidgwick (1981).
39
Karl Jaspers remarks that “Nietzsche attacks morality in every contemporary form in which he finds it,
not in order to remove men’s chains, but rather to force men, under a heavier burden, to attain a higher rank”
(Jaspers 1997, 140). Nietzsche writes, “Basically I abhor every morality that says: ‘Do not do this! Re-
nounce!’ . . . But I am well disposed towards those moralities that impel me to do something again and again
from morning to evening, and to dream of it at night, and to think of nothing else than doing this well . . . ”
(GS 304).
40
I mean this to be an uncontroversial point about Nietzsche, so let me distinguish two claims: (1) in order
to avoid nihilism, we must treat certain values as authoritative; (2) treating a value as authoritative involves or
requires viewing the valuation as justified. Nietzsche clearly holds (1), as the quotations below indicate.
Whether he holds (2) is more controversial. We might, for example, read Nietzsche as attempting to affirm
certain values without thinking that this affirmation can be justified. In Chapter 6, I will argue that Nietzsche
does indeed hold (2).
What is essential ‘in heaven and on earth’ seems to be, to say it once more, that there should be
obedience over a long period of time and in a single direction: given that, something always
develops, and has developed, for whose sake it is worth while to live on earth; for example,
virtue, art, music, dance, reason, spirituality—something transfiguring, subtle, mad, and divine.
The long unfreedom of the spirit, the mistrustful constraint in the communicability of thoughts,
the discipline thinkers imposed on themselves to think within the directions laid down by a
church or court, or under Aristotelian presuppositions, the long spiritual will to interpret all
events under a Christian schema and to rediscover and justify the Christian god in every
accident—all this, however forced, capricious, hard, gruesome, and anti-rational, has shown
itself to be the means through which the European spirit has been trained to strength, ruthless
curiosity, and subtle mobility, though admittedly in the process an irreplaceable amount of
strength and spirit had to be crushed, stifled, and ruined (for here, as everywhere, ‘nature’
manifests herself as she is, in all her prodigal and indifferent magnificence which is outrageous but
noble) . . . Slavery is, as it seems, both in the cruder and in the more subtle sense, the indispensable
means of spiritual discipline and cultivation, too. Consider any morality with this in mind: what
there is in it of ‘nature’ teaches hatred of the laisser aller, of any all-too-great freedom, and
implants the need for limited horizons and the nearest tasks—teaching the narrowing of our
perspective, and thus in a certain sense stupidity, as a condition of life and growth. You shall
obey—someone and for a long time: else you will perish and lose the last respect for yourself—
this appears to me to be the moral imperative of nature which, to be sure, is neither ‘categorical’
as the old Kant would have it (hence the ‘else’) nor addressed to the individual (what do
individuals matter to her?), but to peoples, races, ages, classes—but above all to the whole
human animal, to man. (BGE 188)41
So Nietzsche is mounting a practical challenge for morality, but not the one expressed
by MJI; Nietzsche does not care whether particular normative judgments are necessar-
ily motivating. What concerns him is whether the whole system of normative judg-
ments might become detached from our practical deliberations. What concerns him is
whether normative judgments might come to seem as nothing more than the expres-
sions of mere whims. To put the point in terms of my etiquette story from above:
Nietzsche is worried that all purportedly universal normative claims might come to
seem vestigial, like the norms of etiquette in our hypothetical community. If so,
Nietzsche thinks, the results would be disastrous: if we become incapable of seeing
any value as authoritative, if all valuations are seen as nothing more than expressions of
contingent whims, then we will lose any ability to sustain our commitment to goals. As
Nietzsche puts it in the quotation cited earlier, nihilism means that “The goal is
lacking; ‘why?’ finds no answer” (KSA 12:9[35]/WLN 146). Goals must “inspire
faith” (KSA 12:9[35]), but the nihilist cannot see any reason for this faith: “ ‘Why did
we ever pursue any way at all? It is all the same.’ Their ears appreciate the preaching,
41
Compare the following remark: “it is almost always a symptom of what is lacking in himself when a
thinker senses in every causal connection and psychological necessity something of constraint, need,
compulsion to obey, pressure, and unfreedom; it is suspicious to have such feelings—the person betrays
himself ” (BGE 21).
ASSESSING ETHICAL THEORIES IN LIGHT OF THESE CHALLENGES 25
‘Nothing is worthwhile! You shall not will!’ ” (Z III.12; cf. GS 1, GS 125, GM

III.28).42
In short: we need something that structures our actions, categorizing certain goals as
more important than others, some as worth pursuing, and so forth. Morality aspires to
be just this. So there’s an odd demand: to grip us, we want to say, morality must have
some connection to our motives, goals, and aspirations. But the connection can’t be too
tight: for we want morality to provide some kind of check on our motives, goals, and
aspirations. Or, put differently: we want morality to be related to what we care about,
but we also want it to provide constraints on what we care about. A successful ethical
theory must answer this challenge, by showing how normative claims can attach to our
motives without collapsing into expressions of mere whim. Call this the practical
challenge.
2. Assessing ethical theories in light of these challenges

So we have three challenges for moral philosophy. An adequate account of morality
must answer them. In the following sections, I will consider how four dominant ethical
theories fare with respect to these challenges. My claims here are not meant to be
decisive; far from it. These issues have been hotly debated over the past decades, and
I do not hope to resolve them in a single chapter. I intend rather to survey some
familiar problems with certain popular ethical theories. I will argue that appreciating
these challenges opens us to the possibility of a new kind of ethical theory, which
would avoid them.
2.1 Non-reductive realism
Let’s begin by considering non-reductive moral realism. Non-reductive moral realism
was defended by Plato (on some interpretations), Ross (1930), Moore (1971), and
Sidgwick (1981), among others. According to this view, there are moral facts and these
facts are irreducible. Over the past decade or so, non-reductive moral realism has
enjoyed a striking resurgence. Parfit defends this view, writing that “there are some
irreducibly normative truths” (Parfit 2011, vol. II, 464). Scanlon argues for another
version: he maintains that a consideration is a reason iff it “counts in favor of ” some
action, and he begins his book by stating that he will take this favoring relationship as
an irreducible primitive (Scanlon 2000, 17).
42
Making a related point, Pippin writes that Nietzsche is pervasively concerned with the conditions under
which our “eros, our orienting commitment” is “sustainable and how it could . . . come to fail” (Pippin 2010,
11). Putting this in terms of note 40, Pippin is interested in the extent to which (1) is possible without (2); he
explores the way in which we might treat values as authoritative without linking this authority to questions of
justification. For a gripping discussion of this problem in a concrete context, see Lear (2006). Lear writes, “a
crucial aspect of psychological health depends on the internalization of vibrant ideals . . . in relation to which
one can strive to live a rewarding life. Without such ideals, it is difficult to see what there is to live for” (2006,
140). He documents the collapse and rebirth of ideals in the Crow culture.
Non-reductive realist views claim that certain normative beliefs are true. But how
do we justify these claims about moral truths? The most common approach taken by
realists is intuitionism, which is the view that some moral truths are knowable a priori.43
For example, Sidgwick claims that certain moral truths are self-evident:
the propositions, “I ought not to prefer a present lesser good to a future greater good,” and “I
ought not to prefer my own lesser good to a future greater good of another,” do present
themselves as self-evident; as much (e.g.) as the mathematical axiom that “if equals be added
to equals the wholes are equal.” (Sidgwick 1981, 383)
W. D. Ross agrees: “in ethics we have certain crystal-clear intuitions from which we
build up all that we can know about . . . the nature of duty” (Ross 1939, 144). For
Ross, these intuitions lead us to see that there are five distinct duties: fidelity, reparation
for previous wrongs, gratitude, promotion of the aggregate good, and non-maleficence
(Ross 1930, 19–25). More recently, Shafer-Landau has argued that we have a priori
knowledge that infliction of pain on innocent children is wrong (cf. Shafer-Landau
2003), Parfit has made similar claims about the badness of suffering (cf. Parfit 2011, vol. II,
76–82), and so on.
How do these views fare with respect to the epistemological, metaphysical, and
practical challenges? I contend that they do not offer convincing responses to any of
these challenges. Let’s start with the epistemological challenge.44 As the above quota-
tions demonstrate, when asked to explain why a particular normative claim is true,
realists appeal to truths that are allegedly known a priori. (In a moment I will consider a
complication: some realists avoid this commitment by appealing to reflective equilib-
rium.) But is this warranted?
On this point, Nietzschean critiques seem to me devastating. For Nietzschean
genealogies should make us exceedingly skeptical of the intuitions and convictions
that are being labeled “a priori knowledge.” Nietzsche writes that “whoever ventures
to answer” philosophical questions “by an appeal to a sort of intuitive perception, like
the person who says, ‘I think, and know that this, at least, is true, actual, and certain’—
will encounter a smile and two question marks from a philosopher nowadays. ‘Sir,’ the
philosopher will perhaps give him to understand, ‘it is improbable that you are not
mistaken; but why insist on the truth?’ ” (BGE 16). His point, here, is that genealogies
43
Old-fashioned moral realists, such as Reid, appealed to a “moral faculty”: he claimed “that by an
original power of the mind, which we call conscience, or the moral faculty, we have the conceptions of right
and wrong in human conduct, of merit and demerit, of duty and moral obligation, and our other moral
conceptions; and that, by the same faculty, we perceive some things in human conduct to be right, and others
to be wrong; that the first principles of morals are the dictates of this faculty; and that we have some reason to
rely upon those dictates, as upon the determinations of our senses, or of our other natural faculties” (Reid
1983, 237). More recently, non-reductive realists have preferred to divorce claims about a priori knowledge
from claims about special faculties.
44
There is a well-known challenge for non-reductive realism at this point: we can wonder how we have
epistemic access to these normative facts (cf. Mackie 1977). I will be pressing a different objection.
should debunk our confidence in our own moral beliefs, be these beliefs about
suffering, equality, or what have you.
After all, our intuitions and convictions about morality are strongly influenced by
the moral code under which we have been raised. Those raised in a wealthy, safe,
democratic society that prizes Judeo-Christian values will have the kinds of intuitions
mentioned above; surely Shafer-Landau is right to say that many of us will claim that
we have a priori knowledge that killing innocent individuals is wrong. But consider
Robert Pippin’s objection to a related view. Discussing Mark Hauser’s claim that “do
as you would be done by; care for children and the weak; don’t kill; avoid adultery and
incest; don’t cheat, steal or lie” are “moral universals,” Pippin writes that this list “really
takes one’s breath away.” He continues,
This banal list of modern, Christian humanist values was written by a Harvard professor in a
contemporary world still plagued by children sold into slavery by parents who take themselves to
be entitled to do so, by the acceptability of burning to death childless wives, by guilt-free spousal
abuse, by the morally required murder of sisters and daughters who have been raped, by
“morally” sanctioned ethnic cleansing undertaken by those who take themselves to be entitled
to do so, and one could go on and on. (Pippin 2009, 41–2)
In short, Pippin’s point is that these alleged moral truths are a product of acculturation,
and consequently are not universally shared. The examples discussed above illustrate
this: someone immersed in the moral code of Homeric Greece would have had strong
intuitions that megalopsychia, envy, social hierarchy, and so forth are good.
If we look at the way in which moral beliefs have shifted over time, if we appreciate
the complex links between these concepts, it is hard to see the lists of allegedly a priori
moral truths drawn up by these theorists as anything more than reports of life from
within a particular moral code. The realists may well chart, systematize, and harmonize
our conventional moral beliefs. But we want more than that. We want, if possible, a
reason to maintain our commitment to this system.
Thus, to someone with Nietzschean sympathies—or, to put the point more polem-
ically, to someone with historical sensitivity—the non-reductive realist simply shows us
what follows from within a particular moral code. Scanlon’s version of non-reductive
realism, for example, shows us what follows if we take for granted the idea that we
should act on principles that no one can reasonably reject; utilitarians show us what
follows if we take for granted the idea that we should maximize aggregate utility. There
is nothing wrong with this; it is a monumentally difficult task. But it does not so much
as touch on the epistemological problem: it gives us no reason for confidence in our
moral beliefs.
There is, however, a complication. So far, I have been objecting to the intuitionist
component of non-reductive realism. This is the claim that we have a priori knowledge
of certain moral truths. However, some realists eschew talk of a priori intuitions and
instead appeal to reflective equilibrium.45 Rather than seeking to ground morality in

intuitions, we might pursue a more modest task: we might try to move from the moral
beliefs that we do hold to the moral beliefs that we should hold. On this view, ethical
theory is a matter of increasing the degree of coherence between our beliefs about
particular cases, general principles, and theoretical beliefs. We start with intuitions
about which particular actions are wrong (harming innocents, lying, cheating, and so
forth), which general principles are valid (happiness is to be maximized, harming is
worse than failing to help, etc.), and which theoretical beliefs are true (utilitarianism is
an adequate moral theory, etc.). We strive to increase the coherence among these
beliefs, eliminating inconsistencies and tensions; ideally, we might even come to see
some of these beliefs as providing warrant for others.
Reflective equilibrium views differ from intuitionist views in that the former need
not take any particular beliefs as having a privileged epistemic status. Any belief or
intuition can come up for review and possible rejection. Unfortunately, though, the
epistemological problem arises in just the same way. Reflective equilibrium views need
to treat some set of intuitions and beliefs as having initial credibility. After all,
depending on the starting points we might end up in quite different places: there are
multiple, mutually inconsistent systems of beliefs that are in reflective equilibrium. Put
differently: reflective equilibrium, if faithfully executed, wouldn’t lead to a unique
moral code. Nietzsche’s ancient nobles would have been in reflective equilibrium with
a set of values that differs entirely from our own. But, given the way in which they
arose, why should we grant these initial intuitions any credibility?46,47
So I take it that non-reductive realism fails the epistemological challenge: regardless
of whether it relies on intuitionism or reflective equilibrium, it gives us no reason for
confidence in our moral code.
Let’s now turn to the metaphysical challenge. On the face of things, non-reductive
realism is again in trouble. Non-reductive realism posits irreducible normative truths,
which seem to be paradigms of properties that those with naturalistic sympathies will
45
In his 2009 Locke Lectures (entitled “Being Realistic About Reasons”), Scanlon argues that we must
employ the method of reflective equilibrium in order to defend claims about reasons for action.
46
Some utilitarians have pressed this line: Singer criticizes Rawls for “assum[ing] that our moral intuitions
are some kind of data from which we can learn what we ought to do” (Singer 2005, 346). Singer objects: “A
normative moral theory is an attempt to answer the question ‘What ought we to do?’ It is perfectly possible to
answer this question by saying: ‘Ignore all our ordinary moral judgments, and do what will produce the best
consequences’ ” (Singer 2005, 345–6). Thus, he writes that “there is no point in trying to find moral
principles that justify the differing intuitions to which various cases give rise” (2005, 348). For, “there is
little point in constructing a moral theory designed to match considered moral judgments that themselves
stem from our evolved responses to the situations in which we and our ancestors lived during the period of
our evolution . . . ” (2005, 348).
47
In Political Liberalism, Rawls claims that reflective equilibrium is a search for reasonable grounds of
reaching agreement that can be based on “our conception of ourselves” and in our “relation to society”
(Rawls 1993). But, Nietzsche would point out, our conceptions of ourselves and our relation to society have
histories, and themselves embody normative claims.
find fantastical. Indeed, contemporary realists often acknowledge the counterintuitive

nature of their proposal. For example, Derek Parfit writes,
Many . . . writers ignore the possibility that there might be normative truths . . . Gibbard regards
this possibility as too fantastic to be worth considering. There are good reasons to have this
attitude. Irreducible normative truths, if there are any, are most unusual. As many writers claim, it
is not obvious how such truths fit into a scientific world-view. They are not empirically testable,
or explicable by natural laws. Nor does there seem to be anything for such truths to be about.
What can the property of badness be? Given these points, it is natural to doubt whether these
alleged truths even make sense. If such truths are not empirical, or about features of the natural
world, how do we ever come to understand them? If words like ‘reason’ and ‘ought’ neither refer
to natural features, nor express our attitudes, what can they possibly mean? (Parfit 2006, 330)
Although Parfit champions a view according to which there are irreducible normative
truths about what there is reason to do, he admits that these normative truths may seem
“too fantastic to be worth considering.” I think this is the correct attitude: we should
appeal to irreducible normative truths only if we are driven to that position by the
failure of other ethical theories.48
Let’s end by considering the practical problem. It is very hard to see why we should
care about these alleged normative truths. If non-reductive realism is true, then moral
facts could be completely disconnected from our motives, goals, and aspirations. But
how could these moral facts be of any relevance to us? They look exactly analogous to
the etiquette facts discussed above: for those on whom they have no grip, it is hard to
see what to say. All the realist can say is “It’s a fact that ç-ing is wrong” or “You should
do what’s in accordance with objective morality.”49 We’d like a view that can do more
than this, by showing why moral requirements are something we should care about.50
Realists do have a response. For example, Parfit argues that this kind of objection
conflates normative authority with motivational force: “many people, I believe mis-
takenly, regard normative force as some kind of motivational force” (Parfit 1997, 126).
48
Some realists have attempted to defuse these metaphysical worries. The “partners in guilt” argument has
been a favorite: realists argue that normative properties are no more mysterious than mathematical properties.
For example, Scanlon claims that moral beliefs, like mathematical beliefs, “do not make claims about things
that exist in space and time and the causal relations between them” (2003, 9). Consequently, he reasons,
moral beliefs are not in conflict with science, which Scanlon understands as an “account of the occurrence of
events in the spatio-temporal world and of the causal relations between them” (2003, 9). Science and
morality are about different things. Parfit offers a similar argument, writing that “there are some claims that
are irreducibly normative . . . and are in the strongest sense true. But these truths have no ontological
implications. For such claims to be true, these reason-involving properties need not exist either as natural
properties in the spatio-temporal world, or in some non-spatio-temporal part of reality” (2011, II, 486).
Other realists try to defuse the metaphysical problem in a different way: they argue that normative properties,
though irreducible, are “realized” by natural properties (see, for example, Wedgwood 2007).
49
Thus, when pressed on these kinds of questions, a realist like Clarke can do nothing more than resort to
an ad hominem: “These things are so notoriously plain and self-evident, that nothing but the extremest
stupidity of mind, corruption of manners, or perverseness of spirit, can possibly make any man entertain the
least doubt of them” (Clarke, Boyle Lectures of 1705, reprinted in Schneewind 2002, 296).
50
For similar complaints, see Korsgaard (1996b) and Gibbard (2003, 152–8).
Parfit’s idea is that normative facts needn’t have any motivational force whatsoever:
“even when [normative claims] do not have motivational force, they could . . . have
normative force” (1997, 111–12). Scanlon takes the same approach. He writes, “Sup-
pose a person believes that he has conclusive reason to do X at t. How can this fall short
of what is required?” (Scanlon 2003, 14). Scanlon grants that the person might not be
moved to X. But he does not see this as an objection to the view. I do not find these
responses convincing: they amount to a kind of mysterianism about normative author-
ity. The response consists merely in the assertion that there is a sui generis property,
normative authority, that can obtain independently of any facts about motivation. But
this is precisely what the practical argument calls into question: in light of the historical
considerations adduced above, the mere insistence that there is a sui generis property of
normative force seems, as Nietzsche might put it, rather quaint.
2.2 Aristotelian theories
I have suggested that non-reductive realism has trouble mustering convincing re-
sponses to the three challenges. So let’s turn to a different kind of ethical theory:
Aristotelianism. Aristotelians argue that we can derive norms from facts about the
natures of things. It’s easiest to see how by considering objects that have functions—
motors, toasters, knives, hearts, lungs. For any type with a function, we can evaluate
particular tokens of that type with respect to whether they have the properties required
to fulfill the function. A good knife is one that has the properties necessary for cutting; a
bad knife is one that lacks some or all of these properties. The same goes for parts of
living creatures: a heart is defective if it lacks the properties required for circulating
blood in the requisite way. The central Aristotelian idea is that we can extend this kind
of evaluation to living things, including human beings.
Rosalind Hursthouse gives a nice summary of this approach:
Living things can be . . . evaluated according to all sorts of criteria. We may evaluate them as
potential food, as entries in competitive shows, evens as ‘decorative objects for my windowsill
given my preferences,’ and each noun or noun phrase brings its own criteria of goodness with it.
In the context of naturalism [i.e., Aristotelianism] we focus on evaluations of individual living
things as or qua specimens of their natural kind, as some well-informed gardeners do with respect
to plants and ethologists do with respect to animals. (Hursthouse 1999, 197)
In short, these theorists offer a characterization of “natural kinds,” and then evaluate
particular individuals by determining whether “this individual x is a good x, a good
specimen of its kind” (Hursthouse 1999, 203). Thus, this view “hopes to validate”
ethical claims “by appeal to human nature” (Hursthouse 1999, 193).
There are variations in the details of these views. Hursthouse argues that we can
evaluate aspects of an individual with respect to how well they contribute to four
criteria: the individual’s survival, the continuation of the species, the individual’s
pleasure and freedom from pain, and (with social animals) the functioning of the
group (Hursthouse 1999, 200–3). Bloomfield (2001) contends that moral goodness is
the state of character that disposes human beings to flourish, where flourishing is
defined in terms of biologically determined human purposes. As he puts it, “moral
properties have the same ontological status as healthiness or other biological properties”
(Bloomfield 2001, 28). Thomson (2008) relies on the idea that there are “goodness-
fixing kinds”: kinds such that being a member of that kind establishes standards of
excellence for its members.
How do these views fare with respect to the epistemological, metaphysical, and
practical challenges? They have a good response to the epistemological challenge:
unlike the non-reductive realist, the Aristotelian need not place any great faith in
intuitions, conventional beliefs about morality, and so forth. She can simply appeal to
natural kinds or biologically defined functions. If these functions or kinds can be
specified in a way that does not presuppose the truth of particular moral claims, then
the Aristotelian has a way of stepping outside of and assessing her current moral beliefs.
Provided the critical reflection supports the moral code, the epistemological challenge
will have been answered.51
This brings us to the metaphysical challenge: is the Aristotelian view consistent with
naturalistic strictures? This depends on the particular version of the Aristotelian theory
that we embrace. A burden for the Aristotelian is to show why we should believe that
human being is a normative kind. Aristotle himself bases his argument on claims about
natural teleology that have been discredited by modern science (cf. Williams 1986).
But we can develop a naturalistically respectable version of Aristotelianism. For
example, we could appeal to the tendencies or functions instilled by biology or by
natural selection (cf. Bloomfield 2001). Or we could simply note that there’s nothing
inherently problematic about specifying what it is for a tomato plant or a tiger to
flourish; by extension, the same point should apply to human beings.52 So the
Aristotelian may be able to answer the metaphysical challenge.
The practical challenge is more difficult. Suppose I accept that human beings have a
function, or that “human being” is a normative kind. Why should this matter to me?
Why should I care whether I am a defective instance of my kind? After all, it seems
obvious that I regularly neglect features that are characteristic of the human kind. It is
probable that natural selection has instilled in me a disposition to reproduce as often as
possible, to be distrustful of and somewhat hostile toward those who are not members
of my immediate group, to eat as much sugary food as possible, and so on. Why should
I try to realize these aims? Many of the conditions that made these dispositions
beneficial in the evolutionary past are no longer present. (A disposition to eat as
51
Some Aristotelian views start with a moralized notion of flourishing, and define the human function in
terms of it. Annas (1995) and McDowell (2001, Chapter 1) adopt this strategy. These views do not have the
epistemic advantage mentioned above.
52
As Hursthouse puts it, “we evaluate ourselves as a natural kind, a species which is part of the natural
biological order of things, not as creatures with an immortal soul or ‘beings’ who are persons or rational
agents” (1999, 226). Foot writes “I am therefore . . . likening the basis of moral evaluation to that of the
evaluation of behavior in animals” (Foot 2003, 16).
much sugary food as possible made sense when that aim could be realized only by
eating fruits; it does not make sense when I can realize it by eating cookies.) Even when
the aim hasn’t had its conditions of realization altered, I can decline to fulfill it. To be
sure, reproduction is fitness enhancing; but what’s that to me?53 Absent a convincing
answer to this question, the Aristotelian views fail to answer the practical challenge.54
2.3 Humean theories
I have suggested that non-reductive realism has trouble with the epistemological,
metaphysical, and practical challenges. Aristotelianism has the potential to avoid the
epistemological and metaphysical challenges, but lacks a convincing response to the
practical challenge. In light of this, let’s consider a third view: Humeanism. Taking off
from Hume’s claim that nothing is “in itself valuable or despicable,” apart from the
“particular constitution and fabric of human sentiment and affection” (Hume 1987,
162), Humean views claim that normative facts must be explained by some conative
state of the agent:
Agent A has reason to ç iff A has a conative state of type T that is suitably connected to ç-ing.
This is just a schema; it needs to be filled in by specifying what kinds of conative states
are at issue and what counts as a suitable connection. Different specifications of T and
“suitable connection” will generate different versions of Humeanism.
The simplest version of Humeanism claims that all actual desires provide reasons:
A has a reason to ç iff A has a desire whose fulfillment would be promoted by ç-ing.
This version is probably too simple: for example, it entails that desires based on false
beliefs generate reasons.55 So Humeans typically adopt a modified view. For example,
Bernard Williams argues that A has reason to ç iff A has a conative state that is
connected by a “sound deliberative route” to ç-ing. This “sound deliberative route”
can include the correction of false beliefs, the appreciation of instrumental connections
between one’s desires, and so forth. On this view, we are no longer forced to say that
53
As Copp and Sobel put it, “Why should the constituents of natural goodness for members of my species
(or ‘life form’) determine what counts as morally good for me?” (Copp and Sobel 2004, 542). Indeed, the
idea that something ought to meet the standards of its normative kind seems most plausible when there is a
clear end that has been adopted. That is, we will agree that this object is a defective toaster because we want
this object to serve a particular end: making toasted bread. But if an agent didn’t have that end—if I merely
had the end of collecting shiny metal objects—then the fact that this toaster doesn’t heat up bread would
hardly seem to count as a defect. Put differently, we might argue that talk of functional kinds or normative
kinds is really best explained as talk of items that are presumed to serve a certain end. Absent commitment to that
end, it is hard to see why we should care about them.
54
Proponents of this view typically respond in just the same way as non-reductive realists: they insist that
normative authority need not translate into motivational force. See Foot (2001) and Thomson (2008). This
inherits exactly the same difficulties as the non-reductive realists’ response, discussed at the end of Section 2.1.
55
Take Bernard Williams’ classic example (Williams 1981, 101–13): I desire to drink that cup of liquid,
which I believe is full of gin. However, the cup has actually been filled with gasoline. If I knew that the cup
was full of gasoline, I wouldn’t desire to drink it. In this case, it hardly seems that I have a reason to drink the
liquid. In general, desires that are based upon false beliefs do not seem to provide reasons.
desires based on false beliefs generate reasons.56 And once we set off on this track—
moving from facts about the agent’s actual desire to facts about what the agent would
desire in suitably different circumstances—we might be tempted to go even further.
For example, Michael Smith argues that moral facts are facts about what a perfectly
rational and fully informed version of my actual self would desire that my actual self do
(Smith 1994).57
Let’s consider how Humean views fare with respect to the three challenges. The
metaphysical challenge presents no difficulties: Humeans need not appeal to queer
normative properties, but only to desires and other motivational states. There is
nothing queer or non-natural about the idea that agents have various desires, affects,
and motivational states. Moreover, Humeans have a straightforward response to the
practical challenge: if we analyze normative claims in terms of the agent’s motivational
states, it’s easy to see why agents would be motivated by normative claims. (Of course,
the connection between motivation and reasons will become more tenuous to the
extent that—like Michael Smith—we analyze reasons in terms of merely hypothetical
or idealized motives.)
However, things go less smoothly when we consider the epistemological point.
Some Humeans present their theories as immune to epistemological worries. For
example, Hume himself remarks that although skeptical arguments “may flourish
and triumph in the schools,” in ordinary life “they vanish like smoke, and leave the
most determined skeptic in the same condition as other mortals” (Enquiry, 159). The
skeptic, he tells us, “cannot expect, that his philosophy will have any constant influence
on the mind . . . Nature is always too strong for principle” (Enquiry, 160).58 But there
are reasons to doubt that the Humean theories are secure in light of the Nietzschean
epistemological point. We want to know whether we have a reason for confidence in
our moral code. Above, we saw that intuitions and moral beliefs provide poor grounds
for confidence. An analogous problem arises for conative states. Nietzsche writes:
Feelings and their origin in judgments.—‘Trust your feelings!’—But feelings are nothing final or
original; behind feelings there stand judgments and evaluations which we inherit in the form of
feelings (inclinations, aversions). The inspiration born of a feeling is the grandchild of a
judgment—and often a false judgment!—and in any event not a child of your own! To trust
one’s feelings—means to give more obedience to one’s grandfather and grandmother and their
grandparents than to the gods which are in us: our reason and our experience. (D 35)
56
Williams (1981, Chapter 8).
57
For a sophisticated defense of a Humean view, see Schroeder (2007).
58
Charles Griswold notes that analogous points apply to Adam Smith: “Smith does not hold that as moral
actors we normally treat morality as a skeptic would. Rather, we act as though commonsense moral realism
were valid, that is, as though moral qualities exist objectively in the nature of things, are external to us and
claim authority over us” (Griswold 1999, 165). As Griswold puts it, for Smith and Hume “skeptical
theorizing ought not to budge our everyday beliefs” (1999, 165).
Here, Nietzsche claims that feelings often originate in judgments.59 Let’s illustrate this
with an example. Marrying one’s first cousin was quite common in the ancient world,
and is still widely practiced in certain parts of the Middle East and Sub-Saharan Africa.
However, most individuals in the United States and Europe view cousin-marriage as
disgusting or even repellent. We tend to concoct justifications for this emotion. For
example, we tell ourselves that cousins who marry are more likely to have children
with birth defects. However, this is demonstrably false; cousins are no more likely to
have genetically defective children than non-cousins.60 I suspect that even upon
appreciating the falsity of this belief, most individuals in Western societies will continue
to view cousin-marriage as disturbing or even disgusting. This provides an example of
the way in which an evaluative belief—that marrying one’s cousin is wrong—gradually
generates a variety of affects (disgust, revulsion, etc.), which are resistant to transform-
ation, persisting even in the absence of evidence for the belief.61
If our conative states are influenced by our evaluative judgments, then we cannot
treat them as providing any grounds for confidence in these evaluative judgments.
Consider the dialectical situation: we want to know why we should keep our promises,
be compassionate, value equality, and so forth. The Humean says: abide by these rules
because doing so accords with your conative states. And the Nietzschean responds: it’s
true that our conative states and avowed normative claims will tend to be in general
conformity with one another, but this is simply due to the fact that they are reciprocally
influencing. If we lack confidence in the normative claims, we should also lack
confidence in the conative states.62
In this respect, the Humean and non-reductive realist projects can be seen as mirror
images of one another: the Humean project of constructing morality through an
examination of the conative states won’t inspire confidence in our normative claims,
given that we can modify the conative states, and a non-reductive realist project of
59
The same point is made in the following passage: “whence come evaluations? Is their basis a firm norm,
‘pleasant’ or ‘painful’? But in countless cases we first make a thing painful by investing it with an evaluation”
(KSA 10:24[15]).
60
For a helpful discussion of this case, see Prinz (2007, 240). The genetic data are given by Bennett et al.
(2002).
61
The reader is invited to try this experiment in a class: ask students whether marrying one’s first cousin is
wrong. Students almost inevitably say that it is. When asked why cousin-marriage is wrong, students typically
respond by citing the alleged potential for genetic defects. When told that this belief is false, students tend not
to revise their moral judgment. Instead, they resort to saying that cousin-marriage is revolting or disturbing.
Here we have exactly the process that Nietzsche describes: a moral evaluation—based on any superstition,
custom, or false belief—generates a strong affect; the affect is then taken to justify the moral evaluation that
caused it. For a classic discussion of this phenomenon, see Haidt (2001).
62
In fact, things are even worse. Nietzsche posits an additional explanatory factor: he maintains that it is
not only judgments which influence our conative states, but a drive to imitate our fellows. He writes, “It is
clear that moral feelings are transmitted in this way: children observe in adults inclinations for and aversions to
certain actions and, as born apes, imitate these inclinations and aversions; in later life they find themselves full
of these acquired and well-exercised affects and consider it only decent to try to account for and justify
them . . . ” (D 34). Nietzsche calls this the “herd instinct”: he believes that individuals have a strong drive
toward conformity and imitation. For an illuminating discussion, see Richardson (2004, 81–95). If Nietzsche
is correct about the herd instinct, we have even less ground for confidence in our conative states.
justifying morality through intuition of irreducible normative truths won’t inspire

confidence, given that we can modify our intuitions.
Indeed, Nietzsche speculates that appreciation of this fact will gradually transform
our normative judgments and conative states. Nietzsche writes:
We have to learn to think differently—in order at last, perhaps very late on, to attain even more: to
feel differently. (D 103)
Just as our moral beliefs and intuitions are nothing “final and original,” nothing worthy
of trust or confidence, so too our conative states must be called into question. The
Nietzschean critique reveals that our beliefs and conative states are deeply intertwined
and subject to the vicissitudes of history. So neither beliefs nor conative states provide a
stopping-point for questions about justification. Showing that some evaluation is
consistent with our conative states does not show that we have any reason to embrace
the valuation; showing that some conative state is consistent with our evaluations does
not show that we have any reason to embrace it.
Certain Humeans will accept this argument and claim that it only shows that there
are no universal reasons. This is one possible response. It amounts to a denial that we
have any grounds for confidence in our current moral system. Perhaps this is the best
we can do. If so, we won’t have any answer to the epistemological challenge.
Other Humeans do try to meet this challenge. For another way to avoid the worry is
to show that there are some universal reasons. For example, on Michael Smith’s view
normative claims are contingent upon the motives of particular agents; however, he
argues that universal moral claims could be true if all rational and fully informed persons
would converge on a common set of desires (Smith 1994, 187–9). Below, I will argue
that something like this approach can work. First, though, I want to consider one last
approach to ethics.
2.4 Kantian theories
Let’s end by considering Kantianism. Kant attempts to anchor universal normative
claims in facts about agency. An outline of the Kantian argument would go something
like this: we are committed to acting autonomously. Acting autonomously requires
acting on a law or principle. The law cannot be hypothetical, i.e., tied to the realization
of some goal or the satisfaction of some inclination, because the will would then be
determined to action by something external to itself (i.e., an inclination or goal).
Instead, the law must be categorical; it must be unconditionally valid. Kant states the
content of this law as follows: “act only in accordance with that maxim through which
you can at the same time will that it become a universal law” (G 4:421). He argues that
this law—the Categorical Imperative—rules out certain actions, thereby yielding
determinate constraints on permissible actions. So, Kant moves from a claim about
agency—that we are autonomous—to a normative claim about what we have reason
to do (i.e., act on laws that are in accordance with the Categorical Imperative).
This argument is notoriously difficult. Chapter 4 will examine a version of it in more

detail. For now, the details of the argument won’t matter: we need only consider the
basic structure of the argument in order to see how it fares with respect to the three
challenges.
If Kantianism worked, it would avoid the epistemological problem: we would start
with facts about our agential nature and show that universal normative claims issue from
them. As these norms follow from facts about the nature of rational agency as such, they
will apply to all agents, regardless of the particular moral code that these agents currently
embrace. Kantianism would thus give us a way of stepping outside of and assessing our
current moral beliefs. Likewise, Kantianism has no difficulty with the practical challenge:
normative claims get their grip on us because we are committed to acting autonomously,
and these norms simply specify what we have to do in order to be autonomous agents.
Finally, the Kantian theory might avoid the metaphysical problem as well: if we can
ground normative claims in a naturalistically respectable account of agency, then there
will be no need to appeal to queer metaphysical properties.
I have critiqued non-reductive realism, Aristotelianism, and Humeanism for facing
certain structural problems that render them at best unlikely to answer the three
challenges. Kantianism avoids this problem: it has the right structure to answer the
various challenges. Should we then be Kantians?
I think not. For the Kantian theory faces a number of internal problems. (The
following chapters discuss some of these problems in detail; for now, I simply mention
them.) These center on its analysis of agency and its attempt to extract normative
content from this analysis.
Start with the first point. Although some Kantians present their accounts of agency
as naturalistically respectable, it is far from clear that all is well on this score.63 Consider,
for example, Kant’s reliance on the idea that reason can “of itself, independently of
anything empirical, determine the will” (Critique of Practical Reason 5:42); or that all
actions are done on maxims (Groundwork 401n, 421n). Both of these claims look
implausible in light of recent empirical psychology.64 Additionally, some argue that
the Kantian theory actually does require substantive and implausible metaphysical
commitments, such as a claim that we have a kind of freedom consisting in “independ-
ence from everything empirical and so from nature generally” (Critique of Practical
Reason 5:29, 97).65 In short: problems may arise in the Kantian’s reliance on a highly
questionable model of agency.
Second, even brushing those problems aside, the Kantian arguments that attempt to
move from the nature of agency to normative conclusions face objections at each turn:
it is notoriously difficult to show how commitment to the Categorical Imperative is
63
Korsgaard (2010) is a good example. For arguments that aspects of the Kantian enterprise can be
rendered naturalistically respectable, see Scheffler (1992) and Velleman (2006).
64
I will discuss these problems in Chapter 5.
65
See Wood (1984, 74–83), Allison (1990, 227–9), and Schneewind (1998, Chapters 22–23).
CONSTITUTIVISM 37
supposed to follow from Kant’s initial conception of agency, and even if we can do
that, there are reasons for doubting that the Categorical Imperative can generate any
substantive conclusions about what there is reason to do.66 Thus, although Kantian
ethics has the right structure to avoid the three challenges, there are reasons for doubting
the cogency of the theory.
3. Constitutivism
Above, we have examined four chief competitors in ethical theory: non-reductive
realism, Aristotelianism, Humeanism, and Kantianism. I briefly raised some problems
for each of these theories. (Again, I do not intend the brief discussions above to
constitute decisive refutations of the various theories. I intend them merely to indicate
some potential difficulties that the theories would have to overcome.) I suggested that
non-reductive realism runs headlong into the epistemological, metaphysical, and
practical problems. Aristotelian views avoid the epistemological problem, but may
encounter metaphysical difficulties and lack a convincing response to the practical
problem. Humeanism avoids the metaphysical and practical problems but faces the
epistemological challenge. Kantian theories avoid the epistemological and practical
problems, but may be premised upon metaphysically untenable claims about agency.
In sum:
Epistemological Metaphysical Practical
Non-reductive realism X X X
Aristotelianism ? X
Humeanism X
Kantianism ?
I think a version of this dialectic drives some philosophers toward Kantianism. After
all, Kantianism faces enormous internal difficulties; it can hardly be denied that the
arguments are obscure, the challenges severe. But if you think that all other ethical
theories fail—if you think that in some cases they do not even aspire to answer the
genuine puzzles—then it makes sense to embrace a problematic theory. For Kantianism
at least has the virtue of confronting and attempting to answer the three challenges
above. It doesn’t shirk from these challenges and doesn’t reduce its ambitions. If its
success looks dubious, it at least sets off on the right track. So, at any rate, it seems to me.
66
Hegel is the locus classicus for this objection; see his Philosophy of Right Section 135. In the Phenomenology,
Hegel puts the point this way: “It would be strange, too, if tautology, the principle of contradiction, which is
admitted to be only a formal principle for the cognition of theoretical truth, i.e., something which is quite
indifferent to truth and falsehood, were supposed to be more than this for the cognition of practical truth”
(Phenomenology, Section 431). Wood (1990) offers a helpful discussion of the formalism objection.
Should we then cast our lot with the Kantians? I think not. For lately a distinct kind
of ethical theory has emerged. This theory, which is often called constitutivism, offers a
fresh start. It has the advantages of Kantianism without the problems. Below,
I introduce the theory and indicate how it has the potential to overcome the three
challenges. I then ask how closely related constitutivism and Kantianism are.
3.1 Introducing constitutivism
To see how constitutivism works, we need to reorient ourselves. We have been
examining universal normative claims that apply to action as such. But let’s approach
our topic from a different angle: let’s consider norms pertaining to more restricted
kinds of activities.
Certain kinds of activities are distinguished by the fact that participants in these
activities necessarily have a particular aim. There are simple examples of this phenom-
enon, such as the game of chess. Arguably, it is not sufficient to count as playing chess
that one simply moves one’s chess pieces around on the board in accordance with the
rules of chess. In addition, one must aim at achieving checkmate.67 If you do not have
that aim—if you are just moving pieces, without aiming to win—then you are not
really playing chess. Thus, the aim of checkmate is non-optional for chess players: if
you are playing chess, then you have the aim.
Of course, the aim of checkmate can be influenced and modified by other factors.
But it cannot be wholly abandoned. Consider an example. If you are playing chess with
a child who is just learning the game, you may also adopt the aim of letting her have a
fair chance at winning. This aim will modify the way in which you pursue the aim of
achieving checkmate. For example, you may see a way to achieve checkmate, but
decline to take it, in order to give the child a better chance of winning. But this kind of
deviation from the activity’s aim can only go so far, lest you cease to engage in the
activity of playing chess. If you are not making any effort to achieve checkmate, then
you are not really playing chess. Instead, you are engaged in a more complex activity,
with a different aim: you are engaged in the activity of teaching a child how to play chess, or
some such. (Notice that if you are not pursuing the aim of checkmate at all, the child
could justifiably complain that you are not really playing chess.)
Similarly, some philosophers have argued that the attitude of belief aims at truth.68
For it seems that each instance of belief aims at truth, and aiming at truth is part of what
constitutes an attitude as a case of belief. After all, if an attitude had absolutely no
tendency to be responsive to indications of its truth value—if, for example, an attitude
with the content that p persisted despite the agent’s appreciation of conclusive evidence
that not p—then the attitude would not be a belief.
67
I am simplifying a bit: one could also aim at achieving a draw.
68
For two examples, see Shah (2003) and Shah and Velleman (2005). For more skeptical discussions, see
Wedgwood (2002) and Owens (2003).
CONSTITUTIVISM 39
Let’s be more precise. We can define constitutive aim as follows:
(Constitutive Aim) Let A be a type of attitude or event. Let G be a goal.

A constitutively aims at G iff
(i) each token of A aims at G, and
(ii) aiming at G is part of what constitutes an attitude or event as a token of A.69
For example, suppose we let A be the attitude of belief and G be truth. Then belief has
a constitutive aim of truth iff (i) each token of belief aims at truth, and (ii) aiming at
truth is part of what constitutes an attitude as a belief.
We now have an account of constitutive aims. But what would follow from the fact
that chess, belief, or some other type of attitude or event has a constitutive aim? Well,
suppose we accept a relatively uncontroversial claim:
(Success) If X aims at G, then G is a standard of success for X.

For example, if chess players aim at checkmate, then we can evaluate chess players with
regard to whether their actions are conducive to their goal of achieving checkmate. Or,
if belief aims at truth, then we can evaluate processes of belief formation in terms of
how well they promote the goal of believing truths.
Note that Success simply claims that aims generate standards of success. It applies to
all aims, not just constitutive aims. Whenever you have an aim, you have a standard of
success.70 Take our aforementioned chess player. Suppose she has the aim not only of
checkmating her opponent, but also of enjoying her game. Then we get two standards
of success: we can evaluate a particular move with regard to whether the move brings
her closer to checkmate, and whether it makes the game enjoyable. These aims can
interact with and modify one another: if move A would promote checkmate yet would
be boring, while move B would be fascinating yet somewhat more risky, then the
player may have reason to make move B. Thus, the reasons induced by the constitutive
aim will be one source of reasons among many others.
So what’s special about constitutive aims? The reasons derived from the constitutive
aim differ from these other reasons in that they are intrinsic to the activity in question.
You can play a chess game without aiming to enjoy it, and a chess game is not
necessarily defective if not enjoyed. But you can’t play a chess game without aiming
to achieve checkmate, so a move in a chess game is necessarily defective if it does not
69
Here it is worth making two points about the definition. First, condition (i) is implied by condition (ii).
Strictly speaking, then, condition (i) is superfluous. I include (i) as a separate condition merely for the sake of
clarity. Second, nothing important hinges on my restriction of A to attitudes and events; I would be happy to
include other categories that might have constitutive aims. I cite attitudes and events simply because these are
the categories that have been thought to possess constitutive aims. (I am including actions under the broader
category of events.)
70
In the next chapter, I will consider some objections to this claim. I will also explicate the difference
between aims and desires. In Chapter 7, I show that the version of constitutivism that I defend relies only on
the following principle, which is even less controversial than Success: if an agent aims at G, and the agent
endorses this aim, then G is a standard of success for the agent’s action.
promote the goal of checkmate. Thus, the interesting feature of constitutive aims is
that, being inescapable, they generate intrinsic standards of success.
As a result, these standards readily meet challenges to their authority. If someone is
engaged in an activity that these standards govern, then there is a ready answer to the
question “why should I care about these standards?” The answer is just this: insofar as
you are committed to this activity, you are committed to those standards. For example,
a person who is playing chess has a good reason to abide by the standards constitutive of
chess: if he doesn’t abide by them, he will no longer be playing chess.
To see why this is important, it helps to contrast constitutive standards with other
types of standards. Consider familiar rules such as “No smoking in this restaurant” or
“Provide 24-hour notice of any changes to your doctor’s appointment.” These rules
govern activities such as dining and making appointments. But one does not need to
obey these rules in order to participate in the activities: I can light up in the restaurant,
and I can cancel my appointment an hour in advance. To be sure, I may face penalties
for failing to respect these rules. However, I do not cease to participate in the activity of
eating at a restaurant simply because I light a cigarette; nor do I cease to engage in the
activity of making a doctor’s appointment simply because I change the appointment an
hour in advance. Chess is different: if I do not govern myself with chess’s constitutive
standards (by trying to capture pieces, move bishops on diagonals, and so on), then
I will not be playing chess at all.71
Suppose someone asks, “Why should I care about providing 24-hour notice when
changing my doctor’s appointments?” Of course, there are answers to the question—
answers invoking the financial penalties that the canceled appointment will produce,
the inconvenience to the doctor and her other patients, and so on. But notice that these
answers invoke external standards. The standards apply because medical appointments
are related to other activities, goals, and practices that concern the agent. The standards
governing chess do not have that feature: we can answer the question “Why should
I care about capturing your king?” simply by referring to the rules constitutive of the
game. Thus, the chess player should care about capturing the king because if he doesn’t
govern himself by this standard, he won’t be playing chess.
So this is the intriguing feature of the standards induced by constitutive aims: they
are internal to the activity in question. Accordingly, we need not invoke external facts
in order to legitimate their claim to authority.72
71
One question that arises concerning constitutive rules and games is whether games have any non-
constitutive rules. To answer this question, we would need to determine whether we could eliminate rules
without thereby changing the game. For example, if we eliminated the rule stating that pawns can be moved
two spaces on the first move, would participants in the resulting activity still be playing chess? I will return to
this question in the next chapter.
72
External facts will be relevant, of course. If I am engaged in a game of chess, and suddenly notice that
my house is burning down around me, then there’s a very real sense in which my reasons for capturing my
opponent’s queen are outweighed or silenced by my reason to stop playing and call the fire department. I will
return to this point in Chapter 2.
CONSTITUTIVISM 41
Let’s now take a step back. We have been considering particular types of action, such
as chess-playing. But suppose we could show that action itself has a constitutive aim. If
every agent shares a common aim, then every agent shares a common standard of
success. The reasons generated by this standard will be universal: they will apply to all
agents, regardless of facts about the agents’ contingent desires and circumstances. Just as
all chess players have a reason to checkmate, so too all agents will have a reason to fulfill
the constitutive aim of action. Accordingly, the reasons generated by action’s consti-
tutive aim would have the right form to be moral reasons; they would be universal.73
So we have a two-step recipe for a new moral theory. First, we need to show that
action has a constitutive aim. Second, we appeal to some version of Success in order to
derive reasons from this constitutive aim. This would anchor universal reasons in facts
about aims that are constitutive of agency.
Easier said than done, of course. Showing that action has a constitutive aim is going
to be enormously difficult, and defending a version of Success will raise puzzles of its
own. But let’s set these worries aside, for a moment, and ask whether it’s even worth
embarking on these tasks. Would a successful version of constitutivism answer the three
challenges?
3.2 Constitutivism and the three challenges
We can start with the epistemological challenge: would constitutivism give us a reason
for confidence in our normative beliefs? It would. If action had a constitutive aim, then
this aim would generate normative claims with a universal status. Cultural and histor-
ical variation in moral codes would not be troubling, for we would have a standard
against which we could measure and critique these variations. In short, we could say
that some variations are mistakes.
Moreover, notice that Success is an exceedingly spare claim. It can serve as a kind of
Archimedean point in debates about ethics: we disagree about whether we have reason
to be compassionate, whether happiness is more important than duty, whether suicide
is wrong, and so forth. But we can set aside this disagreement on substantive ends and
agree on this entirely procedural or structural conception of rationality: we can agree
that if you have an end, you should strive to fulfill it, while disagreeing about what
those ends are.74
Next, consider the practical challenge. Again, this is easily overcome: the con-
stitutivist has no more trouble with this than does the Humean or Kantian. Norms
will issue from our aims, and hence will be things that we are motivated to meet.
There is no puzzle about why chess players are motivated to achieve checkmate; just
73
It is, of course, a further question whether the universal reasons generated by the constitutive aim will
be the ones we expect. Pre-theoretically, we expect the universal reasons to include claims such as “you
should not murder” and “harming innocent people for fun is wrong.” Whether action’s constitutive aim
entails these particular normative claims will depend upon what, exactly, action’s constitutive aim is.
74
I will consider objections to Success in the next chapter.
so, there would be no puzzle about why agents are motivated to achieve action’s
constitutive aim.
What about the metaphysical challenge? Here things are a bit more complex.
Whether constitutivism meets this challenge depends on the particular version of
constitutivism that we embrace. In order to provide an account of reasons that is
compatible with naturalistic strictures, the constitutivist account must be grounded in a
naturalistically acceptable account of agency. Suppose Nietzsche and other philoso-
phers are correct in claiming that Kant’s theory of agency is indefensible; then it would
do no good to show that we can extract a constitutive aim from the Kantian concep-
tion of agency, precisely because nothing in the world answers to that conception.
What we need to do, instead, is start with an accurate description of actual human
agency.
For this reason, we should be wary of trying to demonstrate the presence of this aim
via conceptual analysis. As I will explain in the next chapters, the dominant versions of
constitutivism often appear to start with claims about our concept of agency, and to show
that we can extract a constitutive aim from this concept. For example, Korsgaard writes
that “it is essential to the concept of agency that an agent be unified,” and attempts to
derive normative conclusions from this alleged fact (2009, 18; emphasis added).75 If
constitutivism relies on an uncritical faith in our current conception of agency, then it
won’t answer the epistemic challenge. Our concept of agency is something that itself
has a history: like our beliefs about morality, our intuitions about agency have
undergone substantial changes over time.76 Thus, Nietzsche criticizes thinkers who
“accept concepts as a gift . . . as if they were a wonderful dowry from some sort of
wonderland,” rather than recognizing that they are “the inheritance from our most
remote, most foolish, as well as most intelligent ancestors,” and therefore stand in need
of “an absolute skepticism” (KSA 11:34[195]/WLN 13). Put simply, our current
concept of agency cannot be taken for granted and used as a starting point. Intuitions
about agency have the same status as intuitions about morality: unless they can be
independently supported, they give us no reason for confidence.
For these reasons, an adequate version of constitutivism must defend the account of
agency upon which it relies. I believe that the most promising way of doing so is by
relying on an empirical account of agency, rather than attempting to divine the
structure of agency in an a priori or conceptual manner. I will pursue that strategy in
the following chapters: I will argue that a roughly Nietzschean account of agency is not
only empirically convincing, but also allows us to see that action has a constitutive aim.
75
In the previous sentence, I said that constitutivist theories often appear to be starting with claims about
the concept of agency. When we examine these theories in more detail, in Chapters 3 and 4, we will see that
this appearance is misleading. For example, while Velleman does begin with some claims about what is
essential to our concept of agency, he is concerned to this concept is “realized in the world”—that is, whether
our concept of agency matches the reality (2000, 129).
76
For helpful analyses of these changes, see for example Taylor (1992) and Williams (1993).
CONSTITUTIVISM 43
3.3 Constitutivism’s relationship to Humeanism and Kantianism

So we can see that constitutivism has considerable potential: a successful version of the
theory would overcome the three challenges to morality. In the next chapter I will ask
whether constitutivism can be defended against a series of recent objections. Before
proceeding to that, though, let’s ask how constitutivism relates to the ethical theories
surveyed above.
There is a clear parallel between constitutivism and Kantianism: both theories
attempt to ground normative claims in facts about agency. Perhaps for this reason,
constitutivism is often thought to be a Kantian theory. This is a mistake: in its most
straightforward form, the constitutivist theory actually has more in common with
Humeanism than Kantianism.
This claim may come as a surprise. Currently, the literature contains two worked-
out versions of constitutivism. Christine Korsgaard’s version of constitutivism is unam-
biguously Kantian: she seeks to show that “Kant’s two imperatives of practical
reason”—the Hypothetical and Categorical Imperatives—are “constitutive principles
of action, principles to which we are necessarily trying to conform insofar as we are
acting at all” (Korsgaard 2009, xii). David Velleman denies that we can extract Kant’s
Categorical Imperative from the concept of agency, but nonetheless claims that “the
aim constitutive of agency can be seen to have pushed us in the direction of our moral
way of life,” making morality a “rational development, a form of rational progress”
(Velleman 2009, 149). Velleman calls this a “Kinda Kantian strategy” (Velleman 2009,
149). In light of these remarks by Korsgaard and Velleman, we might assume that
constitutivism necessarily takes a Kantian form.
But that would be a mistake. Constitutivism is based on the idea that action has a
constitutive feature whose presence yields substantive normative content. Whether
particular constitutivist theories take a Kantian form depends on the content of this feature
(and perhaps also the way in which the theorist argues for the presence of the aim).
To see this, consider an example. Suppose we start with the most minimal concep-
tion of action: to act is simply to bring something about. On this interpretation, the
paradigmatic case of action has the following form: I desire some end X, I see that
I could get X by doing Y, so I do Y. Action aims merely at effecting a change in the
world, so that the world conforms to my desires.
Mill endorsed this conception of action. As he put it,
All action is for the sake of some end, and rules of action, it seems natural to suppose, must take
their whole color and character from the end to which they are subservient. (Mill 2002, 2)
So, to act is simply to try to bring about some desired end; the rules of action, the
standards of success for action, pertain solely to how well the action brings about this end.
Let’s translate Mill’s point into the terminology of constitutivism. What is consti-
tutive of action is simply what is constitutive of bringing about ends. Is there anything
that is constitutive of bringing about ends? Well, a condition on bringing about an end
is taking the necessary means to that end. Thus, if you aim to bring about an end, you
must aim to take the means to that end. So:
(1) An agent’s ç-ing is an action iff in ç-ing the agent aims to bring about some end.
(2) An agent aims to bring about an end iff the agent aims to take some of the
necessary and available means to this end.
(3) Therefore, an agent’s ç-ing is an action iff in ç-ing the agent aims to take some
of the necessary and available means to her end.
From (3), it follows that in each token of action, the agent aims at taking the necessary
and available means; it also follows that aiming at taking the necessary and available
means is part of what makes an event an instance of action. By the definition of
Constitutive Aim, this is just to say that the taking the necessary and available means is a
constitutive aim of action.
But now consider the instrumental principle: if an agent aims to E, and M is a
necessary and available means for E-ing, then the agent has prima facie reason to
M. Notice that we can derive this principle from the conjunction of Success and claim
(3): claim (3) entails that agents constitutively aim at taking the necessary and available
means to their ends; Success claims that if an agent aims at X, she has prima facie reason
to X. Together, these claims entail the instrumental principle.
In short: we can derive the instrumental principle from a very minimal conception
of action.77 We might describe this as Humean constitutivism. As we saw above, Humean
views maintain that all reasons are derived from facts about our contingent aims: I have
reason to ç only if I have some motive that is suitably connected to ç-ing.78 They face
the question of why—what grounds this commitment? One possible answer is the
argument above: taking the means to one’s ends is a constitutive aim of action.
Thus, a Humean version of constitutivism would claim that there is only one
universal normative principle, directing us to take the means to our ends. The content
of these ends is given by our contingent, subjective motivational states. Humean
constitutivists would therefore deny that any substantive universal content follows
merely from facts about agency. They maintain, instead, that in order to derive
substantive content, we need to appeal to contingent and variable facts about the
motives of particular agents. So there will be no universal reasons derived from facts
about the nature of agency; all reasons will be subjective. The variation in reasons will
be as wide as the variation in motivational states between agents.79
77
James Dreier (1997), Christine Korsgaard (1997), and Michael Smith (2012) defend similar claims.
78
There are two complications. First, some Humeans reject the idea that there are rational requirements
of any form, including the instrumental principle. See Millgram (1995) and Hampton (1998). These
Humeans will want to deny Success. Second, some claim that the instrumental principle expresses a rational
requirement rather than a source of reasons: see, for example, Broome (1999). The idea, here, is that the
instrumental principle governs combinations of attitudes, telling us either to give up our end or to take the
means to it. I will address this issue in the next chapter.
79
If some particular motivational states are ubiquitous, then the reasons derived from these motives will
also be ubiquitous. For example, sympathy might be widespread enough to generate nearly universal reasons;
CONSTITUTIVISM 45
At the other end of the spectrum, we have Kantian constitutivism. Korsgaard’s version
of constitutivism posits both a version of the instrumental principle (Kant’s hypothet-
ical imperative) and the Categorical Imperative as constitutive of agency. This is a
maximally ambitious version of constitutivism: it aspires to show that the entirety of
our current moral code can be extracted from facts about what is constitutive of
agency. As Korsgaard puts it, “Enlightenment morality” can be derived from facts
about agency (1996b, 123).
So a maximally ambitious version of constitutivism claims that the entirety of
“Enlightenment morality” can be extracted from what is constitutive of agency,
whereas a minimally ambitious version claims that only the instrumental principle
can be so extracted. This leaves plenty of room for more moderate views, which would
fall between the two extremes. Velleman’s theory is a good example. He denies that
we can extract our moral code from facts about agency (Velleman 2009, Chapter 5).
But Velleman does think that we can extract some substantive normative content,
including various norms that “favor morality without requiring or guaranteeing it,”
from facts about agency (Velleman 2009, 149). In short, we can get more than the
instrumental principle but less than Enlightenment morality.80
But there is a complication: it’s not just the content of the constitutivist views that
distinguishes them. There are also differences in the characterization of the constitutive
feature that purportedly yields this content. Humean versions of constitutivism focus on
the presence of a constitutive aim, as I explained above. But a more properly Kantian
version of constitutivism would view mere aims as inadequate for generating normative
content; aims, along with associated motives, are (so the Kantian story goes) external to
the will, so acting upon them would result in heteronomy.81 Thus, when we turn to
Korsgaard’s Kantian version of constitutivism, we find reliance on a somewhat different
constitutive feature: not a constitutive aim but a constitutive principle. I will explain this
point in Chapter 4, when we consider Korsgaard’s Kantian theory. For now, it simply
bears noting that the Kantian version of constitutivism is based on the idea that action
constitutively involves commitment to certain principles. The normative principles that
govern the will are the principles that are constitutive of action itself. So, the Humean
constitutivist shows that actions has constitutive aims, and appeals to Success in order to
derive substantive normative content. The Kantian constitutivist, by contrast, argues
that action requires commitment to constitutive principles, and derives substantive
normative content from the agent’s commitment to these principles.
alternatively, if long-term self-interest were universal, we could appeal to Hobbesian considerations to justify
certain reasons. My point is simply that the nature of agency itself won’t underwrite these conclusions.
80
Michael Smith (2012) briefly (and tentatively) argues for a rather different version of constitutivism,
which centers on the claim that rationality requires that people intrinsically desire that there is as much
happiness as possible.
81
Kant writes, “if the will seeks the law that is to determine it anywhere else than in the fitness of its
maxims for its own giving of universal law . . . heteronomy always results” (Groundwork 4:441). I will address
this issue in Chapter 4.
In sum, constitutivism is a label for a broad range of views united by their attempt to
ground normativity in facts about what is constitutive of agency. There is conceptual space
for many different versions of constitutivism: we can be Humean constitutivists, Velle-
manian constitutivists, Kantian constitutivists, and much else besides. In the following
chapters, I will be arguing for a version of constitutivism that—like Velleman’s view—falls
between the Humean and Kantian extremes. This will be a Nietzschean constitutivism.
4. Conclusion
In this chapter, I have suggested that the familiar ethical theories—non-reductive
realism, Aristotelianism, Humeanism, Kantianism—have difficulty overcoming the
epistemological, metaphysical, and practical challenges. Each of the familiar theories
has trouble with at least one of these challenges. A constitutivist ethical theory, by
contrast, would overcome the challenges with ease. For constitutivism extracts norma-
tive content from the structure of agency, by showing that all actions share a common
aim. In deriving normative content from this aim, it has the potential to avoid the
practical and metaphysical problems; and in giving us a way to assess our current moral
beliefs, it can avoid the epistemological problem.
But can we develop a successful version of constitutivism? The theory is ambitious,
attempting to extract normative content from the bare idea of agency. In light of this, it
has been subject to a number of seemingly powerful objections. The next chapter
examines and assesses these objections.
2
Normativity as Inescapability
The previous chapter introduced the constitutivist project and explained the advan-
tages that it enjoys over competing answers to the foundational question in ethics.
Constitutivism promises to avoid the epistemological, metaphysical, and practical
problems by showing that universal reasons can be derived from an aim that is
inescapably present in all action.
But can this strategy succeed? Recently, a number of objections have been levied at
the very possibility of a constitutivist theory. Kieran Setiya, David Enoch, and others
claim to have shown that constitutivism is hopeless: as Enoch puts it, “normativity
cannot be grounded in what is constitutive of agency” (Enoch 2006, 192). In this
chapter, I examine the extant objections to constitutivism. Ultimately, I argue that all
of these objections can be answered, for they turn on misunderstandings of the
constitutivist project. Thus, addressing these objections will show that constitutivism
can succeed while also clarifying constitutivism’s commitments.
Section 1 considers the objection that in order to show that a constitutive aim is
reason-providing, we must show that there is reason to participate in the activity
governed by the aim. Section 2 discusses David Enoch’s claim that even if there are
constitutive aims of agency, we can escape the reasons engendered by these aims by
engaging in some alternative form of agency. Section 3 considers a more recent
objection from Enoch: the mere fact that something is inescapable does not entail
that it is normative. In Section 4, I turn to the objection that the constitutivist engages
in the allegedly wrongheaded attempt to derive an ought from an is. Section 5 considers
a final objection. This objection asks why, even if some aim is inescapable, we should
strive maximally to fulfill it: why not, instead, fulfill it only reluctantly and half-
heartedly? While I contend that all of these objections can be answered, the fifth is
by far the most serious.
In addressing these objections, I will return time and again to a common theme: the
constitutivist maintains that certain aims or standards have a privileged normative status
precisely because they are inescapable. I hope to show that this is constitutivism’s core
idea: the authority of universal normative claims arises from a certain form of inescap-
ability. Thus, in addition to showing that constitutivism survives various objections,
this chapter will elucidate constitutivism’s motivating idea.
48 N O R M AT I V I T Y A S I N E S C A PA B I L I T Y
1. First objection: the standards of a practice are

reason-providing only if one has reason to participate
in the practice
In presenting the constitutivist theory, I employed an analogy between action and
games. Games have constitutive aims or constitutive rules, and their normativity seems
unproblematic. The hope was that we could say something similar about action itself.
However, upon further scrutiny the analogy with games might seem problematic. For
there are two possible objections to the idea that the rules or aims of games are
automatically normative. First, we might think that the rules or aims of games generate
reasons only if one has reason to play the game in the first place. Second, we might
point out that the constitutive rules of games can be escaped quite easily: if you don’t
like the rules of chess, you can abandon some of these rules and play a slight variant of
chess. Below, I explain how each of these observations can be transformed into an
objection to constitutivism.
1.1 Distinguishing two claims about reasons
Consider the claim that participants in a game have reason to abide by the rules that are
constitutive of the game. A possible objection arises at this point. The reasons that
spring from an agent’s participation in a game interact with, and in some cases seem to
depend upon, reasons that do not derive from the game. Consider an agent who plays
baseball. Insofar as the agent is playing baseball, she has reason to aim at scoring more
runs than her opponent. But there might be other reasons that interact with and
modify this reason. If some members of the opposing team are children who are just
learning the game, then she might adopt the aim of letting them have a fair chance at
winning. So, for example, she might throw them pitches that are easy to hit, or she
might allow them four strikes instead of the usual three.
Alternatively, suppose that the agent is playing baseball in her yard and suddenly
notices that her house is burning down behind her. She certainly has a very strong
reason to stop trying to score more runs than the opposing team and to call the fire
department instead. So here we have a case in which the reasons arising from the
constitutive standard are rendered irrelevant by the fact that there is a very strong reason
not to continue participating in the game.
Recognition of these points sometimes generates an important question. Might the
reasons that appear to derive from one’s participation in a game really be dependent
upon one’s having reasons to participate in the game? There are two possibilities. One
possibility is that the constitutive aim of a game is originative of reasons:
(1) Constitutive Aims as Originative of Reasons: If you participate in an activity A,
then the constitutive aim of A is reason-providing.
Another possibility is that constitutive aims merely “transfer” normativity, in the
following sense:
FIRST OBJECTION 49
(2) Constitutive Aims as Transferring Reasons: If you have reason to participate in A,

then the constitutive aim of A is reason-providing.
An example will help to clarify these possibilities. Take baseball:
(Baseball as originative of reasons) If you participate in baseball, then the aim of out-
scoring your opponent provides you with reason to try to outscore your opponent.
(Baseball as transferring reasons) If you have reason to participate in baseball, then the
aim of outscoring your opponent provides you with reason to try to outscore your
opponent.
I call the first option “Originative” because it views reasons as arising merely from the
fact that one participates in a practice. I call the second option “Transferring” because it
views the reasons internal to a practice as having normative force transferred to them by
reasons to participate in the practice.
To see why this distinction is important, suppose (2) is correct. The constitutivist
wants to show that action itself has a constitutive aim, from which we can derive
universal reasons. But if (2) is correct, then the constitutivist will need to do a bit more:
not just show that action has a constitutive aim, but also show that we have a reason for
action itself. That might raise several problems. First, it would prevent constitutivism
from being a fully general account of reasons for action: there would be at least one
reason that could not be analyzed in the constitutivist fashion. Second, (2) would require
that we state constitutivism as follows: if you have reason to engage in action itself, then
you have reason to X. If the reason in the antecedent is aim-dependent or otherwise
contingent, then all reasons springing from the constitutive feature will be contingent as
well. In short, the reasons deriving from action’s constitutive feature would be analo-
gous to the reasons deriving from the constitutive features of baseball—not genuinely
universal after all, since not everyone has reason to play baseball.
On the other hand, if (1) is correct then it is possible for constitutivism to be a fully
general account of reasons for action and to yield universal reasons. So which claim is
correct?
In his objection to constitutivism, Enoch assumes that (2) is correct. He writes,
“even if you somehow find yourself playing chess, and even if checkmating your
opponent is a constitutive aim of playing chess, still you may not have reason to (try to)
checkmate your opponent. You may lack such a reason if you lack a reason to play
chess” (Enoch 2011, 211). Here, he relies on the idea that constitutive aims are reason
providing only if one has reason to engage in the practice that they govern. As he puts
it, in addition to showing that an activity has a constitutive aim, “what is also needed is
that you have a reason to engage in that activity” (2011, 211).
So Enoch rejects (1) and relies on (2). But what is Enoch’s argument for this position?
He rests his case on the following point: whenever I play chess, I can also be described as
engaging in slight variants of chess. For example, define chess* as a game analogous to
chess in all but one respect: you only win if you checkmate your opponent in an even
number of moves. The early stages of a chess game will look exactly analogous to the
early stages of a chess* game. So, in satisfying the constitutive aim of chess, I will also be
satisfying the constitutive aim of chess*. But, Enoch notes, “you don’t have reason to
play chess*” (2011, 211). For, as he puts it, “it’s not generally true that engaging in
some activity—satisfying some relevant description—suffices for having reason to
direct oneself at its constitutive aim” (2011, 211). In other words: the mere fact that
I can be described as participating in an activity with a constitutive aim does not imply
that I have reason to fulfill that aim.
This objection is misguided. It has been familiar since Anscombe (2000) and
Davidson (1980) that we need to distinguish true descriptions of the agent’s action
from descriptions under which the agent’s action is intentional. “Playing chess*” may be a
true description of a person’s action, but it is not a description under which the action is
intentional. Just so, “moving air molecules,” “causing a shadow to fall on the pawn,”
and “irritating my opponent” may be true descriptions of what I am doing, but they are
not descriptions under which my action is intentional. Given that constitutivists are
attempting to show that intentional action has a constitutive aim, these reflections on
chess* are simply beside the point. Put differently: the agent isn’t intentionally playing
chess*, so it is unsurprising that the constitutive aim of chess* isn’t reason-providing.
So Enoch’s argument in favor of (2) is flawed. But might there be other arguments?
Well, notice that (1) and (2) mirror contrary positions that arise in disputes about
instrumental reason. The instrumental principle can be put as follows: if you have an
end E, and M is a necessary and available means to E, then you have reason to
M. A spirited debate has arisen over whether merely by intending an evil end you
can give yourself a reason to take the means. For example, suppose I intend to murder
Bob: does this give me reason to purchase a gun? Some philosophers are uneasy about
saying yes. In light of these sorts of examples, some have defended alternative versions
of the instrumental principle. For example:
(A) If [you have an end E, and you have reason to have this end, and M is a
necessary and available means to E], then you have reason to M.1
(B) Rationality requires that if you have an end E, then [either you give up this end
or you take the necessary and available means to E].2
Analogous reflections might lead us to think that there are counterexamples to (1).
After all, there are activities with constitutive aims that some philosophers are reluctant
to see as reason-providing. For example, consider the activity of being an assassin.
1
For this version of the instrumental principle, see for example Raz (2005). He states his “facilitative
principle” as follows: “When we have an undefeated reason to take an action, we have reason to perform any
one (but only one) of the possible (for us) alternative plans that facilitate its performance” (2005, 5–6; italics
removed).
2
For this version of the instrumental principle, see for example Broome (1999). Broome calls this a wide-
scope requirement: it expresses a requirement not to have certain combinations of attitudes, rather than to
perform particular actions.
FIRST OBJECTION 51
There seems to be a constitutive aim here: say, killing one’s target without being
detected. This aim seems to generate reasons to be stealthy, to avoid capture, to have
good aim, to kill one’s target rather than merely wounding him, and so forth.3 But we
are also inclined to say that there is reason not to be an assassin. Perhaps this entails that
even if you participate in the activity of being an assassin, you have no reason to govern
yourself with assassination’s constitutive aim. If so, we have a case in which there is a
constitutive aim, but the constitutive aim is not reason-providing.4
Perhaps, then, (2) is the right diagnosis: the constitutive aim of a practice is reason-
providing only if you have reason to engage in the practice. This would be a version of
principle (A), above. Or perhaps principle (B) is correct: perhaps violations of consti-
tutive standards are violations of rational requirements. In other words:
(A*) If [you are A-ing, and you have reason to A, and A-ing has constitutive
aim E], then you have reason to pursue E.
(B*) Rationality requires that if [you participate in A, and A-ing has constitutive
aim E], then [either you stop participating in A, or you pursue E]
Any arguments in favor of (A) would carry over to (A*); likewise, defenses of (B) could
be translated into defenses of (B*).
These issues are hotly disputed, so it would be nice if the constitutivist could avoid
taking a stand. And indeed, I think she can. To see this, start with (B*). This would be
no problem for the constitutivist: no one will deny that you can avoid the constitutive
aim of chess by ceasing to play chess. So, if (B*) is right, the constitutivist has no
problem. Now consider (A*). This is the principle that Enoch endorses. Initially, this
seems problematic for the constitutivist: for, when we consider the case of action, it
implies that you would need a reason to engage in action. But I’m going to argue that
even if (A*) were true in general, it could not apply to the case of action as such.
3
I am assuming that in order to be an assassin, one must aim at killing one’s targets without being detected,
and that having this aim is part of what constitutes a stretch of activity as an instance of assassination-behavior.
This constitutive aim would generate the standards mentioned above.
4
Cohen (1996) raises a version of this objection to Christine Korsgaard’s attempt to derive ethical claims
from facts about what is constitutive of certain practical identities (see Korsgaard 1996b). Cohen’s objection
can be put as follows: certain practical identities have constitutive features that do not seem to be reason-
providing. He uses the example of a Mafioso; the standards constitutive of being a Mafioso are not ones that
we would typically regard as reason-providing. For example, adopting the practical identity Mafioso seems
constitutively to involve commitment to intimidating witnesses, lying to the police, and so forth; but we
don’t want to say that individuals have reasons to do these things. Korsgaard responds to this objection by
maintaining that (1) all practical identities do generate reasons, but (2) every individual is committed to the
practical identity human being, and (3) the reasons generated by human being conflict with and overrule some of
the reasons generated by more specific practical identities (such as Mafioso). As she puts it, the Mafioso’s
“obligations are real”—in other words, he really does have reason to intimidate witnesses and so forth—but
“his obligation to be a good person is . . . deeper than his obligation to stick to his [Mafioso] code” (Korsgaard
1996b, 257–8). Thus, Korsgaard relies on the idea that more general (or universal) practical identities can
overrule less general ones. This Kantian response to the problem seems to me flawed, for reasons that I will
explain in Chapter 4. Accordingly, I offer a different response in the text above.
1.2 Why the answer to our question doesn’t matter

I just argued for two points. First, Enoch’s objection to constitutivism depends on (2),
but we lack any convincing arguments for (2). Second, a Broomean conception of
instrumental rationality, encapsulated in claim (B*), could be deployed to argue in
favor of (2). So we need to determine whether this argument would succeed. I think
we can develop an argument that shows that at least in the case of action itself, (2) must
be false.
To argue for this position, let’s simply grant that (2) is true for practices other than
action itself. Even if the reason-providing force of a practice’s constitutive aim were in
general dependent upon one’s having a reason to engage in the practice, this could not
be true when the practice in question is action itself. For I will suggest that action is
crucially different from other, more particular types of practices with constitutive aims.
To bring out the difference between action and more particular practices, consider
the meaning of the question, “Do I have reason to act?” It is important to be clear on
how this question should be understood. This isn’t a question about whether I should
perform particular actions; it isn’t a question about whether I should play chess, or study,
or take a stroll through the park. Nor is it a question about whether I should refrain from
particular actions; it isn’t a question about whether I should refrain from playing chess,
or studying, or taking a stroll. Rather, the question is whether I should perform any actions
at all.
But what exactly is a person asking, when she asks whether there is reason to act? It
certainly makes sense to ask whether there is reason to engage in particular forms of
action, such as chess-playing. Sometimes there is good reason to play chess; other times
there is good reason not to play chess. However, the question whether there is a reason
to A seems crucially different when A does not denote some particular practice, but
denotes action itself.
Asking whether there is reason to A presupposes that there is some alternative to
A-ing. For example, an agent who asks whether he has a reason to play chess is asking
whether, rather than playing chess, he should perform some alternative action (perhaps
he will go for a walk instead). But when an agent asks whether there is reason to
perform any actions at all, the question must have a different sense. It is not as if the
agent can do something other than performing actions; action is inescapable. After all,
constitutivists use the term “action” to pick out our intentional activities. Anything
that we do intentionally counts as an action. If action has a constitutive feature, this
feature will be present in everything we do intentionally. “Action,” as the constitutivist
uses it, just means intentional activity. Accordingly, the question whether there is a
reason to perform any actions at all is moot.
It might seem that there is an alternative to acting: one could decide not to perform
any actions at all. For example, one could decide to go to sleep, in order to avoid
performing any actions for a few hours. However, notice that deciding not to act is itself
an action. Although one can put oneself into a position in which one ceases to perform
S E C O N D O B J E C T I O N : A G E N C Y, S C H M A G E N C Y 53
actions, the decision to put oneself in this position, and the execution of the decision,
will themselves be actions. In this sense, action is inescapable: any attempt to avoid
acting will itself be an action.
Thus, if I am correct in suggesting that the question whether there is reason to
A presupposes that there is some alternative to A-ing, then the question whether there
is reason for action as such does not make sense. There are reasons to perform particular
actions, and there are reasons not to perform others; but there is no such thing as a
reason to perform action as such.5
Suppose this is right: there is no such thing as a reason to perform action as such. We
can now return to the objection from Section 1.1. The objection is this: the consti-
tutive aim of A is reason-providing only if there is a reason to A. If this were right, then
the constitutivist about action would have to show that there is a reason to act.
However, the argument above suggests that the objection is misguided. It is not that
there could be, but turns out not to be, a reason for action as such. Rather, the very
idea of a reason for action as such is incoherent. So the objection, which demands that
the constitutivist about action demonstrate that there is a reason for action, fails. At least
in the case of action itself, (1) is the correct diagnosis.6
2. Second objection: agency, schmagency

Constitutivists need not establish that there is a reason for action as such. There isn’t—
indeed, there cannot be—a reason for action as such, but the constitutive aim of action
nonetheless generates reasons. The first objection to constitutivism therefore fails.
Let’s now consider a second objection. David Enoch has recently argued that even if
action has a constitutive aim, one can escape the reasons generated by this aim by
engaging in a slight variant of action. If this is so, the reasons generated by action’s
constitutive aim won’t be universal, after all. Below, I examine this objection, arguing
that it rests on a misunderstanding of the constitutivist’s project.
2.1 Enoch’s objection to constitutivism
One rule of chess is this: the first time a pawn is moved, it may be moved either one or
two spaces forward; on subsequent moves, a pawn may be moved one space forward.
Suppose I notice that Sarah and Ted seem to be playing chess but ignoring this rule:
5
Consider an analogous case: there cannot be a theoretical reason to have beliefs as such. There can be
theoretical reasons to have particular beliefs, but not to be a believer. (There can, however, be a practical
reason to be (or not to be) a believer. For example, according to a common interpretation, the Pyrrhonian
skeptics attempt to give us a practical reason for withholding assent from all beliefs: by doing so, we attain
ataraxia [tranquillity].)
6
Enoch considers a version of this response, which he attributes to Velleman. Enoch takes Velleman to
argue as follows: the inescapability of agency “renders moot the question of whether one should care [about
the constitutive aim]” (2011, 216). But what’s important is not whether we should care about the constitutive
aim; what’s important is whether the idea of a reason for action as such is intelligible. I will return to this point
in Section 3, below.
they are moving their pawns two spaces forward whenever they like. A stickler for the
rules of chess, I say to them, “You know, you’re playing a defective game of chess.
After all, one of the rules of chess is that, after their first move, pawns may only be
moved one space forward.” They respond, “Oh, we know. But we’ve decided to
modify that rule. It’s more fun if you give the pawns a greater range of movement.”
How would a constitutivist about chess react to this case? I might point out to Sarah
and Ted that, strictly speaking, they are not playing chess, but are instead playing a slight
variant of chess; call it schmess. Nothing wrong with that; they will presumably accept
my characterization. Moreover, precisely because they are not trying to play chess, the
rules pertaining to chess don’t apply to them; it is no criticism of them that they’re
ignoring the pawn rule, because the pawn rule doesn’t apply to them.
This case illustrates a more general phenomenon: you can escape the standards of a
practice by engaging in a new practice. If you don’t like one of the rules of a practice,
you can simply engage in a slight variant of the practice that doesn’t include the
disagreeable rule. But if that’s true in the case of chess, might it also be true in
the case of action? Suppose that action has a constitutive aim of self-constitution (as
Korsgaard argues). If Sarah doesn’t like the constitutive aim of action, maybe she
could just stop engaging in actions, and start engaging in schmactions. Schmactions are
just like actions, except that schmactions don’t include an aim of self-constitution.
Enoch argues that this possibility shows that constitutivism is hopeless. For,
presented with the idea that, say, self-constitution is constitutive of agency, someone
can respond, “I am perfectly happy performing schmactions—nonaction events
that are very similar to actions but that lack the aim (constitutive of actions but not
of schmactions) of self-constitution” (Enoch 2006, 179). If that is right, then the
reasons generated by constitutive aims are not universal after all, and constitutivism
fails.
In fact, though, the objection misses its mark, for it misconceives the constitutivist
project. As the prior section explained, the point of constitutivism is that action is
inescapable; we have no alternative to performing actions. One can decide to play
schmess instead of chess, but one cannot decide to perform a schmaction instead
of an action. For the very process of deciding or trying to produce a schmaction would
itself be an action, and would therefore manifest action’s constitutive aim. After all, as
I noted above, by “action” the constitutivist just means intentional activity.
Any intentional activity that the agent performs will count as an action. Thus, the
idea that there could be a schmaction—an intentional activity that is not an action—is
self-contradictory.
Notice that I am here just reiterating a point I made earlier: constitutivists about
action rely on the claim that action is inescapable. Games and practices have escapable
constitutive standards; the standards cease to apply once you stop playing the game. So,
too, action would have escapable standards, if we could stop acting. But we cannot stop
acting, so the constitutive aim of action is inescapable. Put simply: you can decide to
S E C O N D O B J E C T I O N : A G E N C Y, S C H M A G E N C Y 55
play schmess instead of playing chess, but you cannot decide to produce schmactions
instead of actions.7,8
2.2 The glimmer of truth in the schmaction objection
The schmaction objection is easily answered, because it simply fails to address the
constitutivist claim that action is inescapable. However, there is a glimmer of truth in
the objection; it does raise one important point. The constitutivist is relying on the idea
that action is inescapable. So we have to show that the constitutive aims pertain to
something that really is inescapable.
Consider an analogy. Suppose I claim that obeying the posted speed limit is consti-
tutive of driving: what it is to drive is, in part, to obey the speed limit. This would of
course entail that an aim is built into the structure of driving. But the account of driving
is completely implausible. Of course the bare notion of driving doesn’t just by itself
contain any aim pertaining to speed limits. So it would be perfectly legitimate for
someone to object: “Fine, reserve the word ‘driving’ for driving at the speed limit.
I don’t drive at the speed limit, so you can call what I’m doing schmiving. That’s nothing
to me.”
The above objection would be perfectly sound. And it would be sound because you
can drive without aiming to obey the speed limit. The case of action is not supposed to
be like that. Constitutivists purport to show that we really cannot act at all without
aiming at the constitutive feature of action. But, just as an overzealous account of
driving might implausibly claim that speed limits are part of the concept of driving, so
too an overzealous constitutivist might implausibly claim that some optional feature is
constitutive of action.
So the lesson of the schmaction objection is this: we have to show that the consti-
tutive feature really is inescapable for agents. In other words, we have to show that
“action” isn’t just an honorific that we bestow on certain cases of agency, but that every
case of agency manifests the constitutive feature.9
7
You can, of course, set out to act and fail. One way of failing to act might be producing a schmaction.
The important point, though, is that trying to produce a schmaction is itself an action.
8
Ferrero (2009) responds to Enoch’s objection in an analogous manner, writing: “If one can stand outside
chess and question whether there is any reason to play this game, why couldn’t one stand outside of agency
and wonder whether there is any reason to play the agency game? The problem with this suggestion is that
the analogy does not hold . . . Agency is special under two respects. First, agency is the enterprise with the
largest jurisdiction. All ordinary enterprises fall under it . . . The second feature that makes agency stand apart
from ordinary enterprises is agency’s closure. Agency is closed under the operation of reflective rational
assessment” (2009, 308). In other words, Ferrero notes that there is always the possibility of stepping outside a
practice and reflecting on its justification. But this isn’t possible with agency. Reflecting on agency is itself an
action, so any attempt to step outside of it will itself be an action. As Ferrero puts it, “even reflection about
agency is a manifestation of agency” (2009, 309). Thus, he concludes that “the combination of these features
is what makes agency inescapable” (2009, 309).
9
The language that Enoch employs in his article suggests that he conceives of constitutivists as employing
the label “action” as a mere honorific. For example, consider the following passage: “Korsgaard writes as if
she believes that the threat that your inner (and outer) states will fail to deserve folk-theoretical terms (such as
‘action’) is indeed a threat that will strike terror in the hearts of the wicked” (2006, 179–80). Here, Enoch
Thus, the schmaction objection does help to remind us that constitutivist accounts of
action must avoid building too much normative structure into the starting notion of
action. Of course, once we put the “objection” in this form, we can see that it isn’t an
objection at all. Rather, it simply discloses a criterion of adequacy for a successful
constitutivist theory: the constitutivist needs to start with a conception of action that
makes action inescapable.
3. Third objection: inescapability does not

imply normativity
My responses to the above objections rest on the idea that action is inescapable. But
Enoch has recently objected to the constitutivists’ appeal to inescapability, writing,
The move from “You inescapably ” to “You should ” is no better—not even that tiniest little
bit—than the move from “You actually ” to “You should ”. (Enoch 2011, 216)
An example can illustrate Enoch’s point. Suppose that I have a firmly entrenched,
inescapable addiction to smoking. It hardly seems to follow that I should smoke. So
inescapable tendencies don’t always generate reasons. Does this imply that the con-
stitutivist is making a grave error?
Not at all. The first thing to notice is that Enoch has misstated the constitutivist
strategy. The move is not from “you inescapably ” to “you should .” Rather, the
aim-based version of constitutivism moves from “you inescapably aim at -ing” to
“you should .”10
This is a subtle but crucial distinction. For recall that I pointed out, in the last
chapter, that the constitutivist relies on the following claim:
(Success) If X aims at G, then G is a standard of success for X.

The mere fact that I do something doesn’t generate standards of success. If I catch my
foot on a crack in the sidewalk and trip, it does not follow that there is any standard of
success for my tripping. If I forget my keys in the morning, it does not follow that there
is a standard of success for forgetting my keys.
However, Success claims that aims generate standards of success. If I aim to catch my
foot on a crack in such a way that it causes me to trip (perhaps I want to feign an injury,
or perhaps I’m acting in a play) then there is a standard of success. If I aim to forget my
seems to be suggesting that constitutivism is a point about language: the agent should strive to conform to the
constitutive aim of action because, if he doesn’t, he won’t deserve to have his behavior labeled “action.” If
this is what Enoch means, he is simply misinterpreting constitutivism. The constitutivist isn’t claiming that we
should strive to fulfill certain aims so that our behavior merits the label “action”; rather, the constitutivist is
claiming that we cannot do anything but strive to fulfill a certain aim. Put simply, constitutivism is not a point
about language. It is a point about the structure of action.
10
Notice that the move is not from “you inescapably aim at -ing” to “you should aim at -ing.”
Rather, it is from “you inescapably aim at -ing” to “you should .”
F O U RT H O B J E C T I O N 57
keys in the morning (perhaps, the night before, I put them in an out of the way place
rather than in their usual spot by the door, and hoped that I wouldn’t remember in the
morning), then there is a standard of success. It is aims, not mere activities, that generate
standards of success.
That is the first crucial response to Enoch’s claim. But there is more to say. Notice
that there are different ways of formulating the Success claim. A strong claim would be
that inescapable aims, and only inescapable aims, generate standards of success.
A weaker claim would be that all aims generate standards of success. I employ only
the weaker claim. Constitutive aims differ from ordinary aims only in that constitutive
aims are inescapable, whereas ordinary aims are not. It isn’t the inescapability that is
reason-providing. The aim itself—any aim—is reason-providing. The inescapability is
just a point about how ubiquitous the aim is, not about why it is reason-providing.
With these clarifications in mind, we can see that constitutivism rests on a very spare
claim about reasons. All that we need, in order for the constitutivist project to work, is
the claim that aims in general generate standards of success. While no philosophical
thesis is entirely uncontroversial, this claim is widely accepted.
4. Fourth objection: the mere fact that we have an aim

does not generate normative conclusions
In the previous sections, I have argued that constitutivism is not susceptible to a range
of recent objections. However, one might still worry about the foundational norma-
tive claim that the constitutivist employs: how, exactly, does the mere fact that we have
an aim generate normative conclusions? In particular, one might object that this
strategy threatens to reduce normativity to brute compulsion, or that it illegitimately
derives an “ought” from an “is.” The follow sections examine these objections in turn.
4.1 Does constitutivism reduce normativity to compulsion?
Constitutivism relies on the idea that aims generate reasons: because we aim to A, we
have reason to A. This might seem to reduce constitutivism to nothing more than the
kind of naturalism inveighed against by G. E. Moore (1971). This form of naturalism
claims that because we aim to A, or because we are disposed to A, we ought to
A. Moore claimed that any such naturalism would be susceptible to the open question
argument: even if a naturalist can establish that we do aim at A, we can legitimately ask
whether it is good to aim at A. This, Moore claims, is a decisive objection to the
naturalist proposal.
Developing these ideas, Connie Rosati writes,
Reflect for a moment on one of the kinds of brute naturalism examined by Moore, the account
that appeals to evolution and says, roughly, that we ought to follow the direction in which we
have been developing. An account of our ethical notions that treats any ethical property as
identical with the property of being more evolved or following the direction set by evolution, or
any other brute, natural property, construes our ethical notions and ethical properties themselves
in a way that bears no relation to our agency. It is no part of the nature of such properties that
what has them reflect, comport with, or embody the evaluations or critical responses of any
persons, actual or hypothetical, imperfect or ideal. It is no part of the nature of such properties
that what has them fit with any fundamental concerns or reflectively supportable aims persons
might have. The property of being good for a person, for instance, is on this account entirely
unrelated to our sentiments, choices, or decisions. But it seems utterly implausible that personal
good could be entirely unrelated in this way to anything we care about. (Rosati 2003, 508)
In this passage, Rosati poses Moore’s problem: merely aiming to A or being disposed
toward A can come apart from our choices, critical responses, and concerns. To clarify
this point, consider an example that one often finds in popularizations of science:
evolutionary biologists show that human beings aim at spreading their genes as far as
possible. Grant that I have a tendency or disposition to do so. I can still ask whether
I should do so. I can ask whether I want to pursue this end, or whether I should reject it
in favor of other ends. After all, why should a biologically induced disposition matter to
me? Or, to pick a more invidious example: suppose a child with a racist upbringing is
strongly disposed to fear and hate members of other races. Surely, the child should not
take the presence of this disposition as a reason for being racist. In general, the mere fact
that I am disposed to A seems irrelevant, normatively speaking: we can grant that I’m
disposed to A and reject the claim that I have reason to A.
It might seem that a similar objection applies to constitutivism. We can grant that we
constitutively aim at A, and nonetheless question whether we should pursue
A. However, Rosati argues that this is mistaken, for there is a crucial distinction
between constitutive aims and mere dispositions. As she puts it,
Unlike our other motives and capacities, our autonomy-making motives and capacities are not
arbitrary but, rather, make self-governance possible: they are motives and capacities without the
effective operation of which we would not be agents and evaluators at all. Insofar as we are
agents, the effective exercise of these motives and capacities matters to us, and our caring about
them involves no identifiable mistake. Their operation, we might say, is self-vindicating, and
efforts to challenge them cannot even get going without relying on them. (Rosati 2003, 521–2)
Rosati makes two points in this passage. First, the constitutive aim (the “autonomy-
making motive”) is ineliminable: without it, we would not be agents at all. Second,
because the constitutive aim is an essential aspect of our nature as agents, Rosati
contends that we are justified in caring about it. Her idea, here, is that we care about
our own agency, and hence are justified in caring about the conditions of our own
agency.
To clarify these points, consider how they apply to chess and its constitutive aim of
checkmate. First, insofar as we are chess players, we aim at checkmate. Second, because
chess players tend to care about playing chess, they are justified in caring about chess’s
constitutive aim.
F O U RT H O B J E C T I O N 59
I think Rosati’s points are correct, but also incomplete: the constitutivist can make a
much stronger response than this to the Moorean objection. To be sure, caring about
the constitutive aim needn’t involve a mistake, and manifesting the aim will tend to
matter to us. However, the more important point is that, as I’ve explained above,
questioning the aim simply doesn’t make sense. Asking whether you should pursue
A assumes that you have some alternative to pursing A; it assumes that you could,
instead, pursue B. But, as I explained above, that is not the case with the constitutive
aim of action. Any particular action that the agent chooses to perform will be an action,
and will therefore fall under the standards governing action. By analogy, any particular
chess move that the chess player makes will still be a move in chess, and will therefore
fall under the standards governing chess. It doesn’t make sense for the chess player to
ask why he should aim to checkmate his opponent, unless this is construed as a question
about whether he should continue playing chess. For, so long as he is playing chess, he
is committed to aiming at checkmate; that’s part of what it is to play chess. Just so, so
long as an agent is acting, she will be committed to the standards constitutive of action.
In this respect, constitutive aims are distinct from ordinary dispositions and aims. But
let’s now turn to Rosati’s second point, concerning whether caring about the consti-
tutive aim is justified. Grant that agents tend to care about agency, and hence to care
about the conditions of their own agency. This is, as Rosati notes, one respect in which
constitutive aims differ from many other aims. However, the constitutivist again has a
deeper response to the Moorean objection.
To see this, consider what is meant by asking whether I should care about my aiming
at A. In posing this question, I might be asking whether I should aim at A. But if this is
my question, then it is moot: the aim will be present in all that I do.
Alternatively, the question might be whether I should manifest sentiments of approval
toward my pursuing A. Granted, this is a question that one can ask about any aim. The
question is typically quite important, because if I answer it negatively my motivation to
pursue A tends to dissipate. For example, if I decide that I should not approve of
pursuing a certain relationship, then I am at least marginally less likely to pursue it.
However, not pursuing a constitutive aim is not an option, so the question seems
decidedly less important when applied to constitutive aims.
To see this, consider the operation of constitutive aims in the game of chess. Imagine
two people playing a game of chess. Player One cares deeply about chess, loves playing
chess, would prefer playing chess to anything else. Player Two detests chess, but has no
alternative to playing—say, he has a gun to his head. Player One is going to be happier
than Player Two, Player One is going to enjoy the game more than Player Two, and so
on. But those aren’t the relevant questions with respect to practical reason. The
relevant question is just whether the constitutive aim is reason-providing for both of
these players. And clearly it is. What matters is just that they have the aim, not that they
approve of the aim.
So approval of the constitutive aim, or caring about the constitutive aim, would
not be normatively relevant. Caring does serve a function in the case of optional,
non-constitutive aims, for these aims often weaken once the agent’s approval dissipates.
In the case of the constitutive aim, though, approval cannot have this effect: if the aim is
constitutive of agency, it will be ineluctable. Approval of the aim might be nice, but it
is not necessary.
4.2 Does constitutivism attempt to derive an “ought” from an “is”?
For the above reasons, constitutivism is not susceptible to Moore’s objection. Let’s now
consider a closely related objection. One might argue that the constitutivist project
couldn’t possibly work, because it attempts to derive an “ought” from an “is.” In other
words, the constitutivist starts with a description of the nature of action, and tries to
show that this description yields normative claims. Doesn’t this violate that old
Humean chestnut about deriving an “ought” from an “is”?
Not at all. Here is an argument that would yield a violation of the Humean
principle:
(1) Action aims at A.
(2) Therefore, A is valuable.
This argument would move from a purely descriptive claim to a normative claim. But
that’s not what the constitutivist does. The constitutivist isn’t arguing that the goal set
by the constitutive aim of action is valuable, or that the goal is something that,
independently of its status as a goal, would be such that it ought to be pursued.
To see this, consider the analogous point about belief:
(1) Belief aims at truth.
(2) Therefore, truth is valuable.
This is a bad argument. It is also not an argument that a constitutivist about belief needs
to rely upon. It doesn’t matter whether truth is valuable. All that matters is that when
we are engaged in the process of forming beliefs, we must in fact regulate our
cognitions with the aim of rendering them true. Nothing of interest is added by (2).
The believer takes truth as valuable or to-be-pursued in the limited sense that he
treats truth as a standard in light of which he is to regulate his cognitions. This need not
imply that he also treats truth as a value, where this connotes treating truth as something
worthy of respect, admiration, or praise. More generally, valuing typically involves
having the above feelings, but these feelings need not be present for a standard to be
normative.
Or take another example: a constitutivist about chess needn’t show that checkmate
is valuable, or something that ought to be pursued. All she needs to show is that if you
are playing chess, then you are regulating your activities in order to achieve that aim.
For part of what it is to be a chess player is to take that standard as normative.
In fact, there is an even simpler way of making this point. All but the most ardent
Humeans accept the claim that if you aim at A, you have reason to take steps toward
realizing A. This principle—which I labeled Success in the prior chapter—is all that the
FIFTH OBJECTION 61
constitutivist needs in order to generate normative results. So long as we accept

Success, the mere fact that we have an aim does entail that we have reasons to fulfill it.
5. Fifth objection: constitutivism cannot account

for bad action
Let me summarize the argument so far. Constitutivism relies on the claim that aims
generate reasons. The constitutive aim differs from other aims merely in that it is
inescapable: so long as you perform actions, you will manifest the constitutive aim, and
you will thereby be subject to the reasons generated by this aim. Constitutivism
therefore has the requisite structure for generating universal reasons. The various
objections to this idea—that it relies on the idea that we have a reason to be agents,
or assumes that we cannot be schmagents, or reduces normativity to compulsion, or
derives an “ought” from an “is”—have been shown to fail. Constitutivism retains its
promise.
However, there is a remaining difficulty for constitutivism. Constitutivism may
encounter trouble in attempting to account for a very simple fact: agents often perform
bad actions, actions that are contrary to the balance of reasons. Below, I explain and
address this problem.
5.1 Can constitutivism account for bad action?
At its simplest, the constitutivist project could be seen as consisting of two claims:
(i) Action has constitutive feature F. (F might be autonomy, self-understanding,
self-constitution, etc.)
(ii) F is the constitutive standard of action. That is, an action is good qua action
insofar as it manifests F.
Peter Railton (1997) and Philip Clark (2001) have raised an objection to the constitutivist
strategy. The criticism is straightforward: (i) claims that every action has constitutive
feature F; (ii) claims that F is the standard of success for action; but, if every action
manifests feature F, then every action is a success. In other words, (i) and (ii) seem to entail
that the constitutive standard is achieved in each instance of action. If no intentional
action can fail to manifest the constitutive features of action, then no intentional action
can be defective as an action.11
11
Clark writes, “One noteworthy implication of Velleman’s position, then, is that no fully intentional
action can fail to achieve the constitutive aim of action. The constitutive aim is autonomy, but as we’ve just
seen, it is a premise of the argument that every fully intentional action is autonomous . . . This makes it
difficult to see how a fully intentional action could ever be rationally criticizable, on Velleman’s view” (2001,
582). He continues, “Consequently, Velleman must deny that any fully intentional action is contrary to the
weight of reasons” (2001, 583). This is a version of a familiar problem for roughly Kantian accounts of
practical reason: if Kant is right, then no action can be both fully autonomous and contrary to the demands of
practical reason.
Constitutivists derive standards from an account of action’s constitutive features.

Our problem is that there is no distance between the standard’s being applicable and
the standard’s being met. Call this the bad action problem: it seems that there is no such
thing as a bad action.
Although this appears to be a terrible problem, the constitutivist actually has a
straightforward way of answering it. In order to see how the constitutivist can answer
this objection, let’s distinguish two possible views:
(a) Aiming at A is what constitutes an event as an action.
(b) Achieving A is what constitutes an event as an action.
The first view is better. Compare belief. An attitude doesn’t have to achieve truth to
count as a belief; it just has to aim at (or be regulated for) truth. So we should accept (a)
and reject (b). A similar solution is available for action. We should say that aiming at
A is what constitutes something as an action.
But then how are we to account for defective actions and beliefs? If all that
something has to do in order to be a belief is to aim at truth, then how does the nature
of belief yield a standard of correctness? Well, the constitutivist just needs to make an
important distinction: what constitutes an attitude as a belief is aiming at truth. But
what constitutes an attitude as a good belief is achieving that aim, by being true.12
Analogously, in the case of action:
– What constitutes something as an action: aiming at A.
– What constitutes something as an exemplary or good action: achieving A.
Once we draw the distinction between what is required to be an action, and what is
required to fulfill action’s standard of success, we see that the bad action objection can
be answered. The constitutivist argues that every instance of action shares a common
aim. But many actions fail to fulfill this aim. These actions are still actions; but they are
defective actions.
Alternatively, the constitutivist can make the criterion for an event’s being an action
somewhat more robust:
– What constitutes something as an action: aiming at and to some extent
achieving A.
– What constitutes something as a good action: achieving A completely.
For example, on David Velleman’s view, an event has to be known to some extent in
order to qualify as an action. But, in order to qualify as a good action, it has to be
known more completely.
12
Here I am relying on a straightforward externalist epistemology. Internalists would have a more
complex standard.
FIFTH OBJECTION 63
So the constitutivist can answer the bad action problem, at least in principle. The
constitutivist simply relies on one of two distinctions. Either she distinguishes having an
aim and fulfilling this aim, or she distinguishes different degrees of fulfilling the aim.13
5.2 The “why bother?” objection
I have just suggested that in order to account for bad action, the constitutivist needs to
admit a distinction between the criterion for an event’s being an action, and the
criterion for an event’s being a good action. But this move opens the constitutivist to
one final objection, an objection that I believe points to a much more serious difficulty
than any mentioned so far.
I will call this the “why bother?” objection. It goes as follows. Suppose that aiming at
A is constitutive of acting. It is, however, true that I might aim at A while disavowing
or regretting my aiming at A.14 Or I might aim at fulfilling A only to a minimal degree.
In either case, the fact that I am aiming at A will constitute my behavior as an action. So
why bother to go the extra step, and make it a good action? That is, why not aim at
what is minimally constitutive of action, rather than what is constitutive of good or
excellent action?
An example may help to illustrate this point. In Self-Constitution, Korsgaard develops
an example of an agent who is engaged in the activity of building a house. She claims
that house-building has certain constitutive standards, which are determined by the
function of a house. She writes, “the function of a house is to serve as a habitable
shelter” (Korsgaard 2009, 27). In order to produce a habitable shelter, there are certain
things that one must do: one must make sure that the walls are weatherproof, that the
roof doesn’t leak, that the house isn’t going to fall over when subjected to a gust of
wind, and so on. If you don’t do these things, you will not have produced a habitable
shelter—a house—at all. Thus, Korsgaard claims,
If you fall too far short of the constitutive standard [for building a house], what you produce will
simply not be a house. In effect this means that even the most venal and shoddy builder must try
to build a good house, for the simple reason that there is no other way to try to build a house.
Building a good house and building a house are not different activities: for both are activities
guided by the teleological norms implicit in the idea of a house. Obviously, it doesn’t follow that
every house is a good house, although there is a puzzle about why not. It does, however, follow
that building bad houses is not a different activity from building good ones. It is the same activity,
badly done. (Korsgaard 2009, 29)
13
Note that my claim is not that every version of constitutivism will be able to overcome the bad action
problem. In order to overcome it, the theory needs to yield an aim with the above structure. In Chapter 4,
however, I will argue that Korsgaard’s constitutivist theory actually fails to yield a constitutive aim with this
structure.
14
Enoch considers something like this objection, writing, “Perhaps . . . I cannot opt out of the game of
agency, but I can certainly play it half-heartedly, indeed under protest” (2006, 188).
I think we should agree with Korsgaard that, in order to count as a house, an object
must meet certain standards. For example, it must provide protection from the
elements, it must have walls and a roof, and so on. And I think we should agree that
in order to count as building a house, an agent must try to build a structure that meets
those standards. So Korsgaard is correct in claiming that even the most venal and
shoddy builder must aim at meeting the standards that are minimally constitutive of a
house. For example, if you are not aiming at building something that has walls, is
somewhat weatherproof, and so on, then you are not aiming at building a house at all.
However, there is a great deal of distance between the standards minimally consti-
tutive of househood and the standards of an excellent house. There is a great deal of
distance between a plywood shack in the forest that will dissolve back into the
environment in a few years, and a sturdy stone house that will last for centuries.
Both of these count as houses, but it seems natural to say that the stone house is a
better house. We might describe this by saying that the same standards apply to the stone
house and the shack: the standards of househood. But there are better and worse ways
of fulfilling these standards. At some point, a structure will have met the standards to a
sufficient degree to qualify as a house, but will not have met the standards to a sufficient
degree to qualify as an excellent house.
This is, of course, a scenario that arises in everyday life: a person who considers
buying a house wants to make sure not just that the structure she is examining meets
the minimal requirements of househood, but that it will be sturdy, long-lasting, and
well-constructed. There are, after all, many houses that turn out to be slapped together
out of shoddy materials, that rest on cracked foundations, that leak, and so forth. In
short, there are houses that will provide habitable shelter for several years, but will
encounter a host of problems. And it is obvious that there are builders who deliberately
aim to produce these sorts of houses: in the interest of maximizing profits, it sometimes
pays to produce a merely passable house rather than an excellent house. It would be the
height of absurdity to claim that these builders aren’t really aiming to build houses.
They are aiming to build houses, just mediocre ones. (This point is perhaps even clearer
in the case of other artifacts, such as furniture. Contrast the cheaply made, particle
board bric-a-brac that one finds at discount mega-stores with the sturdier, solid wood,
longer lasting furniture that one finds at more expensive stores. The people who
produce the particle board desk know that it will look battered and beaten in a few
years, that it will not withstand repeated moves, and so forth. They know that in this
sense, they are producing mediocre furniture. Still, they are producing furniture.)
This creates a problem. Presumably an agent could aim to meet the standards
minimally constitutive of househood without aiming to meet the standards to a high
enough degree to qualify as building an excellent house. (It is obvious that an agent could
try to build an excellent house, but end up producing only a passable house. This is not
the case that we are considering. We are considering a case in which an agent
deliberately tries to build a passable-but-not-excellent house.)
FIFTH OBJECTION 65
So Korsgaard is right that a builder must aim at building a good house, if by “good”
she means passable. That is, she is right that a builder must aim at building a good house,
if by “good” she means a house that is at or above the cut-off point for househood. But it is, of
course, false that a builder must aim at building an excellent house, a house well above
the cut-off point for househood.
This is a perfectly general point about constitutive aims. Suppose we grant that there
are activities that have aims that must be fulfilled to some degree, lest one fail to
participate in the activity. We can still imagine someone asking, “Why bother doing
more than the bare minimum? Why not aim to hit the cut-off point, and nothing
more? Or, if there’s no precisely defined cut-off point, why not just do a good enough
job, rather than an excellent job?”
5.3 Answering the “why bother?” objection
The “why bother?” objection grants that there are activities that have aims that must be
fulfilled to some degree, lest one fail to participate in the activity. But, quite simply, the
objector says, “Why bother doing more than the bare minimum? Why not strive to
fulfill the constitutive aim to a minimal degree, and nothing more?”
While I think this objection is important, there are several ways in which we might
try to answer it. First, one could accept the objection and show that the minimal
standards still generate results. Second, one could attempt to show that certain aims
have a peculiar structure: in certain cases, an agent cannot aim at less than full
realization of the aim without ceasing to have the aim at all. Third, one could argue
that in order for the aim to generate sufficiently robust claims about reason, the
standard induced by the aim must be ratified by the agent. I will explain these responses
in turn.
Start with the first response. Suppose the constitutivist agrees that if an activity has
constitutive aim, this need not imply that an agent who participates in the activity must
strive to fulfill the aim to the highest degree possible. Even if the constitutivist grants
this point, the theory might still establish substantive conclusions about practical reason.
Consider again the case of house-building. Suppose we agree that in order to be
engaged in house-building, an agent merely needs to aim at fulfilling the standards of
house-building to some unspecified degree. Still, this generates a number of important
results. For example, if the agent asks, “Why should I make this structure have a roof?”
we will be able to respond, “If you are committed to house-building, you are
committed to at least that much. Houses are structures with roofs, so if you are
going to build a house then you are going to build a roof.” Analogously, suppose
that it is constitutive of action that agents aim at some standard such as acting
autonomously. Even if agents are only committed to acting autonomously to some
degree, this fact might generate substantive results.15 The details will, of course, depend
15
The constitutivist who pursues this line of reasoning might want to argue that fulfilling the constitutive
standards of action to the highest possible degree is supererogatory. By analogy, the excellent house, with
on the nature of the aim that the constitutivist employs. I will be investigating this
possibility in the next two chapters.
Turn now to the second possible response. The constitutivist could argue that
certain aims have a peculiar feature: the agent cannot aim at less than full realization
without ceasing to have the aim at all. This amounts to a Kantian strategy for answering
the objection, and it is the one that Korsgaard pursues. To see how it works, take the
aim of acting justly. One can fulfill this aim to different degrees, in the sense that actions
can be more and less just. But consider an agent who deliberately and self-consciously
aims to be only a little just. This seems incoherent: if you aim to be only a little
just, then you aren’t aiming to be just at all.16 After all, to be just-a-little-but-not-
completely just, one would have to be unjust in certain circumstances. But to aim at
being unjust in certain circumstances is precisely not to aim at being just.
Other aims might have an analogous structure. For example, consider the aim of
being faithful in a romantic relationship. It is not as if you can aim at being just somewhat
faithful, faithful unless someone especially attractive comes along. For then you
wouldn’t be aiming at faithfulness at all. Or, to choose an example with more relevance
to the constitutivist, consider the aim of autonomy: although an agent can be more and
less autonomous, there is something paradoxical about the idea of aiming to be just
slightly autonomous. After all, if an agent had that aim, it is hard to see how she would
be aiming at something other than the opposite of autonomy, heteronomy.17
However, many aims lack this feature. For example, take the aim of knowledge.
There is nothing problematic about aiming to achieve merely some knowledge. I might
aim to know a bit about physics, without aiming to know everything about physics. Or
I might aim to know a bit about myself and my motivations, without aiming to know
everything about myself and my motivations. (We will return to this point in the next
chapter, when we discuss Velleman’s claim that action constitutively aims at self-
knowledge.)
As these examples illustrate, some—but not all—aims are such that one cannot aim
at less than full realization without ceasing to have the aim at all. Aims with this feature
will not be susceptible to the “why bother?” objection. This point will become
important in the following chapters.
Finally, let’s consider the third possible response: one could grant the “why bother?”
objection, and argue that it shows that the constitutive aim is only going to yield the
right results if the agent wholeheartedly accepts the aim, and makes it her own. I won’t
say much about this possibility here, for I will explore it in Chapter 7. There, I discuss a
dispute between Schopenhauer and Nietzsche, and argue that it illustrates the problems
solid walls that will last for two hundred years, is—so to speak—a supererogatory house. It goes above and
beyond what is required. The slapped-together house, with a leaky roof, drafty walls, and joints that begin to
collapse as soon as they are constructed, is nonetheless a house, and it is all that we are really committed to
creating.
16
Thanks to Chris Korsgaard for suggesting this example.
17
Korsgaard makes an analogous point in her example of Emma and Harriet (Korsgaard 2009, 162–3).
CONCLUSION 67
and prospects for this third line of response. Schopenhauer and Nietzsche agree that
there is an aim present in all episodes of action. For Schopenhauer, this is the “will to
live,” whereas for Nietzsche it is the “will to power.” Schopenhauer argues that this
inescapable aim should be repudiated—we should renounce the will to live, becoming
so indifferent to our own motives that we wither away of voluntary starvation
(Schopenhauer was not one to shy away from a pessimistic conclusion). Nietzsche,
on the other hand, argues that the inescapable aim (will to power) should be affirmed
and treated as reason-providing. What is interesting, here, is that Nietzsche may think
that in order for the will to power to generate the right kind of normative results, the
will to power must be affirmed. As I will be arguing that the will to power is a
constitutive aim of action, this amounts to the claim that the constitutive aim only
yields the right results if the agent wholeheartedly affirms the aim.18
So, to mark a third possibility: the lesson of the “why bother?” objection may be that
we need an additional step, an act of affirmation of the constitutive aim, and that
without that we will not achieve the correct normative results. This is the position
that I will be arguing for in the following chapters.
I have sketched three ways in which the constitutivist can address the “why bother?”
objection. Whether these responses ultimately succeed depends on the details of the
particular constitutive feature that the theory posits. Accordingly, I will be investigating
the success of these responses as we examine particular constitutivist theories in more
detail.
6. Conclusion
This chapter has examined a range of objections to constitutivism. I have argued that
the objections fail. For the most part, they fail because they do not appreciate the
implications of the constitutivist’s claim that normativity is grounded in inescapable
features of agency. This, I have suggested, is the core idea of constitutivism: the
authority of universal normative claims arises from their inescapability.
Of course, so far I have argued only that if we can show that action has a constitutive
aim, then constitutivism would succeed. The difficult part—showing that action really
does have a constitutive aim—still confronts us. In the next two chapters, I will
consider two insightful attempts to establish that action has constitutive features.
David Velleman has argued that action has the constitutive aim of self-understanding,
whereas Christine Korsgaard has argued that action has a constitutive standard of self-
constitution. I will examine these theories in turn, asking whether they succeed.
18
Korsgaard inclines toward a similar view: she sometimes suggests that the agent must endorse the
constitutive features of agency. See, for example, Korsgaard (1996b, 102–9) and (2009, 197).
3
Constitutivism and Self-Knowledge
The previous chapters argued that if we could show that action has a constitutive aim,
then we might be able to solve the foundational problem in ethics. In particular, we
could derive normative conclusions from that constitutive aim, and thereby show how
claims about universal reasons for action might be justified.
While this strategy seems promising, it faces a glaring difficulty: isn’t it implausible to
suggest that action has a constitutive aim? Our intentional actions have the most diverse
range of aims, encompass the most multifarious forms of activity, and take place in the
most dissimilar contexts. Could it really be true that the shopper eyeing the wares along
Park Avenue, the diplomat flying to Australia, the convict plotting his escape, the
dancer spinning in a pirouette, the boy wearily brushing his teeth, and the mathemat-
ician laboring over his proof all share a common aim?
If constitutivism is to succeed, then we will have to show that these agents do, in
fact, share a common aim. In this chapter, I will examine David Velleman’s attempt to
show that all agents share the common aim of self-knowledge. The next chapter will
investigate Christine Korsgaard’s claim that all agents aim at self-constitution.
The structure of the present chapter is as follows. Section 1 reconstructs Velleman’s
argument for the claim that all actions share a common aim of self-knowledge.
Section 2 assesses Velleman’s attempt to derive an account of reasons from this consti-
tutive aim. I argue that Velleman’s account fails on this score: even if we grant his claim
that all actions aim at self-knowledge, no substantive conclusions about reasons for
action would follow. Sections 3 and 4 consider some ways in which Velleman might
respond to this challenge. Finding these potential responses unsuccessful, I conclude,
in Section 5, that Velleman’s version of constitutivism encounters insurmountable
difficulties.
1. Velleman’s argument that action constitutively aims

at self-knowledge
Velleman’s claim that all actions aim at self-knowledge is, on the face of things,
wildly implausible. The shopper is aiming at finding a nice pair of shoes, not at self-
knowledge; the boy brushing his teeth is aiming at preventing cavities, not at knowing
anything about himself.
V E L L E M A N ’ S A R G U M E N T T H AT A C T I O N C O N S T I T U T I V E LY A I M S 69
However, Velleman’s claim becomes more plausible when we consider a peculiar

feature of intentional action: on many accounts, there is a necessary connection
between acting intentionally and possessing a certain kind of self-knowledge. To be
precise, many philosophers accept the following claim:
(1) If an agent intentionally A’s, then the agent knows that she is A-ing.
To illustrate this point, consider a case in which the consequent is false. Suppose the
person sitting beside you begins tapping his foot, creating an irritating noise. Bothered
by this, you ask him to stop. Startled, he replies, “Sorry, I didn’t even realize I was
doing that.” We would typically take this as an indication that his foot-tapping was not
an intentional action.
While claim (1) is not entirely uncontroversial, it is widely accepted in the literature
on action.1 But let’s move from (1) to a more controversial claim. Some philosophers
believe that we can refine (1) by offering a more precise characterization of the type of
knowledge that intentional action requires. An agent acting intentionally knows what
she is doing, but this knowledge has a distinctive feature: it seems to be immediate. It is
easiest to illustrate this point by contrasting two cases: if I want to know what my
neighbor is doing, I must observe her, attend to her movements, postulate motives, and
so on. But I do not stand in the same epistemic relation to my own intentional actions.
I seem not to need to reflect on or attend to my intentional actions in order to
determine what I am doing. Let’s put the point this way:
(10 ) If an agent intentionally A’s, then

(a) the agent knows that she is A-ing, and
(b) the agent’s knowledge is immediate.
For now, the term “immediate” is just a placeholder; I haven’t yet said what it means. If
we are to understand (10 ), we need to explain the sense in which the knowledge is
immediate. Moreover, we will want an explanation of how this knowledge arises.
Thus begins Velleman’s project: he accepts (10 ), and wants to offer a determinate
characterization of “immediate” knowledge, together with an explanation of its
origins. In order to do so, he postulates the existence of a particular desire: a desire to
know what one is doing. Velleman calls this the desire for self-knowledge.
Suppose an agent harbors this desire. Then the agent will be averse to acting in ways
that she doesn’t understand, and will be motivated to act in ways that she does
1
There are some variations in the details. For example, some philosophers prefer to substitute “believes”
for “knows” in (1). Moreover, some philosophers argue that we should restate (1) in a weaker fashion: if an
agent intentionally A’s, then the agent knows he is A-ing under some description. (For example, the agent might
know that he is walking to his office, but not know that he is walking along Main Street, although on some
views these are descriptions of the same action.) These complications will not be relevant here, so I do not
discuss them. The philosophers endorsing some variant of claim (1) include Davidson (1980, 50), O’Shaugh-
nessy (1980), Searle (1983, 90), Ginet (1990, 87), Anscombe (2000, 13–15), Falvey (2000), Wilson (2002),
and Bratman (2007, Chapter 9). Of course, there are also those who reject all versions of (1); a prominent
example is Arpaly (2004).
70 C O N S T I T U T I V I S M A N D S E L F - K N OW L E D G E
understand. For example, I know that I have a desire for a cup of coffee; so, were I to
act on that desire by walking to the coffee shop, I would know what I was doing. My
desire for self-knowledge therefore inclines me to perform that action. On the other
hand, as far as I can tell, I have no desire that counts in favor of tossing my computer out
the window. Were I to do that, I would be acting in a way that I did not understand. So
my desire for self-knowledge makes me averse to performing that action.2
This aspect of Velleman’s theory is initially counterintuitive. We are inclined to
think that there is only one way to fulfill a desire to have true beliefs about one’s own
behavior: first you observe your behavior, and then you form beliefs in light of this
observation. But Velleman’s point is that there is also a second way to fulfill this desire:
rather than modifying my beliefs in light of my behavior, I can modify my behavior in
light of my beliefs.
This is clearest in the case of habitual actions: if I believe that I always jump over
puddles, then one way in which I can confirm this belief and secure a bit of self-
knowledge is by jumping over the next puddle that I see. If I believe that I wear
sweaters whenever it is cold, then one way that I can confirm this belief and fulfill my
desire for self-knowledge is by wearing a sweater the next time it is cold.
Velleman’s central claim is that this mechanism may be far more pervasive than we
are inclined to think. It needn’t be limited to cases of habitual action, but can operate
more widely. For example, whenever I make a decision, I will expect to act in
accordance with that decision. My desire for self-knowledge will provide some
motivation for conforming to that decision: if I act otherwise, I will violate my
antecedent belief that I will act as I have decided to act.
Despite its initially counterintuitive feel, then, Velleman concludes that we have
good reason to accept the above picture of action. In short: we sometimes fulfill our
desire for self-knowledge by modifying our behavior in light of our antecedent beliefs,
rather than motivating our beliefs in light of our behavior.3
Now let’s add a twist. Suppose, as often happens, I have conflicting motives: some of
my desires incline me to continue writing this chapter, while others incline me to go
out for coffee. I would of course understand what I was doing, either way. But suppose
I form the expectation that I will continue writing this chapter. Then my desire for self-
knowledge will incline me to continue writing; after all, if I find myself going out for
coffee, my expectation will have been flouted.4
2
Of course, I also have many other desires that bear on these two actions, so the desire for self-knowledge
is not going to determine, all on its own, which action I am inclined toward. It is simply one factor amongst
many others.
3
Velleman bolsters his case by surveying a range of empirical psychology that supports this claim. See
Velleman (2006, Chapter 10).
4
Let me emphasize again that this is not a description of how we ordinarily conceive of our actions. It’s
obvious that the above thoughts do not run through the heads of ordinary agents as they act. Nevertheless,
Velleman argues that there are good reasons for taking this to be an accurate explanation of action’s etiology.
For empirical evidence, see the work cited in the prior note.
V E L L E M A N ’ S A R G U M E N T T H AT A C T I O N C O N S T I T U T I V E LY A I M S 71
So, once I form an expectation about what I will do, my desire for self-knowledge
will incline me to act in a way that fulfills the expectation. For this reason, my
expectations will tend to be self-fulfilling: in some cases, I will be able to make it the
case that I will A simply by forming the expectation that I will A.5
Note that if intentional actions were always caused by self-fulfilling expectations,
then an agent would always possess a distinctive kind of knowledge6 of his own
intentional actions: as Velleman puts it, “the agent attains contemporaneous know-
ledge of his actions by attaining anticipatory knowledge of them” (2004b, 277; cf.
2009, 18–20). So Velleman invites us to entertain the hypothesis that intentional
actions are behaviors caused by self-fulfilling expectations. This hypothesis recom-
mends itself in two ways. First, it offers a determinate characterization of the vague idea
that intentional actions involve immediate knowledge (i.e., the knowledge in question
is knowledge resulting from self-fulfilling beliefs about one’s forthcoming actions).
Second, it explains why actions always involve this distinctive kind of knowledge (i.e.,
intentional actions just are behaviors caused by self-fulfilling beliefs).
Thus, Velleman’s proposal takes the following form: if agents were constituted so
that their behaviors were sometimes caused by self-fulfilling beliefs, then these agents
would sometimes possess a distinctive kind of knowledge of their own behaviors. The
behaviors involving this kind of knowledge would be labeled intentional actions.
Notice that Velleman’s proposal has an interesting implication: it entails that inten-
tional action has a constitutive aim. For, on the analysis just offered, every intentional
action does share a particular, higher-order aim: the aim of knowing what one is doing.
Moreover, it is in virtue of having this aim that a stretch of behavior qualifies as
an intentional action. But, by the definition of Constitutive Aim, this is just to say
that intentional action constitutively aims at self-knowledge.7 So, if we accept Velle-
man’s account of the nature of intentional action, it turns out that intentional action has
the constitutive aim of self-knowledge.
5
Notice that it does not matter why I form the expectation that I will A. I might carefully survey all of my
motives, imagine the consequences of potential actions, and, after much scrutiny, form an expectation that
I will go to the coffee shop. Or I might simply notice that I have a desire for coffee, and immediately form the
expectation that I will go get some. Or, in the limiting case, I might form an expectation for no reason at all;
I might just arbitrarily form the expectation that I will right now lift up my hand. None of this matters: the
expectation will tend to play the same causal role regardless of the way in which it is formed. (Of course,
expectations can be outweighed by other motives. For example, if I form the expectation that I will jump off
the cliff, I will have some inclination to do so; nevertheless, the inclination is probably going to be greatly
outweighed by other motives, such as fear, the desire for self-preservation, and so forth.)
6
It may be a bit hasty to call these self-fulfilling expectations “knowledge.” After all, they are not justified
by anything (except, perhaps, the fact that there is a reliable connection between their formation and their
fulfillment). Although I think this concern is valid, nothing in this chapter depends on whether these
expectations constitute knowledge. So I do not address this difficulty. For a critical discussion of this point,
see Langton (2004).
7
Chapter One defined the notion of constitutive aim as follows: Let A be a type of attitude or event. Let
G be a goal. Then A constitutively aims at G iff (i) each token of A aims at G, and (ii) aiming at G is part of
what constitutes an attitude or event as a token of A.
Of course, agents rarely have a conscious aim of making their actions conform to
their expectations. If Velleman were claiming that whenever an agent acts, she expli-
citly thinks, “I expect to do such and such, so I will do such and such,” then his theory
would be phenomenologically absurd. The appearance of absurdity is dispelled if we
take account of an important point: the process that Velleman describes is not supposed
to take place in a fully reflective fashion. Rather, the aim of self-knowledge is supposed
to function in the background of deliberation and action. Consider, by analogy, the
way in which the aim of avoiding pain figures in actions such as taking a jog. A jogger
will make subtle shifts in the placement of her limbs, the arc of her toes, the angle of her
heels, as she runs, in response to feedback from nerve receptors. In the normal case the
jogger won’t be coordinating her movements with the explicit aim of avoiding pain; in
fact, it may not even occur to her that she has such an aim. Nevertheless, the aim is
present, influencing the way in which the action proceeds. Velleman argues that the
aim of self-knowledge could function in an analogous fashion (Velleman 2000, 21, and
2009, 27–8).
We could, of course, question this model of action. But let’s grant it for the sake of
argument. What would follow from the fact that action has a constitutive aim of self-
knowledge?
2. Velleman’s attempt to derive an account of reasons

from the constitutive aim
So far, we have examined Velleman’s attempt to establish that action has a constitutive
aim. Suppose this argument succeeds. Still, there is work to be done: establishing that
action has a constitutive aim is only the first step in a constitutivist argument. The next
step is to show that the constitutive aim generates a substantive account of reasons for
action. This second step is crucial: after all, showing that action has a constitutive aim
will be of limited interest if this constitutive aim does not yield substantive normative
conclusions. Thus, in the remainder of this chapter I will grant Velleman the above
account of action, and ask whether it suffices to generate an account of practical reason
(that is, an account of reasons for action). I argue that it does not. Even if Velleman
were right about the nature of action, his account of practical reason would not follow.
2.1 A criterion of adequacy for accounts of practical reason
Let’s begin by asking what we want from a theory of practical reason. First and most
obviously, a theory of practical reason must generate claims of the form “there is reason
to perform action A.” Second, a theory of practical reason must do more than simply
tell us which actions there is reason to perform. It must also give us a way of ranking or
weighing reasons. Stepping back from the details of constitutivism for a moment,
consider a perfectly ordinary case in which an agent is trying to decide what to do.
Suppose Sarah can either study for her exam or go to a friend’s party. Sarah has some
V E L L E M A N ’ S AT T E M P T T O D E R I V E A N A C C O U N T O F R E A S O N S 73
reason to perform each action: studying for the exam would be conducive to her goal
of earning a good grade, whereas going to the party would further her goal of having
an enjoyable evening. We expect a theory of practical reason to give us some guidance
in cases of this sort, by providing us with a way of weighing the reasons for and against
the potential actions.
It is important to be clear here. My claim is not that for any two actions, an adequate
theory of practical reason must entail that there is more reason to perform one of them.
My claim is much weaker:
(More Reason) For some actions A and B, an adequate theory of practical reason
must entail that there is more reason to A than to B.
The above principle states that an adequate theory of practical reason must be capable
of generating conclusions about what there is more reason to do.8 I take it that this is an
uncontroversial criterion of adequacy for a theory of reasons for action.
So we should ask whether a constitutivist theory enables us to make judgments of
the form “there is more reason to A than to B.” I will argue that Velleman’s theory
does not enable us to make judgments of that form. Moreover, I will argue that there is
no plausible way of modifying Velleman’s theory in order to make it produce
judgments of that form. I conclude that Velleman’s theory is unsuccessful.
2.2 Velleman’s derivation of reasons from the constitutive aim
Let’s grant Velleman the claim that the constitutive aim of intentional action is self-
knowledge, or knowing what you are doing. In other words, if you don’t aim at
knowing what you are doing, then you are not acting intentionally. Let’s further grant
that this constitutive aim provides a criterion of success: if you know what you are
doing, then you have performed a successful intentional action. The standard of success
can be used to generate claims about what there is reason to do: you have reason to
perform those actions that would fulfill the standard of success. In Velleman’s case,
this entails that you have reason to perform those actions that would generate self-
knowledge. For example, suppose I notice that I have a desire for some ice cream.
Suppose I were to act on the consideration that I desire ice cream, by purchasing ice
cream. The consideration that I desire ice cream would provide me with a ready explan-
ation of my action: I bought the ice cream because I desired ice cream. 9 So I would
8
When a theory generates conclusions of this form, I will say that it “ranks” or “weighs” reasons. To
avoid confusion, it is important to note that the phrase “ranking of reasons” is sometimes used differently, to
mean that considerations of type A (e.g., moral considerations) outrank or outweigh considerations of type
B (e.g., prudential considerations). I am not using the term in this way; I say that a theory ranks or weighs
reasons when it fulfills the More Reason condition, above.
9
Velleman writes, “The considerations that qualify as reasons for doing something are considerations in
light of which, in doing it, the subject would know what he was doing. They are, more colloquially,
considerations in light of which the action would make sense to the agent. . . . What makes sense to someone
is, theoretically speaking, what he can explain. This is what I mean when I say that reasons for doing
something are considerations in light of which it would make sense. I mean that they are considerations that
certainly know what I was doing, were I to act on this desire. Thus, according to
Velleman’s theory I have some reason to perform this action.
So Velleman’s theory can provide us with an account of reasons for action. But can it
provide us with a way of ranking reasons? It seems so. Self-knowledge comes in
degrees: we can know more or less about our own actions. For example, suppose
Betty decides to take a pre-med sequence in college. She might just have a vague sense
that she would like to be a doctor, without really being able to explain why. Or she
might have more substantial understanding of what makes this career appealing to her:
that she is a compassionate individual, that she enjoys aiding others, that she likes
personal interactions, and so on.
Given that self-knowledge comes in degrees, it seems that actions can fulfill the
constitutive aim to differing degrees: an action that provides a bit of localized and
superficial self-knowledge will fulfill the constitutive aim to some extent, but an action
that provides comprehensive and profound self-knowledge will fulfill the aim to a
higher degree, or fulfill it more completely.
If this is right, then we can generate substantive conclusions about practical reason.
Suppose I could act in one of two ways, both of which would provide me with some
degree of self-knowledge. Then I have reason to act in either way. But I have more
reason to act in the way that will provide me with more self-knowledge. As Velleman
puts it, “reasons will have to qualify as better or stronger in relation to the constitutive
aim of action, which lends reasons their normative force. Roughly speaking, the better
reason will be the one that provides the better rationale—the better potential grasp of
what we are doing” (Velleman 2000, 29).
As an illustration, Velleman offers the following example. Imagine an agent who has
a desire to destroy an inkstand on his desk; he has this desire because he believes that if
he destroys his current inkstand, then his sister will provide him with a new, superior
inkstand as a gift.10 The agent considers whether to act on his desire. If he destroys the
inkstand, he will have some self-knowledge: he will know that he is acting so as to
realize his desire for a new inkstand. However, he will also “realize that such an activity
[is] contrary to other motives of his, as well as to some of his customs, emotions, and
traits of character” (Velleman 2000, 29). For example, the activity would be rather
deceptive and underhanded, and would thus violate his character trait of trustworthi-
ness. Consequently,
even though he would subsequently have known that he was destroying the inkstand . . . he
might still have wondered “What am I doing?” That is, he might still have been puzzled as to
how a person like him, with a makeup like his, would come to act on such motives; and so he
would provide the subject with an explanatory grasp of the behavior for which they are reasons” (2000, 26).
“Considerations weight in favor of an action, I propose, insofar as they contribute to an overall understanding
of the action . . . ” (2009, 19).
10
This is a variant of a case discussed by Freud. See the Introduction to Velleman (2000) for a discussion of
the case.
wouldn’t really or fully have known what he was up to. This lack of self-knowledge would have
indicated to the agent that he would have had a better idea of what he was doing if he had chosen
to do something else instead. That is, he could have adopted, and consequently enacted, a
more intelligible story. And insofar as there was a more intelligible story for him to enact,
by choosing to do something else, there was a better rationale for doing that thing instead.
(Velleman 2000, 29)
This is a perfectly general point: whenever I am faced with two actions, one of
which would generate more self-knowledge, I have more reason to perform the action
that would generate more self-knowledge.
2.3 Two kinds of aims
So, at any rate, it seems. But I will argue that Velleman is not entitled to this
conclusion. His theory is incapable of generating conclusions about what we have
more reason to do. For the mere fact that action constitutively aims at self-knowledge
does not imply that we have more reason to act in ways that generate more self-
knowledge.
The key point is this: the claim that action has a constitutive aim is ambiguous; it
could mean two quite different things. For aims can be divided into two kinds. Some
aims can be fulfilled to different degrees by different actions. Other aims cannot be
fulfilled to different degrees; they can merely be fulfilled or unfulfilled.
Let’s introduce some terminology. A simple aim is an aim that a particular action will
either fulfill or fail to fulfill. A differentially realizable aim is an aim that can be fulfilled to
different degrees.
Consider a few examples. An agent might have the aim of buying a book, or going
to a restaurant, or owning a dog. These are simple aims: they can merely be fulfilled or
unfulfilled. For example, you either go to a restaurant or you don’t. On the other hand,
an agent might have the aim of earning money, or making a delicious cake, or staying
in a hotel near the beach. These are differentially realizable aims: they can be fulfilled to
different degrees. For example, the aim of staying in a hotel near the beach is fulfilled to
a greater degree by staying in a hotel a block from the beach than by staying in a hotel
half a mile from the beach.
With this in mind, notice that the claim that action constitutively aims at self-
knowledge is ambiguous: self-knowledge might be a simple aim, or it might be a
differentially realizable aim.
(Aim 1) Action has a simple aim of self-knowledge. That is, action fulfills the
constitutive aim when it generates self-knowledge.
(Aim 2) Action has a differentially realizable aim of self-knowledge. That is, an
action that yields superficial self-knowledge fulfills the aim to a lesser
degree than does an action that yields comprehensive self-knowledge.
These aims yield different conclusions about reasons for action. Suppose an agent has
two options, A-ing or B-ing. Suppose A-ing would result in the acquisition of more
self-knowledge than B-ing. Then, if action has constitutive Aim 2, the agent has more
reason to A than to B. If action has constitutive Aim 1, the agent has equal reason to
A or to B.
This point can be illustrated with an example. In the following passage, Velleman
considers a situation in which an agent has reasons both for and against an action:
One may have reasons for resisting desires that one understands, since one may be averse to the
desires themselves or to their potential manifestations. If a person knows that he is lazy, for
example, he will understand the lassitude that he feels at the start of any project; but if he also
knows that he wants a project to succeed, or that he wants in general to be more energetic, then
he may also understand refusing to let his lassitude prevail. Once again, if his motives for
counteracting his laziness are themselves intelligible to him, then he will stand to gain even
greater self-understanding from deciding and acting on an alternative preference. (1989, 248)
The agent is deciding between two actions: beginning the project (call this A-ing) and not
beginning the project (call this B-ing).11 The agent knows that he is lazy, and this laziness
would render B-ing comprehensible. According to Velleman, the laziness therefore
counts as a reason for B-ing. But other factors count in favor of A-ing: the agent knows
that he wants to be more energetic, and knows that he wants the project to succeed.
Velleman hypothesizes that because the agent disapproves of his own laziness, and
because the agent wants the project to succeed, the agent will gain “even greater self-
understanding” by A-ing.
In sum, Velleman is claiming that the agent has more reason to A than to B, because
A-ing would generate more self-knowledge than B-ing. However, notice that this
conclusion only follows if action has a differentially realizable constitutive aim. If action
had a differentially realizable aim of self-knowledge, then we would have more reason
to perform those actions that would generate more self-knowledge. But the conclusion
does not follow if action has a simple aim of self-knowledge. If action has a simple aim
of self-knowledge, then the agent has equal reason to A and to B.
Importantly, this is a perfectly general result. Consider any two actions A-ing and
B-ing, each of which would yield some self-knowledge. If action has a simple aim of
self-knowledge, then there will be no case in which the agent has more reason to
A than to B. Thus, in order to generate conclusions of the form “there is more reason
to A than to B,” Velleman will have to establish that action has a differentially realizable
aim of self-knowledge.
So which kind of aim does Velleman establish? In the following sections, I will show
that he only establishes a simple constitutive aim.
11
Some philosophers will claim that not beginning the project does not count as an action. This does not
affect the argument. Those philosophers can let B-ing be watching televison, or some such.
2.4 Velleman’s theory yields only a simple constitutive aim

I will argue that even if we grant Velleman his conception of action, it turns out that
action only has a simple constitutive aim of self-knowledge. There are two ways of
making this point. First, recall how Velleman’s argument functions: he begins by
assuming that intentional action requires the agent to have immediate knowledge of
what he is doing. He accounts for the presence of immediate self-knowledge by
positing a desire to know what one is doing, which adds motivational force to the
actions that the agent antecedently expects to perform. As a result, agents tend to
perform those actions that they expect to perform. If an agent acts in the way that she
expects to act, she will have immediate knowledge of what she is doing. Accordingly,
the postulated desire explains the presence of immediate self-knowledge in intentional
action.
Notice, though, that the desire simply adds motivational force to the agent’s
expectations about what she will do. There is no sense in which the desire favors
more comprehensive descriptions of an agent’s action. The desire’s functioning is
entirely dependent upon the nature of the agent’s expectations: if her expectations
are couched in comprehensive and integrative terms, then the desire will prompt her to
act in ways that realize comprehensive and integrative self-knowledge; if her expect-
ations are couched in superficial and proximate terms, then the desire will prompt her
to act in ways that realize superficial and proximate self-knowledge. In short, from the
fact that agents aim at knowing what they are doing, it does not follow that agents aim
at having comprehensive knowledge of what they are doing.
There is also a second and perhaps more intuitive way of putting the problem.
Velleman begins with the idea that intentional actions are immediately known. In
order to explain how this immediate knowledge arises, Velleman hypothesizes that
actions are causally regulated with an aim of self-knowledge. That is, actions are
screened in advance for conformity with an agent’s antecedent expectation about
how he is going to act. This screening mechanism inclines the agent to act in ways
that he antecedently expects to act, so that the agent “attains contemporaneous
knowledge of his actions by attaining anticipatory knowledge of them” (2004b, 277).
With this in mind, here is another way of putting my objection. In order to
construct an agent whose actions are immediately known, we would not need to
equip the agent with a differentially realizable aim of self-knowledge. All that we
would need to do is to equip the agent with the aim of fulfilling its antecedent
expectations. That is, all that we would need to do is equip the agent with a simple
aim of self-knowledge. So it is conceptually possible that there are agents who perform
immediately known actions, but have only a simple aim of self-knowledge. Thus, from
the bare fact that actions are immediately known, all that we can derive is a simple
constitutive aim of self-knowledge.12 As a result, Velleman’s theory of action does not

yield a substantive account of practical reason.13
In sum, even if we grant Velleman his account of intentional action, he fails to
generate substantive results about practical reason. That is, if an agent is faced with the
choice of whether to A or B, where A-ing and B-ing are actions that could be done
self-knowingly, there will be no case in which Velleman’s account entails that the
agent has more reason to A than to B.
In fact, the problem is worse still. Recall that on Velleman’s theory, every action
generates some degree of self-knowledge, because being immediately known is a
necessary condition for an event’s being an action (see section 2.1). Thus, on
Velleman’s theory, the claim “an agent has reason to perform any action that can be
done self-knowingly” is equivalent to the claim “an agent has reason to perform any
action.” Consequently, Velleman’s theory generates the absurd conclusion that an
agent has no more reason to perform any one action than any other action, and has
equal reason to perform every action.
12
For an illustration of this point, consider Anscombe’s (2000) theory of action. Anscombe claims that in
order for a behavior to qualify as an intentional action, it must be immediately known. However, there is no
sense in which more comprehensively known actions are better than less comprehensively known actions. An
action that is known only under a superficial description is on par with an action that is known under a deep
and comprehensive description. So Anscombe’s theory could be read as claiming that action has a simple
constitutive aim of self-knowledge, but lacks a differentially realizable constitutive aim of self-knowledge.
13
Kieran Setiya (2007) reaches a related conclusion in a different way. Setiya argues that while it is true
that action constitutively aims at self-knowledge, it is false that we have more reason to do what generates
more self-knowledge (cf. Setiya 2007, 112). Setiya’s argument takes the following form. He begins by
rejecting Velleman’s claim that actions are brought about by self-fulfilling expectations, arguing instead that
action is brought about by intentions. According to Setiya, an intention is a desire-like, self-referential belief
about what the agent is doing and why: to intend to A for a reason p is to have a desire-like belief with the
content I am hereby A-ing both because p and because I am in this very state (Setiya 2007, 49). When I act
intentionally, I aim to make this belief true (this is the sense in which the belief is “desire-like”: it has a world-
to-mind direction of fit). If I succeed in executing the intention, I thereby acquire a true belief about what
I am doing and why (Setiya 2007, 49). This account of intention implies that action has a constitutive aim:
whenever I act intentionally, I aim to make true the content of my antecedent belief about what I am doing.
In other words, whenever I act intentionally I aim to make true a belief about what I am doing and why; if
I successfully execute the intention, then I will have a true belief about what I am doing and why. In this
sense, Setiya’s account generates a constitutive aim of self-knowledge. However, notice that this constitutive
aim doesn’t generate any tendency for us to seek more comprehensive knowledge of our actions. Rather,
the constitutive aim is fulfilled so long as one acts as one intends to act; contrary to Velleman’s view, the
constitutive aim doesn’t impose any restrictions on the content of the intention, and in particular the
constitutive aim is not better fulfilled by contents that generate more self-knowledge (cf. Setiya 2007,
112). The constitutive aim just tells us to fulfill our intention, whatever it may happen to be. In other
words, we simply have reason to do whatever we intend to do. So, to put the point into my terminology,
Setiya’s theory of action establishes a simple, rather than a differentially realizable, constitutive aim of self-
knowledge. As the summary suggests, Setiya’s argument differs from my argument in two ways. First, Setiya’s
argument is premised upon the rejection of Velleman’s theory of action: he argues that if we reject Velle-
man’s theory of action, and accept an alternative account, then action has (what I have been calling) a simple
constitutive aim. By contrast, I argue that Velleman’s own theory of action yields only a simple constitutive aim.
Second, in the next chapter I will show that my argument is not dependent upon the details of Velleman’s
theory, nor upon the acceptance of any particular theory of action; rather, it raises a general problem for all
constitutivist theories.
DOES AN APPEAL OR QUEST HELP? 79
3. Does an appeal to improvisational actors or a quest

for intelligibility help?
I have argued that Velleman’s theory of action does not support his theory of practical
reason. But one might wonder whether Velleman has grounds for a response. In his
most recent work, Velleman changes focus to some extent. Rather than focusing
merely on immediate knowledge, Velleman appeals to the notions of intelligibility,
self-conceptions, and improvisational actors (Velleman 2009). In short, he attempts to
use a more complex account of self-knowledge in order to bolster his theory. In this
section, I will argue that this change does not enable Velleman’s theory to overcome
my objection.
In his most recent work, Velleman presents his account of action as being based on
an idea from Anscombe:
Anscombe pinpointed the sense in which we make up our autonomous actions but not our
behavior. The contrast lies in how we stand in relation to our behavior in the forethought about
it. When the fact that we are going to do something makes us think so, then we clearly have not
made it up; the case in which we have made up the fact that we are going to do something is the
case in which our thinking makes it so. And in the latter case but not the former, our behavior
will amount to autonomous action. (Velleman 2009, 130)
In other words, an agent performs an action, as opposed to producing mere behavior,

when the agent’s behavior is caused by his thoughts about what he will do.
Velleman claims that we can be more specific about the content of these thoughts: as
he puts it, “my conception of agency elaborates on Anscombe’s insight by spelling out
the relevant thought and its relation to action” (2009, 131). In particular, Velleman
claims that we can elucidate the nature of this thought by thinking about improvisa-
tional acting. The improvisational actor, Velleman says,
manifests his actual thoughts and feelings, as elicited from his actual makeup by his actual
circumstances, in accordance with his idea of what it makes sense for him to do in light of
them. Because this improviser will seek to enact the dispositions he has, and hence will generally
have the disposition he seeks to enact, he will tend to have two sources of motivation for
whatever he does—the first-order dispositions that belong to him as the character, and the
higher-order motive to make sense by enacting them. (Velleman 2009, 14)
The improvisational actor has various motives belonging to his character. For example,
if the actor is pretending to be a tempestuous mobster whose lover has been murdered
by a rival, his character will have a motive for revenge. The actor also has a second,
higher-order motive: he is motivated to perform an action that would make sense for
his character, given that his character has the motive of revenge. For example, it would
make sense for his character to grab a gun and seek revenge, rather than calmly writing
his rival a dispassionate letter of complaint. So, the actor’s higher-order motive of
doing-what-would-make-sense-for-his-character will lead him, in conjunction with

his character’s motive for revenge, to grab the gun.
Velleman then continues,
When there are conflicts among the former, first-order dispositions, the latter motive will tend to
determine which one prevails, by adding its force to the one that strikes him as making most
sense for his character to manifest. He will then act partly out of the improviser’s aim of giving
the most plausible rendition of his character, but also, like a method actor, out of the character’s
underlying motives, which will fund the enactment. (Velleman 2009, 14–15)
So, imagine that the character is torn between a desire for revenge and a desire to avoid
damaging his business: he wants to avenge his lover, but he also wants to avoid
escalating a conflict that could undermine his business. Velleman is claiming that the
actor’s higher-order motive of doing-what-makes-sense-for-the-character will deter-
mine which of these first-order motives prevails. For example, if the character had
hitherto been dispassionate and calculating, then it would make more sense for the
character to act on the motive of doing what’s best for his business. But suppose the
character has hitherto been tempestuous and emotional. Then it will make more sense
for the character to act on the motive of revenge. Thus, that’s what the actor will do: he
will make the character seek revenge. In this way, the higher-order motive adds
motivational force to the lower-order motive, and thereby determines what the
actor does.
Why are these reflections on improvisational actors relevant? Well, Velleman argues
that the same structure is present in ordinary agents. An agent will have a variety of
motives, which often conflict, and she will also have a higher-order aim of doing what
makes sense in light of those motives. The higher-order aim will add motivational
force to particular lower-order motives, thereby determining what the agent does.
Thus, linking this idea to Anscombe’s thesis, Velleman writes that “the respect in
which we resemble improvisational actors, who are the authors of what they do
onstage,” is this:
When the agent thinks about what it would make sense for him to do in light of his circum-
stances, attitudes, and attributes, he cannot honestly purport to be reading his future in them,
since what he does is going to depend on what he sees as making sense in light of them. His
preexisting motives will be joined, and their balance potentially altered, by the very motive that
leads him to think about them as clues to his next action, since that motive will incline him to do
what those clues render it most intelligible for him to do. (Velleman 2009, 132–3)
Here, Velleman emphasizes that just as the improvisational actor’s thoughts about what
it makes sense to do determine what he does, so too the ordinary agent’s thoughts
about what it makes sense to do determine what she does.
How do we transform this analogy between improvisational actors and agents into
an argument for constitutivism? Velleman reasons as follows:
DOES AN APPEAL OR QUEST HELP? 81
(1) What makes A-ing an action, as opposed to mere behavior, is that A-ing is
brought about by the agent’s expectation that he will A.
(2) These expectations have a particular content: they are thoughts about what it
would make sense to do, given one’s knowledge of one’s character. In other
words, they are thoughts about what is intelligible.
(3) Therefore, every action is in part motivated by a thought about what is intelli-
gible, and being motivated in this way is part of what makes something an
action.
(4) Therefore, action constitutively aims at intelligibility.
Premise (1) simply restates Velleman’s Anscombe-inspired conception of action.
Premise (2) is motivated by the reflections on improvisational actors. If we accept
these premises, they entail claims (3) and (4).
Let’s grant Velleman these claims. Unfortunately, they simply recreate the problem
plaguing his earlier theory. For these arguments concerning intelligibility neglect the
distinction between differentially realizable aims and simple aims. Grant that agents
must seek to render their actions at least somewhat intelligible. Velleman assumes that
this entails that agents will seek maximal intelligibility. We can see this in the last
quotation given above: in the first sentence, Velleman speaks of an agent being
motivated to do what would make sense; in the final sentence, Velleman assumes
that the agent will be inclined to do what is “most intelligible for him to do” (emphasis
added). Once we distinguish between simple and differentially realizable aims, though,
we can see that this assumption is unwarranted. The fact that an agent seeks some
intelligibility does not entail that an agent seeks maximal intelligibility, for intelligibility
could be a simple aim.
(Aim 10 ) Action has a simple aim of intelligibility. That is, action fulfills the
constitutive aim when it generates some intelligibility.
(Aim 20 ) Action has a differentially realizable aim of intelligibility. That is, an
action that yields some intelligibility fills the aim to a lesser degree than
does an action that yields maximal intelligibility.
Velleman’s reflections on improvisational actors lend support only to the claim that
action has constitutive Aim 10 .
To illustrate this point, it may help to consider an example. Return to our impro-
visational actor who is impersonating a mobster torn between a desire for revenge and a
desire to preserve his business. The improvisational actor had better do something that
makes some sense, given these motives. For example, the actor had better not have the
character simply forget about the motives and watch television all day. But from this
fact, it does not follow that the actor must determine and act out the action that makes
most sense. Suppose, for example, that the actor is sure that what makes most sense for
his character, given the character’s motives, is to seek revenge. That’s what would be
most consistent with the character’s past actions and the intensity of this motive.
Nonetheless, the actor might choose to have his character forgo revenge. After all, the
actor may have other motives: he may want to surprise the audience with an unex-
pected move; he may want to make the character less predictable; he may want to
display a character who realizes that his motives weren’t what he took them to be; or,
he may simply want the character to have a change of heart. Despite the fact that
forgoing revenge is not the action that makes the most sense, it is implausible to suggest
that the improvisational actor is somehow a failure for choosing it.14
But if this is true of the improvisational actor, it should also be true of the ordinary
agent. I may have a motive for doing what is intelligible in light of my motives. I may
even be committed to doing something that is somewhat intelligible in light of my
motives. But it does not follow that I always have most reason to do what is most
intelligible in light of my motives.
4. Can we supplement the constitutivist theory

with additional premises?
My argument against Velleman takes the following form: even if we grant the theory
of action, the theory of practical reason does not follow. In other words, Velleman
presents his theory of action as directly entailing a substantive account of practical
reason; I have argued that, once we distinguish between simple and differentially
realizable aims, this alleged entailment fails.
Given the structure of my argument, there is an obvious response on the constitu-
tivist’s behalf: one could weaken the central claim of the constitutivist theory. Rather
than maintaining that the account of action by itself entails an account of practical
reason, one might argue that the account of action coupled with some other premise entails
an account of practical reason.
For example, suppose Velleman admits that his argument concerning immediate
knowledge yields only a simple constitutive aim of self-knowledge. All that the
constitutive aim will tell us, then, is that we have reason to perform any action that
generates self-knowledge. But suppose we then supplement this theory with an
additional premise, such as this: if you aim at self-knowledge, then you ought to aim
at maximal self-knowledge. The theory would then generate claims about what we
have more reason to do. While this move is available to the constitutivist, it faces
several problems, which I discuss below.
14
A potential objection: if the actor has these motives, perhaps having his character forgo revenge is what
makes the most sense for him to do, after all. This objection fails. Consider the claim that we are motivated to
act in the way that makes the most sense in light of our first-order motives. If this claim is not to be vacuous,
what the agent actually does and what makes the most sense in light of the agent’s motives have to be capable
of coming apart.
C A N W E S U P P L E M E N T T H E C O N S T I T U T I V I S T T H E O RY ? 83
4.1 Supporting the constitutivist project with premises about non-constitutive features
The most straightforward way for a constitutivist to argue for a supplementary principle
is this: first, concede that there are no facts about the constitution of action as such that
establish a differentially realizable aim; second, argue that there are other facts about
agency or, more generally, about human life, which indicate that the constitutive aim is
differentially realizable.
Some of Velleman’s writings suggest an argument of the above form. For example,
Velleman argues that there is empirical evidence for the claim that human beings have a
differentially realizable desire for self-knowledge (Velleman 2006, Chapter 10). More-
over, his most recent work argues that facts about the nature of emotion and about
social interaction would be explicable if we had a differentially realizable desire for self-
knowledge (Velleman 2009).
It might seem that Velleman could support his constitutivist theory by appealing to
these considerations. For he could argue that, although the constitutive features of
action establish only a simple aim of self-knowledge, the above evidence indicates that
human beings tend to have a differentially realizable aim of self-knowledge.
Unfortunately, this argument would be of no help. Even if it were true that all
human beings have a differentially realizable aim of self-knowledge, this fact would not
provide any support for constitutivism. There are two problems, which I will address in
turn.
The first problem can be illustrated by an analogy. Suppose that empirical research
suggests that human beings are short-term hedonistic calculators. That is, suppose we
could show that human beings are disposed to choose the action that produces the
most pleasure in the short term. This empirical finding would leave open the norma-
tive question of whether human beings ought to choose the action that produces the
most pleasure in the short term. Indeed, it seems fairly obvious that, if there were such a
tendency, agents ought to resist it. For example, it seems clear that prudential consider-
ations should play some role in choice.
As this example illustrates, the mere fact that agents are inclined to choose which
actions to perform by determining which action has feature F does not imply that
feature F is normatively relevant. (This was the problem with brute naturalism,
discussed in Chapter 2, Section 4.1.) Accordingly, even if there were empirical research
conclusively establishing that human beings have a tendency to perform the actions
that can be done most self-knowingly, this would not provide any support for a theory
of practical reason that claimed that what there is reason to do is what can be done most
self-knowingly.15 (Constitutive aims are different on this score, because they are
inescapable.)
15
Of course, if we supplement the above claim with additional normative principles, it might generate
normative conclusions. But we would then need a defense of these substantive principles.
This brings us to the second problem. Even if human beings did have a differentially
realizable desire for self-knowledge, this would just be one desire among many others.
Recall that constitutive aims are distinctive in that they are essential components of the
activities that they regulate: the presence of the constitutive aim is part of what
constitutes an event as an instance of its kind. Accordingly, one cannot abandon the
aim without abandoning the activity. But most aims lack this feature. In particular,
even if empirical research established that all human beings have a differentially
realizable aim of self-knowledge, we could still abandon this aim without thereby
ceasing to act.16 So this aim would lack the special status of constitutive aims. In other
words, this aim would be on par with the host of other aims that human beings tend to
have, such as the aims of happiness, health, sociality, and so forth. Just as we can ignore
or dispense with these aims in various contexts without ceasing to act, so too we could
ignore or dispense with the differentially realizable aim of self-knowledge. Being
optional, the aim would not be capable of generating a non-optional standard of
success for action.
4.2 Generalizing the point
We are considering whether a constitutivist can appeal to some non-constitutive feature
of agency (or, more generally, of human life) that, when coupled with the claim that
action has a simple constitutive aim, yields a substantive account of practical reason.
I pointed out that if Velleman attempted to supplement the claim that action has a simple
constitutive aim of self-knowledge with the empirical claim that human beings tend to
have a differentially realizable desire for self-knowledge, this would not help, for the
latter desire lacks any distinctive normative standing. This point can be generalized.
Constitutivism is distinctive in that it grounds norms in facts about the essential
nature of action, and thereby renders the question “why should I act on aim X?” moot:
if the aim is constitutive of action, then we have no alternative to acting on it. Its
normativity is vouchsafed by its connection to agency. But, for any supplementary
principle, we can ask why it would have any normative standing. By hypothesis, the
answer cannot be that this supplementary principle derives from a constitutive feature
of agency. So we will have to find some other kind of support for the supplementary
principle. This will not be easy, because we will have to show not just that the principle
is one source of reasons among others, but that it has a privileged status in determining
what we have reason to do. For example, with respect to Velleman’s theory, we would
have to show that if I can either help a friend in need at the cost of some self-
knowledge, or abandon the friend in order to maximize self-knowledge,17 I ought
to do the latter. This would require a substantial argument.
16
Notice that I am only claiming that we could abandon the differentially realizable aim of self-knowledge.
If Velleman’s theory of action is correct, then we cannot abandon the simple aim of self-knowledge.
17
Abandoning my friend would maximize self-knowledge if, for example, doing so were consistent with
my character.
CONCLUSION 85
A more general problem is that if a constitutivist theory requires a supplementary

principle of this sort, then the appeal to constitutive features seems to be doing very
little work. Indeed, if a constitutivist incorporates a supplementary principle that is not
derived from constitutive features of agency, and thereby abandons the original claim
that constitutive features of agency suffice to generate substantive normative conclu-
sions, it is unclear why claims about constitutive features are even relevant. It seems that
we could leave them out without thereby affecting the theory’s normative conclusions.
Perhaps this is why both Velleman and Korsgaard seem reluctant to appeal to non-
constitutive features in their accounts of practical reason. For example, Velleman
rejects the idea that we can generate an adequate theory of practical reason simply by
appealing to an “intuitively plausible criterion of success” (Velleman 2004b, 287).
Instead, he argues that hypotheses about practical reason must be based on facts about
action’s constitutive aim. These “hypotheses about the constitutive aim of action,” in
turn, “earn credibility . . . not by resting on unsupported normative intuitions, but by
explaining otherwise inexplicable aspects of agency” (Velleman 2004b, 287–8). As we
will see in the next chapter, Korsgaard is even more clearly opposed to the idea that
non-constitutive features could be used to establish conclusions about practical reason.
She writes that “the only way to establish the authority of any purported normative
principle is to establish that it is constitutive of something to which the person whom it
governs is committed” (Korsgaard 2009, 32).
5. Conclusion
If the arguments given above are correct, then Velleman’s constitutivist theory is
unsuccessful. We cannot move from the fact that action aims at self-understanding or
intelligibility to the claim that we have more reason to perform those actions that yield
more self-understanding or intelligibility. In short, Velleman’s theory of action does
not yield a substantive account of practical reason.
This brings us to the second constitutivist theory that has been defended in the
literature. Christine Korsgaard has argued that action is governed by the constitutive
standard of autonomy or self-constitution. The next chapter considers the prospects for
this view.
4
Constitutivism and
Self-Constitution
The previous chapter examined David Velleman’s constitutivist theory. I argued that
Velleman’s theory is unsuccessful: even if we grant Velleman’s account of action, his
theory of practical reason does not follow. This chapter will assess the second extant
version of constitutivism, which has been developed by Christine Korsgaard. I argue
that her theory encounters an analogous problem.
Section 1 begins by examining Korsgaard’s argument for the claim that action has a
constitutive principle of autonomy or self-constitution (these, she argues, turn out to
coincide). Section 2 explains how Korsgaard’s theory of action generates claims
about reasons for action. Section 3 argues that Korsgaard’s theory encounters a diffi-
culty that is analogous to the one plaguing Velleman’s theory: it is ambiguous
on whether the constitutive feature of action is simple or differentially realizable. In
Section 4, I argue that this ambiguity renders Korsgaard’s theory unsuccessful: once
we resolve the ambiguity, Korsgaard’s account of action does not yield any substantive
conclusions about reasons for action. Put differently, even if we grant her theory of
action, her account of practical reason does not follow. Sections 5 and 6 consider
potential responses on behalf of Korsgaard. I argue that these responses cannot rectify
the problem with Korsgaard’s theory. Thus, Section 7 concludes that Korsgaard’s
version of constitutivism, like Velleman’s version, is untenable.
1. Korsgaard’s Kantian version of constitutivism

This section will reconstruct Korsgaard’s argument for constitutivism. Before beginning,
though, it will be helpful to note a difference between Korsgaard’s and Velleman’s
versions of constitutivism. Velleman’s theory starts with the non-normative claim that
action has a constitutive aim, and uses this claim to generate a normative result. As
Velleman puts it, “my purpose in trying to identify a constitutive aim of action is to find
a non-normative foundation for our norms of practical reasoning” (2004b, 287; emphasis
added). However, Korsgaard eschews this strategy, preferring instead to start with a
normative claim. Thus, Korsgaard writes that “the very idea of action is a normative
one” (2009, 109). Rather than starting with the non-normative claim that action has a
KOR SGAAR D’S KANTIAN VERSION OF CONSTITUTIVISM 87
constitutive aim, Korsgaard argues directly for the normative claim that action has
a constitutive standard or constitutive principle.
As I pointed out in Chapter 1, Section 3.3, this is a more properly Kantian strategy.
Kantians ground normativity in self-imposed principles, rather than desires. Accord-
ingly, Velleman’s and Korsgaard’s theories generate claims about reasons for action in
quite different ways. I flag this point here; I will discuss it in more detail in Section 4.
With this in mind, let’s reconstruct Korsgaard’s argument for the claim that action’s
constitutive principle is autonomy. Recall that Velleman’s constitutivist theory has the
following structure: first, we locate some essential feature of action; then, we show that
the presence of this essential feature of action is best explained by the hypothesis that
action has a constitutive aim. In Velleman’s theory, the essential feature is immediate
knowledge, and the hypothesized constitutive aim is self-knowledge. Korsgaard’s
theory has an analogous structure. She claims that an essential feature of action is that
actions are attributable to unified agents; and she claims that this feature is best explained
by the hypothesis that action’s constitutive principle is the Categorical Imperative.1
What does it mean to say that actions are attributable to unified agents? Korsgaard
explains her view in the following passage:
An action requires an agent, someone to whom we attribute the movement in question as its
author. And I also believe it is essential to the concept of agency that an agent be unified. That is
to say: to regard some movement of my mind or my body as my action, I must see it as an
expression of my self as a whole, rather than a product of some force that is at work on me or in
me. Movements that result from forces working on me or in me constitute things that happen to
me. To call a movement a twitch, or a slip, is at once to deny that it is action and to assign it to
some part of you that is less than the whole: the twitch to your eyebrow or the slip, more
problematically, to your tongue. For a movement to be my action, for it to be expressive of myself
in the way that an action must be, it must result from my entire nature working as a unified
whole. (Korsgaard 2009, 18)
In this passage, Korsgaard relies on a familiar distinction between two broad classes of
things that people do. On the one hand, we sometimes sneeze, cough, fall asleep, blink,
and so forth. On the other, we sometimes contemplate philosophy, conduct conver-
sations, get married, decide to go to Bermuda, and so forth. Each of these events counts
as something that a person does, in a sense, but there seem to be important differences
between, say, sneezing and getting married. The former is a reflex, something that
happens to me, something that is not entirely under my control; the latter is a product
of choice, something that I do, something that is to some extent under my control. We
1
The more ambitious versions of constitutivism argue as follows: given that action has feature F, it must
be the case that action has a constitutive aim/principle. Less ambitious versions argue for a more modest
claim: given that action has feature F, we have good reason to believe that action has a constitutive aim/principle.
Velleman embraces the less ambitious version: he characterizes his claim that action has a constitutive aim as a
case of “inference to the best explanation” (2004b, 287–8). Korsgaard seems to embrace the more ambitious
version, as we will see below.
88 CONSTITUTIVISM AND SELF-CONSTITUTION
can mark this distinction by calling the sneeze and its ilk mere behaviors, and the marriage
and its ilk actions.
With the distinction between action and mere behavior rendered explicit, we can
put Korsgaard’s claim as follows: the essential difference between actions and mere
behaviors is that the former are attributable to the agent as a unified whole, whereas the
latter are attributable merely to some part of the agent. My blinking is attributable to
my eyes, my sneezing to my irritated nostrils; my deciding to go to Bermuda, or to get
married, is attributable to me as a unified whole.
In sum, Korsgaard is making two claims:
(A) we need a distinction between action and mere behavior.
(B) actions are movements that are attributable to a unified agent, whereas mere
behaviors are movements that are attributable only to some part of the agent.
Claim (A) is completely uncontroversial. A defining feature of action theory is that it
sets out to distinguish actions from other movements: every account of action will
classify certain movements as non-actions. Of course, there is controversy over where,
exactly, to draw the line between actions and non-actions. Thus, every theory of
which I am aware classes non-voluntary movements such as the beating of one’s heart,
the blinking of one’s eyes, and so forth, as non-actions. However, theories disagree
over whether automatisms, reflexes, and non-consciously motivated behaviors are to
be classed as non-actions (for opposing views, see for example Velleman 2000 and
Arpaly 2004).
Claim (B) is more controversial; a number of prominent accounts of action deny it.
For example, Donald Davidson distinguishes intentional action and non-action behav-
ior in terms of its etiology: intentional action is both caused and rationalized in a non-
deviant way by the agent’s beliefs and desires (Davidson 1980, Essay 1). If we assume
that beliefs and desires are parts of the agent rather than the whole agent, then
Davidson’s account accepts (A) and rejects (B).
So an initial task for Korsgaard is to show why we should accept (B). Another task is
to explain what is meant by (B). After all, the idea of being attributable to a unified
whole is not exactly transparent; we could interpret it in a number of different ways.
The success of Korsgaard’s theory will depend on these tasks. If she can show that
accepting (A) and some plausible version of (B) commits us to the Categorical Impera-
tive (hereafter CI), this would be an extremely significant result. But if she can only
show that (A) and a highly substantive and controversial version of (B) commit us to the
CI, this would be far less significant. Everything depends on the starting points: the
more content we pack into our initial characterization of agency, the easier it will be to
extract normative content, but the less interesting that result will be.
So how much content must Korsgaard build into her starting conception of agency?
Can she employ a weak and uncontroversial version of (B), or must she resort to a
substantive and controversial version of (B)? I will argue that the latter is the case. In
particular, I will show that she is faced with a dilemma: either the meaning of (B) shifts
S TA G E T WO 89
over the course of Korsgaard’s argument, in which case the argument is invalid; or the
meaning of (B) is kept constant throughout the argument, in which case the argument
conflates simple and differentially realizable aims and relies on an implausible account
of action. To see this, let’s examine her argument.
Korsgaard’s argument is long and complex, so it will be helpful to start with an
outline—we can then fill in the details as we go. I think Korsgaard’s argument is best
seen as proceeding in three stages:
Stage One: Specify the starting conception of action, from which the theory’s
normative conclusions will be extracted. These are claims (A) and (B): there is a
distinction between action and mere behavior, and an agent’s A-ing is an action iff
A-ing is attributable to the agent as a unified whole.
Stage Two: Explicate and defend claim (B). Korsgaard does so by arguing for a
theory of action that she finds in the work of Kant and Plato. Korsgaard argues that if
we accept a Kantian/Platonic characterization of action, then we will see that a
version of (B) is correct.
Stage Three: At this point, we will have explicated and defended (B). Korsgaard next
argues that in order for the condition specified in (B) to be met—in order for an
action to be attributable to an agent as a unified whole—the principle of the agent’s
action must be the CI. Given that action requires attributability to a unified whole,
and attributability to a unified whole requires acting on the CI, it follows that
action’s constitutive principle is the CI.
That, in outline, is Korsgaard’s argument. We have already examined Stage One.
Below, I will reconstruct Stages Two and Three.
2. Stage Two: explication and defense of (B)

2.1 Self-conscious reflection engenders parts of the soul
Claim (B) says that actions are movements that are attributable to a unified agent,
whereas mere behaviors are movements attributable only to some part of the agent.
We need to know both what this means and why we should accept it. Let’s begin by
determining what the parts that stand in relations of unity and disunity are. In short:
what is it that is being unified?
Korsgaard argues that “self-consciousness . . . introduces what, following Plato, I will
call the parts of the soul” (2009, 119). While the notion of “parts of the soul” may seem
obscure, Korsgaard thinks the idea is borne out by reflection upon the nature of self-
conscious agency. Self-conscious creatures have the ability to reflect on and thereby
distance themselves from their desires. As Korsgaard puts it,
Our capacity to turn our attention to our own mental activities is also a capacity to distance
ourselves from them, to call them into question . . . I desire and I find myself with a powerful
impulse to act. But I back up and bring that impulse into view and then I have a certain distance.
Now the impulse doesn’t dominate me and now I have a problem. Shall I act? Is this desire really
a reason to act? (Korsgaard 1996b, 93; cf. Korsgaard 2009, 126)
Self-conscious reflection introduces “parts of the soul” in the sense that I experience
my reflective thought as distinct from my desires. Consider a typical case of deliber-
ation: I experience a desire to dance, and consider whether I should do so. I don’t
experience my desire as simply forcing me to dance. Reflecting on the desire, I
experience myself as free to decide whether to act on it. In this sense, self-consciousness
“opens up a space” between the desire and the action, and “is therefore the source of a
psychic complexity” (Korsgaard 2009, 125). In particular, self-consciousness opens up
a distance between the reflecting self and the desire it reflects upon.2 We can describe
this by saying that self-consciousness divides us into parts.
The claim that self-consciousness introduces a distance between the agent and her
desire is recognizably Kantian. Yet Korsgaard claims that Plato’s theory of action is
premised upon a similar claim. In the Republic, Plato argues that the soul can be
described as having three parts: Reason, Appetite, and Spirit. Korsgaard argues that
the above remarks on reflective agency can be put in a roughly Platonic fashion: self-
consciousness creates a distinction between Reason and Appetite. Appetite is the locus
of desire: its function “is to supply the whole person with whatever he needs”
(Korsgaard 2009, 141). Reason’s “function is to direct things, for the good of the
whole person” (Korsgaard 2009, 141). The above reflections show, Korsgaard thinks,
that Reason and Appetite are distinct. Moreover, further reflection on the nature of
action reveals a third part of the soul: Spirit. Korsgaard interprets Spirit as what we
would ordinarily call willpower or resoluteness. Consider the connection between
desires, decisions, and actions. Having reflected on my desires and decided to act in a
certain way, I will need to maintain this decision in the face of contrary pressures.
I decide to work, but experience desires to speak with friends, watch television, and so
forth; if I am to carry out my decision, I need to resist these pressures. This is what
Korsgaard’s Plato describes as Spirit. As Korsgaard puts it, Spirit “controls appetite in
the face of temptation, pleasure, pain, and fear” (Korsgaard 2009, 145).
Thus, Korsgaard maintains that Kant and Plato have the same picture of reflective
agency: both philosophers agree that self-conscious agents experience a reflective
distance from their desires, and in this sense have two or three parts of the soul: self-
consciousness and desire for Kant; Reason, Appetite, and Spirit for Plato.3 Of course,
we can certainly question these claims: we can ask whether these characterizations of
2
This claim is controversial, and Chapter 5 will investigate some problems with it. For now, I grant the
claim for the sake of argument.
3
It is a familiar point that Kant’s moral psychology is dualistic, employing a distinction between reason
and passion. Yet Korsgaard claims that we can find an analogue for Plato’s third part—Spirit—in Kant’s moral
psychology. If this is right, then Kant and Plato will be in perfect agreement on the basic structure of agency.
As this point will not be relevant for my argument, I pass over it here. For Korsgaard’s argument in favor of
the idea that Kant recognizes some version of Platonic Spirit, see Korsgaard (2009, 153–8).
S TA G E T WO 91
the nature of self-conscious agency are defensible. In the next chapter, I am going to
argue that some of these characterizations are erroneous, and I will thus defend an
alternative model of reflective agency. For now, though, let’s grant Korsgaard this
description of agency, in order to see what might follow from it.
2.2 Reflection as undetermined by desire
Self-conscious agents have parts of the soul; how do these parts relate in the production
of action? Above, we saw that Korsgaard describes the self-conscious agent as “stepping
back” from her desires. Korsgaard next claims that once the agent steps back from a
desire, she need not act as the desire suggests: “reason need not follow inclination”
(Korsgaard 2009, 125).4 Instead, the agent can make a choice about how to act
(Korsgaard 2009, 119). For example, I might experience desires both to continue
writing this chapter and to stop and take a walk. Korsgaard’s claim is that I can decide
which of these desires to act upon; my action is not determined by them. (Obviously,
this claim is controversial. I grant it for now and critically assess it in the next chapter.)
2.3 The role of principles
I experience various desires, and am free to decide which desires to act upon. But what
happens when I make this choice? Korsgaard tells us that “in order to make that choice,
reason needs a principle” (2009, 213). She explains:
When we become conscious of the workings of an incentive within us, the incentive is experi-
enced . . . as a proposal, something we need to make a decision about. Cut loose from the control
of instinct, we must formulate principles that will tell us how to deal with the incentives we
experience. And the experience of decision or choice . . . is a separate experience from that of the
workings of the incentive itself. (Korsgaard 2009, 119)
It’s easiest to illustrate this with an example. Suppose I am hungry, and experience a
desire for food. I reflect on this desire for food, considering whether and how to act on
it: I could order a salad, but the pizza also sounds appetizing. In reflecting on these
considerations and ultimately making a decision, I can be described as formulating
various principles that would determine how I act on the desire. Here are two such
principles: when I am hungry, I will eat nutritious food; when I am hungry, I will eat whatever
strikes me as most appealing. Depending upon which principle I employ, I will make
4
Korsgaard often speaks of “inclinations” and “incentives” rather than “desires.” She defines inclinations
as “a kind of attraction to something, which is grounded in our sensuous nature” (2008, 46). Incentives, by
contrast, are “features of the represented object that make it, from some point of view, attractive or aversive”
(2009, 120). For example, “if dancing is pleasant, that means there is a natural incentive to dance” (2009,
120). So incentives are features of objects, whereas inclinations are features of agents. Or, to put this in more
familiar terms: inclinations are desires, whereas incentives are the desired properties of objects. For present
purposes, the distinctions between inclinations and incentives will not be relevant, so I will employ the more
familiar term “desire.”
different decisions: the first principle might lead me to order a salad, the latter a pizza. In
this way, my principles interact with my desires in order to determine what I do.5,6
2.4 Principles determine the relationship between the parts of the soul
So far, we have the following picture of self-conscious action: various desires arise (the
work of Appetite); the agent reflects on them, employs a principle determining which
of these desires to take as reasons, and makes a choice (the work of Reason); and, if all
goes well, the agent acts as she has decided to act (the work of Spirit). Korsgaard next
argues that the relationship between the parts of the agent’s soul is determined by the
principle upon which the agent acts.
In support of this claim, Korsgaard appeals to an analogy between city and agent: just
as different laws specify different ways in which the citizens and rulers relate in the
production of a state’s action, so too different principles specify different ways in which
the parts of the agent relate in the production of a person’s action (Korsgaard 2009,
163). Consider an analogy: in an unrestricted democracy, the state’s actions are deter-
mined by whatever the majority wants. Korsgaard claims that the agential analogue
would be a person who acted on the principle, “I will do whatever I most strongly
desire at the moment.” Or consider another example: in a timocratic state, the state’s
actions are based upon the quest for honor and victory. Korsgaard tells us that the
agential analogue would be a person who acted upon the principle, “I will do whatever
is regarded as honorable in the present circumstances.”7 With these examples in mind,
we can see how relationships among parts of the soul can be determined by principles:
certain principles will specify which parts are to have priority in the production of
action. For example, the “democratic” principle gives Appetite complete priority in
determining action; the “timocratic” principle gives one specific appetite (honor) com-
plete priority; and so on.
5
These needn’t be explicitly formulated principles. As Korsgaard puts it, “liberation from the government
of instinct means that it is up to us to decide what justifies what, what counts as a reason for what, what is
worth doing for the sake of what. We don’t need to think of this . . . as a decision made prior to action; as often
as not, it is a decision embodied in the action. Action involves an incentive and a principle . . . If you choose to
run in order to escape your predator, to stand your ground in order to protect your offspring, or to dance for
the sheer joy of dancing, then those are your principles, your conception of what is worth doing for what”
(Korsgaard 2009, 127).
6
Why must we employ principles? Korsgaard’s argument takes the following form: in order to act, we
must think of ourselves as the causes of our actions; seeing something as a cause requires seeing it as operating
in accordance with some principle (or law); therefore, in order to act, we must think of ourselves as in
accordance with some principle. As she puts it, “Determining yourself to be the cause is not the same as being
moved by something within you, say some desire or impulse . . . operating as cause. When you deliberate,
when you determine your causality, it is as if there is something over and above all of your incentives,
something which is you, and which chooses which incentive to act on. So when you determine your own
causality you must operate as a whole, as something over and above your parts, when you do so” (Korsgaard
2009, 72). For an insightful critique of this argument, see Tubert (2011).
7
For Korsgaard’s discussion of the timocratic agent, see Korsgaard (2009, 165–6). For the democratic
agent, see Korsgaard (2009, 168–9).
S TA G E T WO 93
2.5 Principles determine the degree of unity

So Korsgaard has argued that when an agent acts, the parts of the soul can be related in
different ways. The relationships between the parts are determined by the principle
upon which the agent acts (Korsgaard 2009, 162–3). As she puts it: “deliberative action
[i.e., action involving a principle] by its very nature imposes unity on the soul. When
you deliberate about what to do and then do it, what you are doing is organizing your
appetite, reason, and spirit into a unified system that yields an action that can be
attributed to you as a person” (Korsgaard 2009, 179).
At this point, it might seem that we have an answer to our initial question of what
unity is: an action is attributable to a unified agent iff the agent acts on some principle—
any principle. But Korsgaard adds a wrinkle: she next argues that different principles
generate different degrees of unity among the parts (Korsgaard 2009, 163; cf. 175). As
she puts it, “an action can unify and constitute its agent to a greater or lesser degree”
(Korsgaard 2009, 163). For example, she argues that acting on the democratic principle
will generate one degree of unity, whereas acting on the timocratic principle will
generate another degree of unity. Thus, she writes,
the extent to which one is unified, and so is an agent, is a matter of degree. Timocrats . . . are
pretty well unified; oligarchs, who are divided against themselves, still manage to hold themselves
together so long as one part keeps the other firmly repressed; even democrats, who are united
only by the principle of being all in pieces, can hang together if they are lucky . . . in the contents
of their desires. (Korsgaard 2009, 174)
Korsgaard’s “timocrats” are agents who act on the principle “I will do the honorable
thing”; oligarchs are agents who act on the principle “I will do what is prudent”; and
democrats are agents who act on the principle “I will act on my strongest desire.”8
In the above passage, then, Korsgaard is claiming that agents who act on the first
principle are more unified than agents who act on the second, and agents who act on
the second are more unified than those who act on the third. In short, the degree to
which a particular action constitutes the agent as unified depends upon which principle
the agent employs in acting.
Why exactly is this? In other words, what makes it the case that the democratic agent
is less unified than the timocratic agent? This brings us back to a point raised in Stage
One: the concept of unity is not transparent. It can be analyzed in a number of different
ways, and hence needs explication. Surprisingly, Korsgaard never provides an explicit
definition of this crucial notion. However, we are now in a position to see what she
means by unity. For her discussions make it clear that she judges agents to be unified to
the extent that they exhibit a kind of diachronic stability (though she does not use this
term).
8
For these definitions of the various types, see Korsgaard (2009, 165–9).
To see this, let’s focus on Korsgaard’s example of the agent who acts on the
“democratic principle” (“I will do whatever I most strongly desire”). Korsgaard tells
us that “the coherence of the democratic person’s life is completely dependent on the
accidental coherence of his desires” (Korsgaard 2009, 169). She illustrates this with an
example of a college student who acts on whichever desire happens to be strongest: he
begins to study, finds himself distracted, takes a walk, wanders into a bookstore, goes to
a party, and so on. In acting on whichever desire happens to be strongest, the agent
exhibits no stability in his commitments: he will only achieve his immediate end (studying,
enjoying a walk, etc.) if, through sheer accident, he encounters no contrary tempta-
tions. Korsgaard describes this as “chaos” (2009, 169). So this agent is disunified in the
sense that he will exhibit diachronic stability only if his desires happen to be stable.
To be clear, Korsgaard’s claim is not that unity requires uniformity or single-
mindedness. The problem with the democratic agent is not that he changes his mind.
Changing one’s mind is fine. The problem is that the democratic agent’s commitment
to his end is not a genuine commitment at all: the alleged commitment dissolves
whenever his desires change. So unity is having commitments, where having commit-
ments involves being capable of maintaining diachronic stability.
Another example makes the same point: suppose Korsgaard decides that she will
spend the day working on her book. Korsgaard lists a number of potential threats to her
unity: she will lack unity if she fails to “determine myself to stay on its track. Timidity,
idleness, and depression will exert their claims in turn, will attempt to control or
overrule my will, to divert me from my work . . . if I always allow myself to be derailed
by timidity, idleness, or depression, then I never really will an end” (Korsgaard 2009,
69). Here, again, the agent is said to be unified insofar as she has genuine commitments.
That is, she is capable of maintaining diachronic stability.
Thus, summarizing her examples of disunity, Korsgaard writes
I claimed that in the conditions of timocracy, oligarchy, and democracy, your unity, and your
capacity for self-government are propped by external circumstances, by the absence of conditions
under which you would fall apart. (Korsgaard 2009, 177)
In other words, only external contingencies keep the agent diachronically stable; only
external circumstances keep the agent “committed” to his end. So agents are disunified
if some circumstance could arise that would cause them to “fall apart,” in the sense of
undermining their commitments to their ends.9 As she puts it, the unity that a defective
principle such as “timocracy, oligarchy, and democracy” provides “is contingent and
9
Of course, we might question these claims: suppose a “democratic” agent has robust desires that are
consonant with his social setting and that lend unity to him over time. The agent might last for a long while—
indeed, for his whole life—without experiencing disunity. (Consider the non-self-conscious analogue:
employing Korsgaard’s model of agency, we should presumably describe a fish as driven by three desires:
to reproduce, to avoid danger, and to eat. These robust, stable desires structure the fish’s life, lending it order
over time.) I think reflections of this kind undermine Korsgaard’s claim. However, for the sake of argument
I want to grant Korsgaard the claim.
S TA G E T H R E E 95
unstable . . . [For example,] the democratic person drops his projects in the face of the
slightest temptation or distraction” (Korsgaard 2009, 175).
So now we know how to interpret Stage One: when Korsgaard claims that action
requires attributability to a unified agent, she is claiming that action requires attribut-
ability to an agent who is capable of maintaining diachronic stability. We need to know
whether we should accept this claim, but for the moment let’s grant it. Then we can
ask whether acceptance of this claim yields substantive normative conclusions.
3. Stage Three: from agential unity to the

Categorical Imperative
At this point, we can pose a question: which principles constitute us as fully unified
wholes (i.e., agents who are capable of maintaining diachronic stability)? We have seen
that agents acting on democratic, timocratic, and other principles don’t enjoy dia-
chronic stability, for there are circumstances in which these principles break down: the
agent’s commitment to her end dissolves. Korsgaard argues that the Categorical
Imperative is the only principle that avoids this problem, providing the agent with
diachronic stability:
you must operate as a whole, as something over and above your parts . . . And in order to do
this . . . you must will your maxims as universal laws. (Korsgaard 2009, 72)
The CI tells us to do just that: act only on those maxims that we can will as universal
laws. So, in the above passage, Korsgaard contends that if we are to operate as unified
wholes, we must act on the CI. Principles that are inconsistent with the CI either fail to
unify the agent at all, or fail fully to unify the agent.
Korsgaard’s argument for this conclusion proceeds as follows. Suppose we act on
some principle other than the CI. Kantians maintain that principles are either categor-
ical or hypothetical. Hypothetical principles have the form “if you want to X, then you
ought to Y.” So, if I don’t act on a categorical imperative, my principle will be
hypothetical: it will be contingent upon my having some motive. But, as Korsgaard’s
examples of democratic agents emphasize, motives are changeable; it is, as she likes to
put it, “only an accident” if our motives remain the same over time. But this implies
that hypothetical principles are contingent upon changeable elements of our psych-
ology. Accordingly, any diachronic stability that we manifest will be purely accidental;
it could disappear at any time, as soon as our desires change. So, in order to be capable
of diachronic stability, we must avoid anchoring our commitments in mere desires. But
this is just to say that we must adopt non-hypothetical principles. In other words, we
must act on the CI.
In short, the argument is as follows: if I choose some principle other than the CI,
then any diachronic stability that I seem to exhibit will be purely accidental; it could
dissolve at any time. But, if this happens—if I choose a principle that potentially
compromises my diachronic stability—then I am not really unified at all. (For this

argument, see especially Korsgaard 2009, 78ff.)10
Clearly, this argument might be questioned in a number of ways. I will not pause to
examine potential criticisms here. For I am going to argue that an insuperable difficulty
arises at an earlier stage of Korsgaard’s theory: we don’t have to worry about whether a
commitment to unity yields a commitment to the CI, for I am going to argue that
Korsgaard cannot establish a commitment to unity.
4. Korsgaard’s argument for constitutivism

Above, I have discussed the core claims in Korsgaard’s long and complex argument. To
review:
Stage One: Korsgaard states her initial conception of agency, from which she will
attempt to derive substantive normative conclusions. In particular, she claims
(1) An agent’s A-ing is an action iff A-ing is attributable to the agent as a unified
whole. (See Section 1)
Stage Two: Korsgaard explicates and defends this initial claim:
(2) Self-conscious agents experience a reflective distance from their desires, and
in this sense are divided into parts. (Section 2.1)
(3) Self-conscious agents are free to make decisions in a way that is not deter-
mined by their desires. (Section 2.2)
(4) When a self-conscious agent decides how to act, she employs a principle of
choice. (Section 2.3)
10
In addition to the argument given above, Korsgaard offers two rather different arguments for the claim
that full unity requires acting on the CI. One focuses on autonomy, the other on particularistic willing. I will
rehearse the arguments briefly in this note, but will not pause to offer full explications of them, as the details
are irrelevant for our purposes.
The autonomy-based argument proceeds as follows: “Action is determining yourself to be the cause of
some end . . . The Categorical Imperative binds you because what you are determining to be the cause of some
end is yourself ” (Korsgaard 2009, 81). Let’s unpack this. We have seen that action requires attributability to a
unified agent. To say that an action is attributable to the agent is just, Korsgaard suggests, to say that the agent
is the cause of the action. But now focus on the fact that the agent, rather than some part of the agent, is
supposed to be the cause of action. This, Korsgaard tells us, requires that the agent be self-determining or
autonomous: “In order to be an agent, you have to be autonomous, because the movements you make have to
be your own, they have to be under your own control . . . ” (Korsgaard 2009, 213). Korsgaard accepts Kant’s
claim that the CI specifies the nature of autonomy (I won’t rehearse these familiar arguments here). Thus, we
might put Korsgaard’s second argument as follows: first, an action is attributable to an agent as a unified whole
iff the agent is autonomous. Second, an agent is autonomous iff the agent acts on the CI.
The argument from “particularistic willing” proceeds rather differently. Korsgaard begins by claiming that
if our principles didn’t have to be universal, then they could be fully particular. That is, it would be possible to
have a principle that applies only to the case at hand, with no further implications for other cases. Call this a
particularistic reason. She claims that there is no such thing as a particularistic reason. Consequently, our
principles must be universal. But, Korsgaard reasons, the CI simply tells us to act on universal principles. For
this argument, see Korsgaard (2009, 72ff.).
KO R S G A A R D ’ S A R G U M E N T F O R C O N S T I T U T I V I S M 97
(5) The agent’s principle of choice determines how the parts of the agent’s soul
are related. (Section 2.4)
(6) Some relationships between parts of the soul engender unity; others fail to do
so. (Section 2.5)
(7) Unity is (implicitly) defined as having commitments that enable diachronic
stability. (Section 2.5)
(8) Thus, an action is attributable to the agent as a unified whole iff the
principle of the agent’s action renders the agent capable of diachronic
stability.
Stage Three: with the account of unity explicated and various claims about agency
defended, Korsgaard argues that only the CI provides diachronic stability. That is,
(9) The CI is the only principle acting upon which generates the capacity for
diachronic stability. (Section 3)
We can then put these claims together in order to show that the CI is the constitutive
principle of action. First, recall our original definition of constitutive aim, from
Chapter 1: Let A be an attitude or event. A has constitutive aim X iff there is some
aim X such that (i) each token of A is governed by X, and (ii) being governed by X is
what makes an attitude or event qualify as a token of A. We can offer an exactly
analogous definition of constitutive principle:
(Constitutive Principle) A has constitutive principle P iff there is some principle

P such that (i) each token of A is governed by P, and (ii) being governed by P is what
makes an attitude or event qualify as a token of A.
With that in mind, notice that Korsgaard’s argument entails that the CI is the consti-
tutive principle of action. In particular:
(a) An agent’s A-ing is an action iff A-ing is attributable to the agent as a unified
whole.
(b) An action is attributable to the agent as a unified whole iff the principle of the
agent’s action renders the agent capable of diachronic stability.
(c) The CI is the only principle which renders the agent capable of diachronic
stability.
(d) Therefore, an agent’s A-ing is an action iff the principle of the agent’s A-ing is
the CI.
From (d), it follows both that every action is governed by the CI, and being governed
by the CI is part of what constitutes something as an action.11 By the definition of
Constitutive Principle, this is just to say that action’s constitutive principle is the CI. So,
11
There is a complication here. Strictly speaking, it is not clear whether Korsgaard’s argument establishes
that we need full unity in order to act, or whether partial unity would suffice for action. That is, would a
partially unified agent be capable of producing an action? I will return to this point in the next section.
if we accept Korsgaard’s account of the nature of action, it turns out that action’s
constitutive principle is the CI.12
5. Bad action
In the previous chapter, I argued that Velleman’s theory foundered on the problem of
better and worse actions: because Velleman’s account of action entails only a simple
constitutive aim, it cannot generate any claims about what we have more or less reason
to do. In the following sections, I will argue that Korsgaard’s theory encounters a
version of this problem. She, too, is able to establish only a simple constitutive aim; and
this makes her incapable of establishing any conclusions about actions that are bad or
defective.13
5.1 Accounting for bad action
Consider the conclusion of Korsgaard’s argument:
(d) An agent’s A-ing is an action iff the principle of the agent’s A-ing is the CI.
This conclusion seems to imply that there is no such thing as bad action. After all, an
action is good iff its principle is the CI; by (d), A-ing is an action iff its principle is the
CI; therefore, every action is good. In other words, (d) entails that the constitutive
principle is achieved in each instance of action. If no intentional action can fail to
manifest the constitutive features of action, then no intentional action can be defective
as an action.14
Of course, a theory of practical reason that labeled all actions as good actions would
not be useful. The point of a theory of practical reason is to distinguish good and bad
12
A final word on the argument. Action’s constitutive principle is the Categorical Imperative precisely
because action constitutes the agent as a unified whole, and the Categorical Imperative is the only principle
that fulfills that demand. So, instead of claiming that action’s constitutive principle is the CI, we could
describe Korsgaard’s theory as claiming that actions’ constitutive principle is self-constitution. For these notions
turn out to be equivalent. On the other hand, it would also be correct to describe Korsgaard’s theory as
claiming that action’s constitutive principle is autonomy. For Korsgaard claims that an agent counts as
autonomous iff the agent acts as a unified whole. She writes, “In order to be autonomous, it is essential
that your movements be caused by you, by you operating as a unit, not by some force that is working in you
or on you. So in order to be an [autonomous] agent, you need to be unified—you need to put your whole
self, so to speak, behind your movements” (Korsgaard 2009, 213). Korsgaard here claims that acting as a
unified whole just is being determined by one’s self, and being determined by one’s self is a paradigmatic
description of autonomy. Thus, we can describe Korsgaard’s theory as claiming that action’s constitutive
principle is the Categorical Imperative, self-constitution, or autonomy. For, given the way that Korsgaard
uses the terms, these all come to the same thing.
13
Following Korsgaard, I use the terms “bad action” and “defective action” interchangeably, and define
them as follows: action A is bad or defective iff in the present circumstances there is another available action
that there is more reason to perform. For example, if in the present circumstances I can either donate a certain
sum of money to charity or spend it on a luxurious vacation, and if there is more reason to do the former, then
the latter action would be a defective action.
14
Philip Clark was the first to raise a version of this problem for constitutivism. In Clark 2001, he argued
that the bad action problem arose for an early (and now abandoned) version of Velleman’s theory.
BAD AC TI ON 99
actions, or actions that we have reason to perform from actions that we have reason not
to perform. If all actions were good actions, then theories of practical reason would be
otiose.15
For this reason, Korsgaard wants to leave room for bad action. To do so, she appeals
to a point that became clear in our discussion of claims (6) and (7), above: unity comes
in degrees. If unity comes in degrees, we can claim that defective action is action that
fails fully to unify its agent, whereas good action is action that succeeds in fully unifying
its agent. We can see this line of reasoning in the following passage:
bad action is action governed by a principle of choice which is not reason’s own: a principle of
honor (timocracy), prudence (oligarchy), wantonness (democracy), or obsession (tyranny). It is
action, because it is chosen in accordance with the exercise of a principle by which the agent rules
himself and under which he is—in a sense—constitutionally unified. It is bad, because it is not
reason’s own principle . . . Reason’s own principle, in contrast to all of these, is the principle that
truly unifies the soul . . . (Korsgaard 2009, 175, emphases added)
More succinctly, Korsgaard writes that defective action is possible because “an action
can unify and constitute its agent to a greater or lesser degree” (Korsgaard 2009, 163).
Thus, Korsgaard claims that what is required for something to be an action is that it
constitute the agent as unified to some extent. But what is required for something to be
a good action is that it constitute the agent as fully unified.16,17
We can generalize this point. Constitutivists derive norms from an account of
action’s constitutive features. The problem is that there is no distance between the
norm’s being applicable and the norm’s being met. Call this the bad action problem: it
seems that there is no such thing as a bad or defective action.
15
Might a constitutivist accept the idea that bad “actions” are not really actions? No; for there would be
two problems. First, this looks like a reductio ad absurdum. After all, it implies when a liar deliberately misleads
his audience, when shoplifter makes off with his loot, when a murderer plunges a knife into his victim, these
individuals aren’t performing actions. This is very hard to swallow. If the premises leading to this conclusion
had been unquestionable, we might accept these highly counterintuitive conclusions; but I think that even a
hardened Kantian should admit that we are much more confident that murder is an action than that (e.g.)
actions must be attributable to unified wholes. Second, the problem isn’t just that there turn out to be fewer
actions than we thought: the problem is that there will be no such thing as performing an action that there is
reason not to perform. If a theory of practical reason entails that every action is an action that is in accordance
with reason, then the theory of practical reason isn’t doing any work; it is not giving us a way of distinguishing
between rational and irrational actions, or between good and bad actions.
16
In fact, this is one of the pivotal moves in Korsgaard’s theory: indeed, it seems to be her primary
motivation for turning to self-constitution in the first place. For, if Korsgaard did not need a degreed
conception of action, then all the material on self-constitution could be left out; she could simply argue,
directly, that action requires acting on the Categorical Imperative. The appeal to self-constitution, though,
enables Korsgaard to have a degreed aim, and thus to account for the possibility of defective action.
17
There may be an additional problem with Korsgaard’s account of defective action. Douglas Lavin
argues that Korsgaard is committed to imperativalism, which is (roughly) the claim that if one is subject to a
normative principle, then one must be capable of violating it (Lavin 2004, 436). Lavin explains that Korsgaard
relies on this principle in her accounts of defective action (cf. Lavin 2004, 438–41). However, Lavin argues
that imperativalism is incompatible with constitutivism. If this is right, then Korsgaard’s account of defective
action is problematic. I will pursue a different objection.
How can a constitutivist avoid this problem? By incorporating a differentially

realizable constitutive feature. Suppose that action constitutively aims at G, and
suppose further that this aim is differentially realizable. Then we could offer the
following accounts of action and good action, respectively:
– What constitutes something as an action: achieving G to some extent.
– What constitutes something as an exemplary or good action: achieving
G completely.
Once we draw the distinction between what is required to be an action, and what is
required to fulfill action’s principle of success, the bad action problem is easily resolved.
The constitutivist argues that every instance of action to some extent achieves the
constitutive principle, or to some extent fulfills the constitutive aim. But many actions
fail to achieve full conformity with the constitutive principle, or fail to fulfill the
constitutive aim completely. These actions are still actions; but they are defective
actions. So, in order to account for bad action, we need differentially realizable
constitutive features. And this seems to be just what Korsgaard wants, as indicated by
the quotation above.
5.2 The admission of partially unified actions undermines Korsgaard’s theory
While Korsgaard’s degreed conception of unity seems to make room for defective
action, a problem arises when we ask how to classify the movements produced by
partially unified agents. Do these movements count as actions? There are three possible
answers: no, yes, and “sort of.” I will argue that each answer encounters insuperable
problems. I address the “yes” and “no” answers in this section, turning to the (more
complex) “sort of ” answer in the next section.
Start with the first option: movements produced by partially unified agents do not
count as actions. Then we haven’t solved the bad action problem at all: the only
genuine actions are the ones that are done in accordance with the CI. So we can
dispense with this option straightaway.
Suppose, second, that movements produced by partially unified agents do count as
actions. We originally presented Korsgaard’s argument as follows:
(a) An agent’s A-ing is an action iff A-ing is attributable to the agent as a unified
whole.
(b) An action is attributable to the agent as a unified whole iff the principle of the
agent’s action renders the agent capable of diachronic stability.
(c) The CI is the only principle which renders the agent capable of diachronic
stability.
(d) Therefore, an agent’s A-ing is an action iff the principle of the agent’s A-ing is
the CI.
However, if movements produced by partially unified agents do count as actions, then
we need to modify (a). In particular,
BAD AC TION 101
(a*) An agent’s A-ing is an action iff A-ing is attributable to the agent as a partially
unified whole.
An agent will count as partially unified if the agent acts on principles that produce
partial, but not total, diachronic stability. Examples include the democratic principle,
the timocratic principle, and so on. In light of this, we will also need to rewrite (b) as
follows:
(b*) An action is attributable to the agent as a unified whole to the extent that the
principle of the agent’s action renders the agent capable of diachronic stability.
We can keep (c) as is: the agent will act as a fully unified whole iff she acts on the CI,
which provides complete diachronic stability
The problem is that (d) does not follow from (a*), (b*), and (c). Korsgaard has
claimed that a variety of principles generate partial unity. If performing an action
merely requires being partially unified, then performing an action does not require
acting on the CI. Put simply, if all that it takes to perform an action is to act on a
principle that unifies the agent to some extent, then our commitment to action yields
only a commitment to acting on principles that unify us to some extent, not to acting on
principles that unify us to the fullest extent. (We might wonder whether Korsgaard can
somehow establish that aiming at partial unity commits us to aiming at full unity.
I address this possibility in Section 5.4, below.)
I submit that it is a failure to distinguish simple and differentially realizable consti-
tutive features that leads Korsgaard to think that her argument succeeds. For, if all
constitutive features were differentially realizable, then her conclusion would follow.
In other words, if showing that we must aim at some degree of unity sufficed to show that
we must aim at full unity, then Korsgaard’s conclusion would follow. However, once
we recognize the possibility of simple constitutive features, it becomes apparent that
Korsgaard’s argument is unsuccessful.
In sum, then, if Korsgaard gives the “yes” answer, her theory of action establishes
merely that we are committed to acting on principles that unify us to some extent. It
does not establish that we are committed to acting on the CI, the principle that fully
unifies us. So, Korsgaard’s theory simply tells us to act on some principle or other, but
does not provide us with a way of adjudicating between competing principles. In
particular, it fails to establish that we have any more reason to act on the CI than to act
on any competing principle, such as a principle of self-interest.
5.3 Does an appeal to degreed action help?
So far, I have argued that Korsgaard faces a problem when we ask whether movements
springing from partially unified agents count as actions. If she answers “no,” she is
unable to account for bad action; if she answers “yes,” she is unable to establish that
action’s constitutive principle is the CI. This might seem to exhaust the possibilities.
However, there is actually a third option: instead of answering “yes” or “no,” we can
answer “sort of.” Korsgaard writes,
an action that is less successful at constituting its agent is to that extent less of an action. So on this
conception, ‘action’ is an idea that admits of degrees. An action chosen in a way that more
successfully unifies and integrates its agent is more authentically, more fully, an action, than one
that does not. (Korsgaard 2009, 25)
More briefly, Korsgaard writes, “action, as I conceive it, is something that comes in
degrees” (Korsgaard 2009, 163). For example, when Tom takes a drink of water, this
may be more (or less) of an action than when Tom, later that night, takes a drink of beer.
Although the idea that action comes in degrees is no doubt puzzling and controversial,
here I will grant the cogency of the claim and simply ask whether it helps Korsgaard
avoid the problem that I have raised.
The claim that action comes in degrees might seem to give Korsgaard the resources
to respond to my objection. For suppose we start with the following claim:
(a**) An agent’s A-ing is an action to the extent that the principle of the agent’s
A-ing constitutes the agent as a fully unified whole.
And suppose we can show that the CI is the only principle acting upon which
constitutes the agent as a fully unified whole. Then we could conclude that:
(d**) An agent’s A-ing is an action to the extent that the principle of the agent’s
A-ing is the CI.
In short: while less-than-full-fledged action requires only a low degree of unity, full-
fledged action requires full-fledged unity. Full-fledged action requires acting on the
Categorical Imperative.
While this strategy might initially seem to help Korsgaard, attention to the way in
which the conclusion—claim (d**)—differs from claim (d) will reveal a problem.
Originally, the argument purported to establish that action requires that the agent act on
the CI. Now, the argument establishes only that full-fledged action requires that the agent
act on the CI. Unfortunately, the appeal to full-fledged action simply pushes the
problem back to a different point in the theory. My argument in the previous section
took this form: if all that it takes to perform an action is to achieve some degree of
unity, then the fact that we are committed to action does not imply that we are
committed to full unity. Suppose Korsgaard responds to by claiming that if you aim at
less than full unity, you will be performing something less than a full-fledged action. In
order for this response to have purchase, Korsgaard would then have to show that we
are committed to performing full-fledged actions, rather than less-than-full-fledged
actions. But no such premise is employed in the above argument.
Put differently, once Korsgaard introduces a degreed conception of action, an agent
can legitimately ask: why should I perform a full-fledged action, rather than a less-than-
full-fledged action? This is not an idle question, for according to Korsgaard’s account
BAD AC T ION 103
I can do all sorts of things without performing a full-fledged action: I can act on my
strongest desire, I can act on feelings of natural beneficence rather than duty, I can act
on self-interested prudence, and so forth. (Recall that any action that is not in
accordance with the CI will count as less than full-fledged.) Given this fact, what is
supposed to commit me to aiming at full-fledged action rather than these sorts of less-
than-full-fledged actions? Rather than answering this question, the introduction of a
degreed conception of action simply postpones it.
In sum: we can take it for granted that agents are committed to performing actions;
after all, what it is to be an agent is to perform actions. However, once we introduce a
distinction between full-fledged and less-than-full-fledged actions, we cannot take it
for granted that agents are committed to performing full-fledged actions.18
Thus, there is a common problem in both of Korsgaard’s answers. If she says that
events springing from partially unified agents are actions, then she would need to show
that we are committed to aiming at full unity. If she says that events springing from
partially unified agents are partial actions, then she would need to show that we are
committed to performing full actions.
5.4 Can Korsgaard establish a differentially realizable aim?
Section 5.2 argued that claim (a*)—that an agent’s A-ing is an action iff, in A-ing, the
agent acts on a principle that constitutes her as unified to some extent—entails only a
simple constitutive principle of self-constitution. Section 5.3 argued that we cannot
resolve this problem by appealing to a degreed conception of action. However, an
objection might be raised on Korsgaard’s behalf: isn’t it unfair to interpret Korsgaard as
moving directly from claim (a*) or (a**) to a conclusion about action’s constitutive
principle? Put differently, even if claim (a*) or (a**) by itself entails only a simple
constitutive principle, doesn’t Korsgaard have other arguments that, when taken with
(a*) or (a**), jointly entail the presence of a differentially realizable principle?
First, an interpretive point: Korsgaard does in fact present her argument as moving
directly from claim (a*) or (a**) to a conclusion about constitutive principles. Here is
how she summarizes her argument:
18
My objection is related to the point that David Enoch raises in his (2006). As I pointed out in Chapter 2,
Enoch argues that even if action has a constitutive aim of (say) self-constitution, it would be open to an agent
to respond: “I am perfectly happy performing schmactions—nonaction events that are very similar to actions
but that lack the aim (constitutive of actions but not of schmactions) of self-constitution” (2006, 179). In
other words, the agent could escape the reasons engendered by action’s constitutive aim by performing
schmactions instead of actions. In Chapter 2, I argued that if Enoch’s objection is to have any purchase, he
needs to establish that performing schmactions is a real possibility for human agents. With this in mind, notice
that my objection to Korsgaard’s degreed conception of action is, in some ways, an application of Enoch’s
point: I claim that even if full-fledged action has the constitutive aim of full unity, less-than-full-fledged
action doesn’t have this aim; it aims merely at any degree of unity. Moreover, we already know that performing
what Korsgaard is calling less-than-full-fledged action is a real possibility: we know that we can act on our
strongest desire, on a motive of natural beneficence, and so on. So, unlike the claims about schmaction,
Korsgaard will not be able to deny that we can perform less-than-full-fledged actions. The introduction of a
degreed conception of action is therefore unhelpful.
What is an agent? An agent is the autonomous and efficacious cause of her own movements . . .
So the constitutive standards of action are autonomy and efficacy, and the constitutive principles
of action are the categorical and hypothetical imperatives. It’s also true that in order to be
autonomous, it is essential that your movements be caused by you, by you operating as a unit, not
by some force that is working in you or on you. So in order to be an agent, you need to be
unified—you need to put your whole self, so to speak, behind your movements. That’s what
deliberation is: an attempt to reunite yourself behind some set of movements that will count as
your own. And in order to reunite, you have to have a constitution, and your movements have
to issue from your constitutional rule over yourself. (Korsgaard 2009, 213)
Here, Korsgaard moves from a claim about the nature of action—that it requires from a
unified agent—to a claim about action’s constitutive principle: the agent must be
unified, and hence must aim in deliberation at performing an action that unifies him.
But let’s set this interpretive point aside and ask whether Korsgaard might be able to
appeal to supplementary principles that would take us from her account of action
(claim a* or a**) to her conclusion that action’s constitutive principle is the Categorical
Imperative. Suppose we start with claim (a*):
(a*) An agent’s A-ing is an action iff the principle of the agent’s A-ing constitutes
the agent as unified to some extent.
If we could supplement (a*) with a principle of the following form, then it would
entail commitment to the Categorical Imperative:
(S1) If an agent is committed to acting on principles that unify her to some extent,
then the agent is committed to acting on principles that unify her maximally.
If (S1) could be established, then Korsgaard could indeed establish that action’s consti-
tutive norm is differentially realizable.
Analogously, suppose we start with claim (a**):
(a**) An agent’s A-ing is an action to the extent that the principle of the agent’s
A-ing constitutes the agent as a fully unified whole.
If we start with this premise, we would need a principle of the following kind:
(S2) If an agent is committed to performing partial actions, then she is committed to

performing full-fledged actions.
Again, (a**) and (S2) would enable Korsgaard to establish that action’s constitutive
principle is the Categorical Imperative.
The question, then, is whether Korsgaard can establish (S1) or (S2). Korsgaard does
offer arguments that address this point. Consider the following passage, in which she
claims that Plato shows that the just life is the only life we can choose.
Plato’s argument shows that this aristocratic constitution is the only one you can choose. For you
can’t, in the moment of deliberative action, choose to be something less than a single unified
BAD AC TION 105
agent. And that means you can’t exactly choose to act on any principle other than the principle of
justice. Timocratic, oligarchic, and democratic souls disintegrate under certain conditions, so
deciding to be one would be like making a conditional commitment to your own unity, to your
own personhood. And that’s not possible. You can be a timocratic, oligarchic, or democratic
person, in the same way that you can be a just person who fails on the rack. But you cannot
decide in advance that this is what you will be. (Korsgaard 2009, 180)
In this passage, Korsgaard claims that an agent cannot choose to be something less than
a maximally unified agent. That sounds just like premises (S1) and (S2), above. If you
choose to be an agent, eo ipso you choose to perform full-fledged or maximally unified
acts. For, Korsgaard claims, you simply cannot choose to do otherwise. You cannot
choose to perform a less-than-full-fledged or less-than-maximally-unified act.
So Korsgaard wants something like (S1) or (S2). But does she have any argument for
these claims? I will suggest that she does not.
To see this, consider Korsgaard’s claim that I cannot choose to perform a less-than-
full-fledged act or to be a less-than-maximally-unified agent. What is the sense of
“cannot,” here? Obviously, some agents do become less than maximally unified agents.
For example, some agents do choose to act on the principle “I will do the honorable
thing,” and thereby qualify as “timocratic souls.” Timocratic souls are less than fully
unified, and therefore perform less than full-fledged action. So it certainly seems that
agents can choose less-than-maximal unity, or less-than-full-fledged action.
In the face of this, we might interpret Korsgaard in two ways:
– The agent can choose less-than-maximal unity (or less-than-full-fledged action),
but if he does he will be violating some normative principle to which he is
committed.
– It is literally impossible for the agent to choose less-than-maximal unity (or less-
than-full-fledged action).19
We can illustrate the difference between these two readings by considering the
example of doing the honorable thing (which, as I’ve pointed out above, counts as
manifesting less-than-maximal-unity). On the first reading, the agent can choose to do
the honorable thing, but if he does he will be violating some normative principle to
which he is committed. On the second reading, it is literally impossible for an agent to
choose to do the honorable thing. Which of these claims does Korsgaard accept?
Suppose she accepts the first option. Straightaway, a problem arises: rather than
explaining why we must aim at maximal unity, this claim simply asserts that we must aim
at maximal unity. Put differently, we were looking for a principle that would enable us
to argue that if we aim at some degree of unity, we must aim at maximal unity. In this
19
These distinctions are indebted to Fitzpatrick, who argues that an earlier version of Korsgaard’s theory
conflates “literal practical necessity” and “commitment.” X is literally practically necessary if without X-ing
we cannot function as agents; X is a commitment if it is not literally practically necessary to X, but we are
committed to X by other things that are literally practically necessary (Fitzpatrick 2005, 668).
context, stipulating that we are committed to aiming at maximal unity is circular. So

the first option cannot be what Korsgaard has in mind.
Let’s consider the second option. On the face of things, this option might appear
incredible: after all, there are agents who choose to do the honorable thing all the time,
who make this a way of life. Nonetheless, Korsgaard sometimes appears to deny that
such agents can be regarded as genuinely choosing these courses of action. For
example, she claims that appearances notwithstanding, “the tyrannical person doesn’t
really choose actions” (Korsgaard 2009, 172). Presumably, the same goes for the
timocratic agent.
To see what Korsgaard has in mind, compare the case of belief and truth: while there
are many agents who believe falsehoods, there does seem to be a sense in which it is
literally impossible for an agent to choose to believe falsehoods. For, in recognizing that
P is false and then choosing to believe P, the agent would be feigning belief rather than
genuinely believing. We might describe this fact by saying that it is literally impossible
for an agent to choose to believe falsehoods. Korsgaard is suggesting that in an
analogous way, it is literally impossible for an agent to choose less-than-maximal unity.
If Korsgaard could establish that it is literally impossible to choose less-than-maximal
unity (or less-than-full-fledged action), her argument would succeed. However, there
are two problems with this approach.
First, while the claim that it is literally impossible for an agent to choose to believe
falsehoods is plausible, the claim that it is literally impossible for an agent to choose less-
than-maximal unity (or less-than-full-fledged action) is considerably less so. For this
would mean, among other things, that it is literally impossible to choose to act on your
strongest desire. In order to maintain this claim in the face of the overwhelming
evidence that agents choose to act on their strongest desires all the time, we must
give “choice” some special sense: we must decline to label as “choice” those cases in
which an agent appears to choose to act on his strongest desire. Once we have done
that, though, an agent can legitimately question why she should care whether her
apparent choices merit the honorific “genuine choice.” Consider an example: last
night, I thought I should grade some student essays, but had an intense desire to spend
the night watching a good movie instead. I thought about the options, judged that
there was more reason to grade the papers, and nonetheless decided to act on my
strongest desire (watching the movie). What is it to me if this decision doesn’t deserve
the label “choice”?
In sum, the first problem is this: maintaining that agents cannot choose to act on
their strongest desires requires that we employ some idiosyncratic, robust notion of
choice. Once we do so, it is unclear why the agent should care about manifesting that
kind of choice. (This is exactly analogous to the problem canvassed in Chapter 2,
Section 2.2: if someone claims that the notion of driving contains, as part of its content,
the idea of driving lawfully, then of course they can derive normative content from that
notion; but we are not inescapably committed to driving in that sense, so the
conclusion is not interesting.)
S U M M A RY 107
This brings us to the second and more decisive problem. If such an argument
succeeded, it would entail that it is literally impossible for an agent to do anything
but choose to act in the correct way. The claim that it is literally impossible to choose
anything other than an action that maximally unifies us, when coupled with the claim
that actions that maximally unify us are the good actions, entails that it is literally
impossible to choose to perform a bad action. So employing this claim would recreate
the problem raised in Section 5.1: it would render the theory trivial, by making the
theory incapable of distinguishing between good actions and bad actions.
I conclude that Korsgaard cannot establish any supplementary principles that would
takes us from her conception of action (claim a* or a**) to the conclusion that action’s
constitutive principle is the Categorical Imperative. Accordingly, Korsgaard’s consti-
tutivist account is unsuccessful.
6. Summary
I have argued that once we distinguish between simple and differentially realizable
features, we see that Korsgaard’s argument for constitutivism is unsuccessful. For
Korsgaard’s theory encounters a dilemma. On the one hand, she could argue that
A-ing is an action only if it fully unifies the agent. This establishes a commitment to the
Categorical Imperative, but entails that all actions are good actions: if you fail to act on
the Categorical Imperative, you fail to produce an action at all. On the other hand,
Korsgaard could (and does) argue that A-ing can be an action even if it merely
generates some degree of unity. Yet, once we distinguish between simple and differen-
tially realizable features, this account of action fails to yield a commitment to realizing
full unity (i.e., acting on the Categorical Imperative); all that it yields is a commitment
to realizing some degree of unity. As a result, Korsgaard’s theory fails to generate any
substantive conclusions about practical reason. It requires us to act on a principle that
unifies us to some extent, but Korsgaard herself tells us that a host of mutually
inconsistent principles unify us to some extent. These principles range from self-
interest, to acting on one’s strongest desire, to honor, to the Categorical Imperative.
Korsgaard’s theory of action gives us no reason to prefer one of these principles to any
of the others.
This points to a pervasive problem for constitutivist theories. The constitutivist seeks
to show that actions have some feature, F, which both makes them actions and makes them
good actions. In order to account for the possibility of bad action, the constitutivist needs
to open a gap between actions without feature F (the good-making feature) and actions
with feature F. The only way to do this, while preserving the initial claim that all actions
have feature F, is to claim that F comes in degrees, and is possessed to different extents
by different actions. The constitutivist can then claim that good actions manifest a high
degree of F, whereas bad actions manifest lesser degrees of F. However, the constitu-
tivist then faces the question of why we should aim at manifesting actions with the
highest degree of feature F, given that we could perform actions with low degrees of
F without thereby ceasing to act. The failure to distinguish simple and differentially
realizable features makes this problem seem to disappear: if all aims were differentially
realizable, then the mere fact that we aim at manifesting some degree of F would entail
a commitment to manifesting the highest degree of F. But, once we draw the
distinction, the problem is much less tractable. Rather than establishing merely that
we aim at manifesting some degree of F, the constitutivist needs to establish that, in
every action, we aim at manifesting the highest degree of F. Neither Korsgaard nor
Velleman succeeds in showing this.
5
Action’s First Constitutive Aim:
Agential Activity
The previous chapters have investigated the problems and prospects for constitutivism.
A successful version of constitutivism must locate some feature of action whose
presence entails substantive normative conclusions. I have argued that neither of the
current constitutivist theories succeeds in doing so. Even if we grant Velleman and
Korsgaard their accounts of action, these accounts don’t yield substantive conclusions
about practical reason. Velleman’s theory is unsuccessful because even if we could
show that each action aims at self-knowledge, it would not follow that we have more
reason to seek actions that yield more comprehensive self-knowledge. Likewise,
Korsgaard’s theory is unsuccessful because even if we could show that self-constitution
is action’s constitutive principle, this would not entail that we have more reason to
perform actions that more fully unify us. In short, neither theory generates substantive
normative conclusions, in part because neither theory employs a differentially realiz-
able constitutive feature.
However, these results do not indicate that constitutivism as such is a failed project.
Rather, they simply generate criteria of adequacy for a successful constitutivist theory.
First, the theory must employ a differentially realizable constitutive feature; second, the
theory must generate conclusions about weights of reasons. The following chapters
develop a constitutivist theory that does meet these criteria. This chapter begins that
task by defending the account of action upon which the constitutivist theory will be
based.
Constitutivism attempts to extract normative content from the bare concept of
action. Of course, the more content that we build into our conception of action, the
easier it will be to show that action contains constitutive standards. I drew attention to
this point in Chapter 2, using the example of driving: if we define driving as operating a
vehicle lawfully, then it will be straightforward to extract normative content from the
account of driving. But this would be of no interest, for the account of driving is
completely implausible; of course the bare idea of driving doesn’t commit us to driving
lawfully. The same goes for action: if we build a great deal of content into our starting
conception of action, then we will have an easy time extracting normative content. But
this will be of no interest, for no one will have reason to grant us the initial conception
of action.
110 AC T IO N’ S FI R S T C O N S T I T U T IV E AI M: AGE N T I A L AC T I V I T Y
In short: a constitutivist theory of practical reason can only be as convincing as the

account of action upon which it is based.1 The most powerful version of constitu-
tivism, then, would start with a minimal, universally accepted account of action. It
would show that this minimal account of action, which is accepted by everyone, has
substantive normative consequences.
I will not be able to do that; I cannot show that a completely uncontroversial account
of action yields substantive normative content. (Indeed, it is not even clear that there is
a completely uncontroversial account of action.) But I will try to show that we can start
with a fairly minimal account of action, an account which—though not universally
accepted—is very widely accepted even by those who do not themselves have con-
stitutivist sympathies. I will argue that we have good reason for accepting this account
of action. I will then show that the account yields substantive normative content. If my
arguments succeed, then a widely accepted account of action yields a constitutivist
theory of practical reason.
This chapter comprises five sections. Section 1 introduces the widely accepted claim
about action, the claim that will be integral to my constitutivist account: we can
distinguish between activity and passivity in action. I introduce the notions of activity
and passivity, and argue that the nature of reflective agency commits us to accepting
that there is some such distinction. Most philosophers who have argued for a distinc-
tion between activity and passivity have associated agential activity with reflective or
deliberative activity. Section 2 asks why one would think that there is an essential
connection between agential activity and reflective activity. I begin by teasing apart
three distinct claims about reflective agency, which have not been clearly distinguished
in the literature: that choice causes action, that motives do not determine choice, and
that reflection suspends the effects of motives. Sections 2 and 3 assess these three claims,
arguing that while the first and second claims are true, there are philosophical argu-
ments and results from empirical psychology indicating that the third claim is false.
Section 4 argues that the third claim is the crucial one; its truth is necessary in order to
support the idea that reflective agency is paradigmatically active. Thus, I maintain that
once we distinguish these three claims, and see that the third claim is false, we discover
that our assumptions about reflective agency require modification and perhaps even
rejection. Reflective acts are not necessarily more active than unreflective acts.
Accordingly, the distinction between activity and passivity must be rethought. In
light of these results, Section 5 introduces and defends a new account of agential
activity. This account claims that an agent is active iff two conditions are met: (i) the
1
In the previous chapters, I granted Korsgaard and Velleman their accounts of action, in order to show
that these accounts of action did not yield substantive normative conclusions. Of course, there is a more
common form of objection to Korsgaard and Velleman: many have thought that their theories of action are
problematic. Korsgaard’s Kantian theory of action has seemed, to many, unduly reflective or intellectualistic;
Velleman’s self-knowledge based theory has seemed to reverse the knowledge–action relationship. I have not
pursued this kind of objection. For reasons that will become apparent over the course of this chapter, I do not
think that the Korsgaardian and (especially) Vellemanian points about action can be so easily dismissed; there
is something correct in each of these accounts.
AGE N T I A L AC T I V I T Y A ND I T S A L L E G E D C O N N E C T I O N 111
agent approves of her action, and (ii) further knowledge of the motives figuring in the
etiology of this action would not undermine her approval of the action. This account
of activity preserves what is right about Korsgaard’s and Velleman’s theories. To
anticipate, there is a sense in which Velleman is right that action aims at self-
knowledge. And there is a sense in which Korsgaard is right that action aims at
autonomy. But the role of self-knowledge and the nature of autonomy turn out to
be very different than those envisioned by Korsgaard and Velleman. Finally, Section 6
uses the above results to argue that agential activity is a constitutive aim of action.
1. Agential activity and its alleged connection

to deliberation
This section introduces the idea of activity and passivity in action. The idea that there is
some such distinction is widely accepted, but the particular way in which we should
account for the distinction has been hotly debated. I will criticize certain accounts
while refining a new account. Accordingly, my hope is that by the end of this chapter
even those who were initially skeptical of the idea of agential activity will find reasons
for accepting the notion.
1.1 Activity and passivity in action
Action theorists pursue two different tasks. On the one hand, some action theorists seek
to explain the distinction between actions and undergoings. Consider the following
movements: my muscle spasms; my hair grows; my heart beats in order to distribute
blood; my pupils dilate in order that I may see. These are all changes that, in some
sense, I effect. In addition, some of these events are purposive movements. But there
seems to be an important distinction between these events and genuine actions, such as
walking into the kitchen in order to get a cup of coffee, typing in order to write a
paper, or deciding where to take my next vacation. The events in this latter set seem
attributable to me, to my own activity, in a quite different way than do dilations of my
pupils or the growth of my hair. A central task in action theory is explaining this
distinction between actions and mere undergoings.
However, some action theorists undertake a second and more ambitious task. They
seek not only to distinguish actions from other events, but also to grade actions as more
and less exemplary members of their kind. On this view, it makes sense to speak of
“paradigm” or “full-fledged” actions, for there is some essential feature that is fully
manifest in some actions, and less completely manifest in other actions.2,3 The under-
lying idea can be brought out by considering a few cases:
2
Philosophers endorsing this second approach include Harry Frankfurt (1988, 1999, 2004), Christine
Korsgaard (1999, 2009), Joseph Raz (1999), David Velleman (2000), Jay Wallace (2006), and Michael
Bratman (2007).
3
The two approaches to action theory may coincide, in the end. There may be some feature of action that
is realizable to different degrees, so that at the highest degrees of fulfillment we have the paradigmatic, full-
fledged actions, whereas at the lowest degrees, acting shades off into mere undergoing.
Abe is walking along a dark path at night. Out of the corner of his eye, he suddenly sees a shape
lurching through the trees, and he recoils.
Bob has been at the party for a while, and he is trying to decide whether to have another drink.
He reflects on his reasons for drinking, as well as his reasons for declining to drink. He recognizes
that if he has another drink, he won’t be able to drive home in a safe and responsible fashion.
Accordingly, he decides that he will not drink. Nevertheless, when his friend hands him a drink a
moment later, he ends up drinking after all.
Claire is torn between her need to study for the exam that she has tomorrow and her strong
desire to go out to the party tonight. She reflects on her motives for each course of action and
decides that she has most reason to stay at home and study. She stays at home and studies.
These three agents seem to play progressively more active roles in the production of
their movements. Abe doesn’t seem to play much of a role in the production of his
movement; his recoiling seems to be something that he undergoes, something that
happens to him, rather than something that he performs. (On most accounts, this
behavior will not even qualify as an intentional action.) Bob and Claire, on the other
hand, certainly play some role in the production of their actions, but there seem to be
important differences between them. Bob endeavors to act in a certain way, but finds
himself yielding to or being overpowered by his desires. Claire also experiences strong
desires to perform a certain action, but she resists these desires and acts in accordance
with her decision. Accordingly, there is a sense in which Claire plays more of a role in
the production of her action than does Bob. These examples illustrate one way in
which we might wish to distinguish actions that the agent plays more and less active
roles in producing.4
Harry Frankfurt drew attention to these kinds of case in a series of articles beginning
in the 1970s. He offered examples of the following sort: a drug addict wants the drug,
but also wants not to want the drug; nonetheless, the desire for the drug moves him to
action. We can contrast this with a drug addict who wants the drug and wants to want
the drug; he is satisfied with his addiction. The actions of these agents seem importantly
different: the contented addict’s action appears active in a way that the struggling
addict’s action does not. The idea, here, is that in addition to distinguishing between
action and non-action, we must distinguish two different types of actions: those with
which the agent “identifies” or those which the agent performs “wholeheartedly,” as
in the case of the contented addict, and those with which the agent does not identify or
which the agent performs while experiencing inner division, as in the case of the
struggling addict.
4
Of course, I do not intend to suggest that invoking a notion of agential activity is the only way in which
we can mark the distinctions between these actions. There are alternatives. For example, one relevant
distinction is that Bob’s action is akratic, whereas Claire’s is not. Another relevant distinction is that Abe’s
act is not intentional, whereas Bob’s and Claire’s acts are intentional.
AG E N T I A L AC T I V IT Y A N D I T S A L L E G E D C O N N E C T I O N 113
A number of philosophers endorse Frankfurt’s general claim that we must distin-

guish between two types of action, but develop different analyses of the distinction.
Consider a few examples. Michael Bratman writes,
When a person acts because of what she desires, or intends, or the like, we sometimes do not
want to say simply that the pro-attitude leads to the action. In some cases, we suppose, further,
that the agent is the source of, determines, directs, governs the action and is not merely the locus
of a series of happenings, of causal pushes and pulls. (Bratman 2007, 91)
Bratman here claims that we need to distinguish actions that the agent actively
produces from actions that spring from the agent in a more passive fashion. David
Velleman concurs, claiming that “full-fledged” or “paradigm” actions cannot simply
be behaviors that are caused by a belief and a desire, for the process so described
fails to include an agent. In this story, reasons cause an intention, and an intention causes bodily
movements, but nobody—that is, no person—does anything. Psychological and physical events
take place inside a person, but the person serves merely as an arena for these events: he takes no
active part. (Velleman 2000, 123)
Jay Wallace claims that full-fledged agency requires “reflective self-control,” which is
manifest in “choices” and “decisions” (2006, 149). Analogously, Brian O’Shaughnessy
claims that an agent is active when the etiology of her action includes an act of trying or
willing: “the internal constituent of action . . . is trying or (what we might as well call)
willing” (O’Shaughnessy 1973, 367). Joseph Raz distinguishes between activity and
passivity in action in terms of whether the agent’s action is responsive to reasons:
We are active when our mental life displays sensitivity to reasons, and we are passive when such
mental events occur in a way that is not sensitive to reasons . . . (Raz 1999, 11)
And the examples could be multiplied.5

As these examples indicate, a number of philosophers pursuing very different
projects endorse the idea that we need a distinction between two kinds of action.
The distinction is described in different ways: it has been characterized by turns as
identification, wholeheartedness, guidance by the agent, direction by the agent, agential control,
agential activity, reflective self-control, rational control, and so forth. All of these terms have
different connotations, and are analyzed in different ways. But many of the theorists
who employ these terms agree on one point: as the quotations above indicate, these
terms are meant to characterize the idea that agents can be more and less active in the
production of their own actions. Accordingly, I will use “agential activity” as the most
general term for notions that are meant to pick out the agent’s contribution to the
production of action. Agential activity is a genus whose species are notions such as
5
See, for example, Melden (1961), Foley (1977), Thalberg (1978), Ginet (1990), Korsgaard (1996b, 2009),
Ruben (1997), Schapiro (2001), Mele (2003, 2009), Hornsby (2004), Watson (2004), and Zhu (2004).
agential control, agent causation, guidance, direction by the agent, self-government, self-control,
choice, decision, and so on. Then we can say that the philosophers who take up the
second task of action theory, the philosophers who grade actions as more and less
exemplary members of their kind, are claiming that different actions manifest different
degrees of agential activity. To return to our initial examples: Claire is presumably
the most active member of the lot; Bob is somewhat less active; Abe may not be active
at all.6
1.2 What is agential activity? Activity, reflection, and deliberation
So far, we have seen that many philosophers accept the notion of agential activity. Of
course, this leaves open the question of whether they should. Do we really need such a
notion? I will argue that we do; we cannot offer a satisfactory account of agency
without distinguishing between activity and passivity in action. In order to make this
case, though, we first need to say more about what agential activity is.
Philosophers who accept the notion of agential activity almost inevitably associate
agential activity with reflective, self-conscious, or deliberative activity. As noted above,
Velleman claims that deliberative actions are “paradigm cases of action” (Velleman
2000, 124), and Wallace agrees, telling us that full-fledged agency is manifest in
“choices” and “decisions” (Wallace 2006, 149). Wallace continues,
Our motivations divide fundamentally into states of two different kinds. There are, first,
motivations with respect to which we are basically passive, such as conscious desires, inclinations,
yearnings, and various long-term dispositions . . . Second, there are motivations that are not
merely given, but that directly express our activity as agents, such as choices, decisions, and
intentions to act. (Wallace 2006, 149)
Here, Wallace claims that phenomena such as choices and decisions are direct expres-
sions of the agent’s activity.
Many philosophers endorse similar ideas. Christine Korsgaard suggests that an agent
is active when her action is produced by an episode of reflective, self-conscious
deliberation. According to Korsgaard, “when you deliberate, it is as if there were
something over and above all of your desires, something which is you, and which
chooses which desire to act on” (Korsgaard 1996b, 100). In other words, when
I deliberate, I reflect on my potential actions, and choose to perform one of them.
I experience my choice as expressive of me, attributable to me, in a way that my desires
need not be. Richard Foley agrees, writing that “it seems clear that actions which are
preceded by deliberation are paradigms of action” (1977, 59). Finally, to cite just one
more example, Gary Watson claims that “practical decision, ‘deciding to’, is an active
phenomenon” (Watson 2004, 140).
6
To forestall a potential objection, notice that every action will be active in some sense. Even Abe will
have neural activity, muscle activity, and so forth. But this is not the type of activity in which we are
interested. Put differently: we are trying to introduce a sense of activity enabling us to say that just as not every
bodily change counts as an action, so too not every action counts as active.
T E A S I N G A PA RT S E V E R A L D I S T I N C T C L A I M S 115
Here we have a number of philosophers who are engaged in quite different projects,
and who operate with quite different sets of assumptions, yet who agree that actions
produced by deliberation and reflective choice are paradigmatically active.7 In the
following sections, I will investigate this claim. I will argue that this agreement is
founded on a mistake.
In order to produce this argument, it would be easiest if I could simply state, as a
definition, “agential activity is X.” I could then argue that deliberative activity is not X,
and my proof that agential activity and deliberative activity come apart would be
complete. Unfortunately, I cannot do that. Any such definition of agential activity
would be merely stipulative. To see what is meant by agential activity, we need to
investigate the way in which deliberative acts can differ from non-deliberative ones.
So, my strategy is this: I will ask what makes deliberative acts appear different, in a
philosophically relevant sense, from non-deliberative acts. I will then argue that
deliberative acts aren’t really different from non-deliberative acts in that way. How-
ever, I will then show that part of what we want to say about deliberative acts is
tracking a deeper point: a point about the nature of agential activity. In short, seeing
that and how the association of deliberative activity with agential activity is mistaken
will lead us to a new account of agential activity.
2. Teasing apart several distinct claims about

reflective agency
2.1 The components of deliberative agency
To see why philosophers tend to assume that deliberative agency is paradigmatically
active, let’s begin by analyzing the notion of deliberation. The philosophers quoted
above believe that there are two ways in which action can be brought about. Some
actions are directly caused by the agent’s motives; in these cases, episodes of choice are
either absent or causally inert.8 Other actions are caused by the agent’s choice.9 This
view is explicit in Wallace, who writes that the “will,” or the agent’s capacity for
choice, is “a capacity for active self-determination” (2006, 149).
In short, these philosophers assume that in normal circumstances, by choosing to A,
an agent makes it the case that she will A. Of course, this isn’t always true. Sometimes,
7
To be clear, not all of these philosophers think that deliberation is a necessary condition for agential
activity. Velleman, for example, argues that agential activity can be manifest in unreflective, skillful actions,
such as playing a piano or riding a bicycle. However, all of these philosophers maintain that deliberative
action is a paradigmatic case of agential activity, for two reasons. First, in ordinary conditions, deliberative
action is sufficient for agential activity; if the agent deliberates and acts in accordance with this deliberation, she
will manifest agential activity. Second, these philosophers claim that it is by analyzing the structure of
deliberative agency that we will understand other types of agential activity (cf. Velleman 2004b, 281–2).
8
I will use the term “motive” in a very encompassing way, to refer to any folk-psychological state that can
cause action. Examples include desires, affects, moods, and emotions.
9
That is, the episode of choice is one of the causes of the action. It need not be the only cause.
my choices are overpowered by strong motives. I may choose to have only one drink at
the party, and find myself unable to refuse a second or a third. I may decide that I need
to spend three hours grading student papers, and find myself distracted by another task
within the first hour. So, strong motives sometimes seem to trump choice. Addition-
ally, the world can thwart my attempt to carry out my choice: I may choose to eat
dinner at a favorite restaurant, yet arrive to find the restaurant closed. I may choose to
drive to my office, yet find that my car won’t start. In each case, the world foils my
chosen course of action. However, these cases are atypical: in each case, we need to
provide an explanation of why the choice didn’t issue in the chosen action. Either the
motives are unusually strong and overpowering, or the world is unexpectedly recalci-
trant. Thus, these cases support the claim that in the usual case, we are capable of acting
as we choose to act. If I choose to eat a grapefruit for breakfast, I will probably end up
eating it; if not, we need some explanation of why my choice did not eventuate in
action. Thus, the following claim seems true:
(Choice) Typically, if I am faced with two actions that it is possible for me to

perform, A-ing and B-ing, and I choose to A, then I will A.10,11
How does the agent’s choice relate to her motives? In other words, if choice typically
determines action, how do motives relate to choice?
In the tradition, we can find two views: either motives determine what we choose,
or they do not. According to the first view, while a mental episode of choosing or
deliberating may occur in me, the course of these events, and their outcome, are
determined by the motives that I have.12 The philosophers who associate agential
activity with deliberative action reject this picture, instead adopting the following
view:
10
Two notes. First, I intend Choice to apply to actions in the present; if we consider decisions about
future actions, further complications will arise. Second, it is clear that difficulties would arise if we attempted
to spell out what is meant by “actions that it is possible for me to perform.” As the focus of this chapter is
elsewhere, I will not delve into these issues. All that matters, for our purposes, is that there are cases of the
following sort: it is possible for me either to have grapefruit or cereal; I choose to have grapefruit; this choice
eventuates in my having grapefruit.
11
For philosophical arguments and empirical evidence in support of the claim that conscious decisions
typically cause actions, see Mele (2009, Chapter 7). Mele writes, “there is powerful evidence for the truth of
the following thesis: the fact that an agent consciously decided to A or had a conscious intention to
A sometimes has a place in a causal explanation of the corresponding overt intentional action” (Mele
2009, 144).
12
A version of this view is defended by Hobbes, who writes, “When in the mind of man appetites and
aversions, hopes and fears, concerning one and the same thing, arise alternately; and diverse good and evil
consequences of the doing or omitting the thing propounded come successively into our thoughts; so that
sometimes we have an appetite to it, sometimes an aversion from it; sometimes hope to be able to do it,
sometimes despair, or fear to attempt it; the whole sum of desires, aversions, hopes and fears, continued till the
thing be either done, or thought impossible, is that we call deliberation” (Leviathan VI). He continues, “In
deliberation, the last appetite, or aversion, immediately adhering to the action, or to the omission thereof, is
that we call the will . . . ” (Leviathan VI). For a more sophisticated version of this position, see Schopenhauer
(1999, 36–7).
(Non-determination) Choice is not determined by motives. Typically, if we hold

constant the agent’s motives, the agent could choose either to A or to B, where
A and B are different actions.
For example, my decision as to whether to have grapefruit or cereal for breakfast is not
determined by my motives; on two consecutive mornings I might have exactly the
same motives, and yet choose grapefruit one day and cereal the next.
Wallace and Velleman are explicit about their commitment to Non-determination,13
but we can see that anyone who wants to maintain a distinction between actions caused
by choice and actions caused solely by motives will be committed to it as well. After all,
if choice were simply determined by motives, then there wouldn’t really be two types
of actions. Chosen actions would also be determined in a passive fashion, issuing merely
from the agent’s motives.
What reasons might we have for accepting Non-determination? A version of this
claim was proposed by Leibniz, and defended by Locke and Kant. Leibniz famously
claimed that motives “incline without necessitating.”14 On this view, motives do not
directly cause action. Rather, while motives incline or dispose us to perform certain
actions, we are free to choose whether to act as the motives incline us to act.
Locke and Kant offer a sophisticated defense of this position. They claim that when a
self-conscious agent reflects on potential actions, she is committed to viewing her
deliberation as capable of suspending the effects of her motivational states. Locke writes that
the mind has “a power to suspend the execution and satisfaction of any of its desires.”
The mind can “consider the objects of [these desires]; examine them on all sides and
weigh them with others. In this lies the liberty that man has” (Locke 1975, 263).
Kant endorses a similar model of deliberation, writing that human choice “can
indeed be affected but not determined by impulses . . . Freedom of choice is this independ-
ence from being determined by sensible impulses” (Metaphysics of Morals 6:213–14).
Korsgaard describes the Kantian model of deliberation as follows:
Our capacity to turn our attention to our own mental activities is also a capacity to distance
ourselves from them, to call them into question . . . I desire and I find myself with a powerful
impulse to act. But I back up and bring that impulse into view and then I have a certain distance.
13
Wallace writes, “volitional motivations [e.g., choices] are independent of our given desires . . . What an
agent chooses or intends to do is not a function of the given desires to which the agent is subject at the time”
(2006, 150). Velleman’s view is more complex. He argues that when an agent is faced with the choice of
whether to A or B, the agent can sometimes make it the case that he will A by forming a self-fulfilling belief
that he will A. The self-fulfilling belief need not be determined by the motives that the agent is reflecting
upon, so a version of Non-determination is true. However, the self-fulfilling belief will be determined by a
background motive which, Velleman argues, is omnipresent—the motive of self-understanding. This
background motive determines which particular self-fulfilling belief the agent formulates. So, while the
motives reflected upon do not determine choice, a background motive does determine choice. Thus,
Velleman’s view is a qualified version of Non-determination. For this view, see “Epistemic Freedom” in
Velleman (2000).
14
See Leibniz’s Fifth Letter to Samuel Clark, collected in Leibniz and Clark (2000, 37).
Now the impulse doesn’t dominate me and now I have a problem. Shall I act? Is this desire really
a reason to act? (Korsgaard 1996b, 93)15
Locke and Kant are making a claim about motivation: by self-consciously reflecting on
a motive, we can distance ourselves from this motive, thereby making the motive cease
to “dominate” us, or “suspending” the motive. Put differently, reflecting on a motive
suspends the efficacy of the motive. So, in unreflective action, we may simply be caused
to act by our strongest motive. But, in reflective action, our motives operate as mere
inclinations, which are incapable of causing us to act.
Thus, the primary point that Kant and Locke are making can be put this way:
(Inclination) In deliberative agency, motives incline without necessitating. The

agent’s motives could be the same, and yet she could choose differently.16
Further, Kant and Locke support Inclination with a claim about deliberation: they
claim that when an agent reflects on her motives, she suspends the influence of these
motives, and is then free to assess the rationality of acting as the motives suggest. We
can put this claim as follows:
(Suspension) When an agent reflects on her motives for A-ing, she suspends the
influence of these motives upon her assessment of the rationality of A-ing.17
Although the details will have to be worked out, I think we can already see why Kant
and Locke wish to distinguish reflective, deliberative actions from unreflective, non-
deliberative actions.18,19 If the claims about Suspension and Inclination were true, then
15
Compare Tamar Schapiro, who writes, “According to the Kantian picture of action, our inclinations
need not determine what we do. They influence us, but we have the capacity to decide, freely and rationally,
whether or not to act on them” (2009, 229).
16
Inclination differs from Non-determination in only one respect: Inclination includes a claim about how
motives operate.
17
We could state Suspension in a more modest form, by adding the word “typically” to the beginning of
the definition. I will return to this point in section 3.1.
18
There is some scholarly controversy on whether Kant actually endorses the claims that I label
Inclination and Suspension. The dominant view among contemporary Kantians seems to be that Kant
does, indeed, endorse the claims. For example, Allen Wood writes, “Kant holds that in the brutes, impulses
operate mechanically to produce behavior predetermined by instinct . . . This means that a brute cannot resist
impulses, or decide whether to satisfy a desire, or even deliberate about how to satisfy it” (Wood 1999, 51).
On the other hand, “Kant contrasts this with the human power of choice, which is ‘sensitive’ (affected by
sensuous impulses) but also ‘free’ . . . Only a free power of choice is a will . . . Not only do rational beings have
the capacity to resist impulses, but even when the rational faculty of desire acts on sensuous impulses, it is
never determined by them mechanically . . . ” (Wood 1999, 51). Henry Allison notes that “incentives
(Triebfedern) do not motivate by themselves by causing action but rather by being taken as reasons and
incorporated into maxims” (Allison 1990, 51). This “requires us to regard empirical causes (motives) of the
actions of sensuously affected and thoroughly temporal rational agents such as ourselves as ‘not so determin-
ing’ so as to exclude a causality of the will . . . ” (1990, 52). For “I cannot conceive of myself as [a rational
agent] without assuming that I have a certain control over my inclinations, that I am capable of deciding
which of them are to be acted upon (and how) and which resisted” (1990, 41). Marcia Baron writes that
“Kant’s theory of agency is very different [than the familiar causal models]. Our actions are not the result of a
desire or some other incentive that impels us. An incentive can move us to act only if we let it” (1995, 189).
Korsgaard (1996b, 94) and Reath (2006, 154) agree. On the other hand, Frierson (2005), McCarty (2009,
the reflective agent would have the capacity to rise above her motives, cancel their
effects, and choose in complete independence of them. This would give us good
reason for distinguishing unreflective actions, which are produced by the agent’s
motives, from reflective actions, which are produced in independence of the agent’s
motives.20
Versions of this Kantian/Lockean model of choice have become predominant in the
literature on action, and indeed seem to be implicitly or explicitly present in each of
the accounts reviewed in Section 1. In the following sections, I will ask whether the
Kantian/Lockean model is tenable. I will argue that while Inclination is true, Suspen-
sion is false. In addition, I will argue that Suspension is required to support the idea that
deliberative actions are paradigmatically active. Thus, the Kantian/Lockean model of
agential activity will have to be rejected.
67ff.), and Wuerth (forthcoming) develop a very different reading of Kant’s theory of agency, according to
which there is a sense in which motives determine choice. For example, Wuerth argues that we should
distinguish the claim that we have “reflective distance” from our motives from the claim that we have
“affective distance” from them: “We certainly have a reflective distance on both intellectual and sensible
desires due to our capacity for self-consciousness, and we are therefore not necessitated to act on either, but
this reflective distance is not what one could term a complete affective distance, where we are completely
isolated from sensible desires: reflective distance on a sensible desire does not mean that the sensible desire
becomes nothing to us . . . ” (Wuerth forthcoming, 20). He continues, “because Kant actually understands us
as far more than reason alone, even when we are self-consciously reflecting on our choices, and so as subject
to feelings and desires of both the sensible and intellectual sort, he recognizes in these feelings and desires a
rich range of factors that can impede or facilitate moral choices” (Wuerth forthcoming, 33). Resolving these
interpretive difficulties is beyond the scope of this essay. For my purposes, it does not matter whether Kant
himself endorsed Inclination and Suspension; I am simply interested in the fact that versions of Inclination and
Suspension are widely accepted by contemporary philosophers. See also the next note.
19
Consider one potential objection to my claim that Kant endorses Suspension. Kant often remarks that
agents can be mistaken about how they are being motivated. For example, Kant claims that a reflective agent
can believe that he is acting on duty, but actually be acting upon self-interest. This might seem to indicate that
Kant actually rejects Suspension: for, in this case, the agent reflects on his motives, takes himself to be acting
on one of them, but is actually acting on another. Reflection therefore seems not to suspend the effects of the
latter motive (self-interest). (Thanks to Christine Korsgaard for raising this objection.) However, I think these
kinds of example are compatible with Kant’s acceptance of Suspension. I intend Suspension to be read as
follows: when an agent reflects on a set of motives, she suspends the effects of those motives. Kant does not
think that reflection suspends the effects of all of the agent’s motives—in particular, reflection does not
suspend the effects of the motives that the agent is not reflecting upon. This gives us a way of interpreting the
passages on self-interest: Kant is claiming that background motives, upon which the agent is not reflecting,
can influence the agent, and thereby make it the case that the agent is actuated by self-interested motives even
when he takes himself to be acting on duty. (An example: I can reflect on the fact that helping Bob is the right
thing to do, and neglect to attend to the fact that I have a strong desire to ingratiate myself with him; this latter
motive, unnoticed, can influence my action. I take myself to be acting on the motive of duty, whereas I am
actually influenced by a self-interested motive.) If this is the correct interpretation, then it is compatible with
my attribution of the Suspension claim to Kant.
20
This point is often emphasized in discussions of free will. For example, Robert Kane writes, “Free
will . . . is the power of agents to be the ultimate creators or originators and sustainers of their own ends or
purposes . . . when we trace the causal or explanatory chains of action back to their sources in the purposes of
free agents, these causal chains must come to an end or terminate in the willings (choices, decisions, or efforts)
of the agents, which cause or bring about their purposes” (Kane 1996, 4).
120 AC TI O N ’S FI R S T C O N S T IT U T I V E A I M : AG E N T I A L AC T I V I T Y
2.2 Is Inclination true?

Let’s begin with Inclination. Should we accept the Kantian/Lockean claim that we can
choose in independence of our motives? If we reflect on the experience of deliber-
ation, I think it will provide some support for the Kantian/Lockean analysis. Take my
choice of what to have for breakfast. Here, I reflect on and assess my desires for cereal
and grapefruit. Kant and Locke are claiming that when I reflect on these desires, I will
have an experience of distance from them: I will experience myself as free to act on
either of the desires. As a characterization of the phenomenology of choice, this point
seems accurate. Though there are cases in which desire seems to compel choice, such as
addictions, pathologies, and cases of extreme emotion, these cases are atypical. Typic-
ally, when we decide what to have for breakfast, or how to spend our evening, or what
career to pursue, it feels as if we are free to choose between competing options.
But consider an objection: might there be some alternative way of conceiving of
deliberation, such that it does not entail or presuppose Inclination? In other words,
couldn’t an agent take his motives to determine what he will do, and still deliberate?
Hobbes thought he could:
When in the mind of man appetites and aversions, hopes and fears, concerning one and the same
thing, arise alternately; and diverse good and evil consequences of the doing or omitting the
thing propounded come successively into our thoughts; so that sometimes we have an appetite to
it, sometimes an aversion from it; sometimes hope to be able to do it, sometimes despair, or fear
to attempt it; the whole sum of desires, aversions, hopes and fears, continued till the thing be
either done, or thought impossible, is that we call deliberation . . . In deliberation, the last
appetite, or aversion, immediately adhering to the action, or to the omission thereof, is what
we call the will . . . (Hobbes, Leviathan VI)
And William James offers a roughly Hobbesian description of deliberation:

The process of deliberation contains endless degrees of complication. At every moment of it our
consciousness is of an extremely complex object, namely the existence of the whole set of
motives and their conflict . . . Of this object, the totality of which is realized more or less dimly all
the while, certain parts stand out more or less sharply at one moment in the foreground, and at
another moment other parts, in consequence of the oscillations of our attention, and of the
‘associative’ flow of our ideas. But no matter how sharp the foreground-reasons may be . . . the
background, however dimly felt, is always there . . . The deliberation may last for weeks or
months, occupying at intervals the mind. The motives which yesterday seemed full of urgency
and blood and life to-day feel strangely weak and pale and dead . . . we must wait a while, patient
or impatient, until our mind is made up ‘for good and all’. (James 1890, vol. II, 528–9)
The language here—especially James’ last sentence—suggests a kind of passivity with

respect to one’s own decision. James does not say that we must make up our mind;
he says we must wait until our mind is made up. And he emphasizes the way in which
the objects of deliberation impress themselves upon the deliberator with differing
intensities.
What are we to make of these descriptions? Do they show that we can deliberate
while envisioning our deliberation as determined by motives? Here we should draw a
distinction. It’s undeniable that Hobbes and James have offered accurate descriptions of
a portion of human life. We do, sometimes, have awareness of conflicts of motives; we
do, sometimes, merely observe as these conflicts play themselves out; we do, some-
times, learn what we will do, or learn what we really desire, rather than deciding, in a
moment of choice, that we will do something. Indeed, many of the most important
aspects of human life fit this description: romantic relationships, friendships, decisions
about careers, and so forth, often feel just the way that James describes. We weigh the
options, let things develop in our minds, and eventually see that one motive has taken
shape as dominant.
But this admission should not blind us to the fact that these moments are peppered
with episodes of reflective choice. James’ deliberation may well last “weeks or
months,” but during those weeks and months there will be resolutions more and less
effective, decisions, choices; there will be attempts to try out courses of action; there
will be moments when he decides to stop reflecting on his decision, other moments
when he decides to carry on; and there will be that crucial moment when he decides
that reflection has gone on long enough, that his mind is now “made up for good and
all.” (After all, we are all familiar with the person who, like James, lets an important
decision brew for a few weeks or months; and we are also familiar with a more extreme
version, a person who lets these important decisions brew forever, without making a
choice.) Absent these moments, it would be hard to see James’ “deliberation” as
genuine deliberation at all. After all, some of our motives do change over time in a
non-deliberative fashion. As a child I used to enjoy playing in sandboxes; as an adult
I no longer do. It wasn’t deliberation or choice that brought about this change in my
motives. If James’ “deliberation” takes this form, then it isn’t deliberation at all. So
careful attention to the phenomenology reveals that choices do have the character that
Kant and Locke describe.
Of course, the phenomenology could be misleading.21 However, Kant points out
that there is a powerful reason for taking the phenomenology at face value: if there is to
be any such thing as deliberation at all, then agents must be capable of choosing in a
way that is not determined by their motives. Kant argues that I not only experience my
choices as undetermined by my motives; in addition, there is a sense in which I am
committed to viewing my choices as undetermined. This is one way of putting Kant’s
21
Perhaps the most obvious worry about the phenomenology is that it seems to view deliberation as an
uncaused cause. But don’t our choices have to be determined by something? And isn’t the idea that motives
incline without necessitating a denial of this? No. We should distinguish two claims: (1) our choices are not
determined by our conscious beliefs, desires, and intentions; (2) our choices are not determined by anything.
Clearly, we can maintain (1) while denying (2): even if our conscious beliefs, desires, and intentions don’t
determine our choices, something else (such as non-conscious mental processes) might determine them. So
one can divorce Inclination from general concerns about causal determination. One worry about taking the
phenomenology at face value therefore disappears.
point about the necessity of assuming freedom from the practical standpoint: deliber-
ating about whether to A commits one to viewing oneself as having a say in whether
one will A. If one’s action were simply determined independently of deliberation—for
example by one’s motives—then one would not have such a say. So deliberation
presupposes its own efficacy.
To see why the reflective agent is committed to viewing his reflection as undeter-
mined by his motives, consider the difference between deliberating about what to do
and thinking about what someone else should do. Suppose I consider whether my
neighbor should act on his desire for another serving of ice cream. In engaging in this
kind of reflection, I do not take my thoughts to have any bearing on what the neighbor
will actually do. But this is quite different from the ordinary case of first-person
deliberation about action. When I deliberate about whether I should act on a desire
for ice cream, I take my answer to settle the question of what I will do. I do not regard
my reflection as a passive, theoretical inquiry that has no bearing on its subject matter;
I do not regard myself as making predictions while waiting to see how things turn out.
Rather, if I decide that I should act on some desire, I also take this to settle the question
of whether I will.22
That is why I am committed to viewing my deliberative reflection as suspending the
workings of desire: if I did not view my deliberation as settling the question of what
I will do, then my deliberation would be exactly analogous to the way in which
I might think about what my neighbor should do. In other words, it would not be
deliberation at all.
This Kantian argument establishes that we must conceive of our own deliberation as
proceeding in a way that is not determined by the motives upon which we are
deliberating. Moreover, we must conceive of this deliberation as eventuating in action.
That is, we must conceive of our deliberation as causally efficacious; otherwise, it
would be analogous to thinking about what someone else should do. Of course, there
could still be a mismatch between the way that we are compelled to conceive of
deliberation, and what actually happens when we deliberate. But Kant’s argu-
ment establishes a very powerful point: if there is to be any such thing as genuine
deliberation—if anything in the world answers to our concept—then deliberation
must proceed in a way that is not determined by motives. Inclination must be true.23
So we have two reasons for accepting Inclination. First, it is supported by the
phenomenology. Second, the Kantian argument shows that it is an ineradicable feature
22
Of course, any number of things can prevent my judgment that I should A from leading to my A-ing:
weakness of will, unforeseen circumstances, accidents, failures. So the claim is not that when I judge that
I should A, I then A. Rather, the claim is that when I judge that I should A, I take this judgment to settle the
question of whether I will A.
23
Notice that I am not claiming that in order for agents to engage in what they take to be deliberation,
Inclination must be true. As I note above, there could be a mismatch between the way in which we conceive
of deliberation and what actually happens when we deliberate. My claim is simply that in order for
deliberation to be what we take it to be, Inclination must be true. Thanks to David Enoch for pressing me on
this point.
of our conception of deliberation. Still, these considerations are not decisive: the
phenomenology could be misleading, and the fact that we must conceive of a
phenomenon in a certain way does not imply that the phenomenon is that way (a
rough analogy: that we must see a stick in water as bent does not imply that it really is
bent). So we need further support for Inclination. Let’s turn, then, to empirical
evidence.
The evidence from empirical psychology suggests that Inclination is true: we do
have the capacity to override our immediate desires, choosing in a way that is not
determined by them. One body of evidence for this claim comes from a group of
psychologists who have been studying “executive control,” which is defined as agents’
capacity to “call upon goals or standards to regulate their actions and responses, and
make choices” (Vohs 2010, 67). The empirical evidence on executive control indicates
that human beings do have a capacity to override their immediate desires, selecting
which goals to pursue. This capacity is not at work in every instance of action. Rather,
episodes of executive control appear to be intermittent and costly. In particular, a
number of experiments have demonstrated that when agents exercise their ability to
override immediate desires and impulses, they temporarily hinder their ability to
engage in further “reasoning, rational thought, and intelligent decision making”
(Vohs 2010, 67). For example, in one experiment subjects watched a disturbing film.
One group was asked to suppress emotional reactions to the film, maintaining a neutral
facial expression. The control group was given no such instructions. Afterwards, both
groups took a series of tests. The agents who suppressed their emotional reactions to
the film scored worse on tests requiring them to engage in logical reasoning, but
equally well on tests requiring mere recitation of factual information (Schmeichel et al.,
2003; Vohs 2010, 69). Experiments of this sort—of which there have been over a
hundred—have demonstrated that logical reasoning, rational thought, attention con-
trol, emotional modification, and self-control deplete the same “energy resources”:
engaging in any one of these activities makes one less successful at engaging in the
others (Baumeister et al. 2010 and Vohs 2010 provide helpful overviews).24
24
Additional evidence in favor of Inclination is provided in Mele (2009). Mele discusses research by Peter
Gollwitzer on “implementation intentions,” which are defined as intentions that specify when, where, and
how one will achieve one’s goals (Mele 2009, 135). Implementation intentions turn out to have a significant
impact on subjects’ future behavior: when subjects form an implementation intention, they are significantly
more likely than a control group to perform the relevant behavior. For example, in one study, subjects are
given the task of vigorously exercising for 20 minutes during the next week. The subjects are presented with
information about the health benefits of exercise. The subjects are then divided into two groups. The first
group is given no further instructions; during the next week, only 39 percent of these individuals perform the
20 minutes of exercise. The second group is instructed to form implementation intentions (i.e., to decide
when, where, and how they will exercise). During the next week, 91 percent of these individuals perform the
20 minutes of exercise (Mele 2009, 135). These kinds of studies—of which Gollwitzer and Sheeran (2006)
reports 94—collectively provide strong evidence for the claim that implementation intentions impact
behavior. See Muraven and Baumeister (2000); Holton (2009, Chapter 3), Mele (2009), Pockett et al.
(2009), and Baumeister et al. (2010) for additional evidence.
Drawing on experiments of this sort, a standard model of human agency has

emerged in empirical psychology. This is the “dual-process” model. According to
this model, there are two sources of action. Most of our actions are produced,
monitored, guided, and maintained by non-conscious, automatic processes. These
actions tend to be based on either heuristics or affects, make very low demands on
cognitive resources, and are thus capable of occurring in parallel with other processes.
A second stream of behavior is produced by conscious processes. These processes are
comparatively slow, resource-depleting, and intermittent. They can be disrupted by
changes in conscious attention, and require a degree of effort.25
In typical cases of action, these two processes interact. Consider a perfectly ordinary
stretch of behavior: I walk to my office while chatting with a colleague. My conscious
attention is focused on the conversation: I consider how to respond to a philosophical
objection, how to reformulate a position. All the while, my legs move, my feet make
precise adjustments to the terrain, my hand grasps the railing as I climb the stairs. In
typical cases I am not explicitly aware of this latter set of actions. Nonetheless, I am
perfectly capable of switching my attention to these automatic processes and choosing
either to continue or desist; I can remove my hand from the rail, I can decide to stand
still rather than continuing. My conscious decisions are determined neither by these
automatic processes nor by the motives driving the processes (though, presumably,
they are determined by something else.)
According to the dual-process model, then, we can distinguish between reflectively
controlled and automatic processes; yet even the automatic processes can be
overridden by reflection, and are thus subject to reflective control. Although certain
details of this model are controversial, the broad outlines are widely accepted and
well confirmed by experimental data.26 Thus, the empirical evidence is in favor of
Inclination.
In sum, then, we have three reasons for thinking that Inclination is true: the
phenomenology, the Kantian argument from the nature of deliberation, and the
empirical evidence from psychology. Not surprisingly, these arguments fall short of a
deductive proof that Inclination is true. Taken together, however, they give us
exceedingly strong grounds for assuming that this claim is true. The burden of proof
is therefore very much on those who wish to deny the claim. Thus, for the remainder
of this chapter, I will assume that Inclination is true.
25
For good overviews of this model, see Bargh and Chartrand (1999), Timothy Wilson (2002), Stanovich
(2004), and Hassin et al. (2005).
26
Some authors, notably Libet et al. (1983) and Wegner (2002) attack aspects of this model.
However, a number of philosophers and psychologists have pointed out that Wegner’s and Libet’s
conclusions are not warranted by their data. For a good overview of the problems with these studies,
see Mele (2009).
S U S P E N S I O N A N D T H E N AT U R E O F M O T I VAT I O N 125
3. Suspension and the nature of motivation

Having granted Inclination, let’s now turn to Suspension. I will argue that Suspension
is false. The problem with the claim that deliberation suspends one’s motives is that it is
ambiguous; it fails to specify what is meant by “suspending” one’s motives. This
ambiguity arises from a failure to distinguish two models of the way in which motives
can operate: a simplistic model and a sophisticated model. According to the simplistic
model, motives operate as mere pushes and pulls. Below, I will argue that if this were
the only way in which motives operated, then Suspension would be true. However,
I argue that motives also operate in a more sophisticated manner: motives influence
reflection itself. I explicate and defend this model of motivation, showing that it is well
supported both by philosophical considerations and by results from empirical psych-
ology. Moreover, I show that the sophisticated model of motivation entails that
Suspension is false.
3.1 Two ways of being moved
With this in mind, let’s examine two ways in which attempts at deliberative suspension
can fail. There appears to be a familiar type of failure: an agent’s attempt at suspension
can be overpowered. Consider Harry Frankfurt’s description of such a case:
Sometimes people are unsuccessful even in strenuously conscientious efforts to avoid being
moved into action by desires that they would prefer to be motivationally ineffective. For
instance, someone may act out of jealousy, or out of a desire for revenge, although he disapproves
of these motives and would strongly prefer that he not be driven by them. Unhappily, as it turns
out, their force is too great for him to withstand; and in the end he submits to it. Despite his
resistance, the unwelcome desire is effective in moving him to act. (Frankfurt 2004, 19; emphasis
added)
Frankfurt describes a self-conscious, reflecting agent who is confronted with an attitude

of which he disapproves. The agent struggles to avoid being moved by this attitude, but
in the end, he “submits” to it; he is, perhaps, carried along by its force.
What kind of process is Frankfurt envisioning, when he imagines the agent con-
fronting and submitting to his motive? The picture seems to be as follows:
1. A motive tempts the agent to pursue some end.
2. The agent reflects on this motive and his potential actions, considering whether
he wants to act on the motive.
3. The agent decides not to act on the motive.
4. The agent struggles to avoid being caused to act by the motive.
5. The agent’s resistance is overcome, and he is caused to act by the motive.
There are cases that seem to fit this description. Addictions provide good examples: an
alcoholic struggles to resist his desire for another drink, but finds himself picking up the
bottle once more. Perhaps cases of this sort occur even in more mundane, everyday
situations: a decadent dessert lies before me on the table, and I crave it. I disapprove of
126 AC TI O N ’ S FI R S T C O NS T IT U TI VE A I M: AG E N T I A L AC T I V IT Y
this craving, decide not to eat the dessert, and struggle to resist the temptation. I try to
put the thought of dessert out of my mind. But the temptation is too much for me: the
thought of eating keeps recurring, my eyes are continuously drawn to the dessert.
Eventually, I give in and pick up the spoon.
If this sort of case is a real possibility, then Suspension, in its unqualified form, would
be false: there would be at least some cases in which deliberation failed to suspend the
influence of motives. However, proponents of Suspension could respond in two ways.
First, they could admit that this kind of case is possible, but point out that it is atypical,
occurring only when the motive in question is unusually strong. Accordingly, they
could qualify Suspension by claiming that it is true only in cases in which the motive is
not unusually strong. Second, they could argue that Frankfurt simply mischaracterizes
the case. In the jealousy and dessert cases given above, I submit to the desire not in the
sense that it overpowers me, but in the sense that I decide to give in to it, as I might give in
to a pesky child, when resistance no longer seems worth the effort. In short, I am not
overpowered by the desire; rather, I at first resist, but eventually consent to be moved by
it.27 If this is right, then Frankfurt’s examples are compatible with Suspension.
Examples of the above sort do not impugn Suspension, then. However, I am going
to argue that there is a different way in which an agent can be moved, unwillingly, by
an attitude. I will show that this second form of motivation is far more problematic for
proponents of Suspension.
Let’s return to the example of jealousy. Frankfurt considers the case in which an
agent clear-sightedly reflects on his jealousy and tries to resist its grip. But surely one of
the most distinctive features of our attitudes is that they influence reflection itself. Part of
what it is to be in the grip of jealousy is to see reasons for jealousy everywhere: in the
fact that Melissa arrived home a few minutes later than usual; in the fact that she got off
the phone rather quickly last night; in the fact that she is a bit quiet tonight.
Accordingly, jealousy and other attitudes can move an agent not simply by overpower-
ing his capacity to resist their pull, but by influencing his judgment. A jealous agent’s
attention will be drawn to certain features of his environment that another agent would
scarcely notice. A jealous agent’s trains of thought will return to details that another
agent might regard as inconsequential. A jealous agent’s deliberative process itself can
be influenced by these attitudes; they can incline him to draw conclusions that are not
supported by the evidence, to give excessive weight to certain features, and so on. All
of this may occur without the jealous agent’s recognizing that it is occurring.
Precisely because attitudes influence reflective thought, agents often fail to grasp the
ways in which attitudes are motivating them. An agent who succumbs to jealousy is not
always an agent who consents to be moved by his jealousy; indeed, an agent who
27
There are cases in which the person literally has no control over the behaviors that his body performs.
For example, in cases of anarchic hand syndrome, the person’s hand engages in goal-directed behaviors over
which the agent has absolutely no control. But I see no reason to think that the case of desire is in any way
analogous to these pathological cases.
succumbs to jealousy need not even recognize, much less consent to, a fully formed
attitude of jealousy. More often, the jealous agent will struggle to resist the jealousy,
but succumb to it in subtler ways. The attitude influences the agent’s reflective thought
itself: the agent experiences herself as having a reflective distance from the attitude, as
scrutinizing the attitude and asking herself whether there is a reason to act on it; but, all
the while, the attitude influences the agent’s reflective thought in ways that she does
not grasp.
This type of influence is easiest to detect when we look back at an action retrospect-
ively. A person can be dissatisfied with his past actions not because he submitted to or
was overcome by a recalcitrant attitude, but because his attitude blinded him, leading
him to have a restricted, impoverished, or distorted conception of the options that
were open to him. Looking back on my jealous spat with Melissa, the problem was not
that I yielded to jealousy: the problem was that, in the grip of jealousy, I made so much
of so little. The problem was that I saw my rage as warranted by the fact that Melissa got
home a few minutes late. I now see that the rage was entirely unwarranted, that I was
driven to rage in a way that I did not comprehend. In this way, an agent can act
reflectively, yet still be moved by attitudes that operate in the background.
So we have distinguished two ways in which Suspension could fail:
(Motives as forces) Motives, if they operate as brute forces, could cause us to act in a
way that we choose not to act.
(Motives as influences) Motives could influence our reflection in such a way that even
when we attempt to achieve a distance from them, we fail to do so.
In cases of influence, motives operate through reflection, rather than independently of
it. In a moment, I will show how these claims bear on the idea of deliberative
suspension. First, though, let’s explicate the idea that motives might operate as
influences.
3.2 The way in which motives can operate as influences
There are at least three ways in which motives can act as influences upon reflection
itself:
(i) Motives can affect perceptual saliences.
(ii) Motives can affect the way in which we conceive of our circumstances, our
reasons for action, and the potential actions that lie open to us.
(iii) Motives can affect the course of deliberation itself.
Influence of type (i) is perhaps the most obvious: motives affect perceptual saliences.
When I’m hungry, my attention tends to be drawn to food. When I’m angry, my
attention tends to be drawn to features of my environment that might justify or
perpetuate the anger. The effects will probably be proportional to the strength of the
motive: mild hunger won’t generate as much effect on perceptual saliences as extreme
hunger. I take it that these points are uncontroversial and familiar.
Type (ii) influence occurs when motives affect the way in which we conceive of our
circumstances, our reasons, and our potential actions. This requires more explanation.
First, we need to distinguish the agent’s circumstances from the way in which the agent
describes or conceives of his circumstances. Anyone who reflects on a case will be
entertaining some description of the facts of the case; this description will be partial,
couched in these terms when it might have been couched in those terms. Moreover,
this description will influence, constrain, and sometimes even determine the way that
the person goes on.
This shows up even in relatively simple cases: until you stop seeing the geometrical
problem in this way, you won’t be able to solve it. It is also obvious in the case of
discussion: often, changing someone’s mind is achieved by getting him to use different
descriptions of the same facts (“Don’t think of it as stealing, think of it as taking a little
something from an incredibly rich, greedy corporation”; “Don’t think of it as making a
joke, think of it as hurting Tom’s feelings”).
Stuart Hampshire makes this point quite clearly:
It is misleading to speak of ‘the facts of the situation’ in such a way as to suggest that there must be
a closed set of propositions which, once established, precisely determine the situation. The
situations in which we must act or abstain from acting, are ‘open’ in the sense that they cannot be
uniquely described and finally circumscribed. Situations do not present themselves with their
labels attached to them . . . (Hampshire 1949, 476)
In a footnote to the above passage, he adds:

The word ‘fact’, here as always, is treacherous, involving the old confusion between the actual
situation and the description of it; the situation is given, but not ‘the facts of the situation’; to state
the facts is to analyze and interpret the situation. And just this is the characteristic difficulty of
actual practical decisions, which disappears in the text-book cases, where the ‘relevant facts’ are
pre-selected. (Hampshire 1949, 476)
Descriptions of situations, just in virtue of the fact that they characterize the situation in
a determinate way, involve simplification and incompleteness. Certain details are
emphasized at the expense of others. In this sense, descriptions are inescapably partial.28
Psychologists have studied this phenomenon for a number of years under the
heading of “framing.” Researchers have demonstrated that the way in which a
situation is described has a profound impact on the decisions that agents make. For
illustrative purposes, I will mention the classic example of this phenomenon: Kahne-
man and Tversky’s “dangerous disease” case (Kahneman and Tversky 1981).
Case 1: There has been an outbreak of a dangerous disease. 600 people are infected. Doctors can
adopt treatment Program A or treatment Program B, but not both. If Program A is adopted, 200
28
Frederic Schick also remarks on this phenomenon. Schick argues that an adequate psychological
account of action will make room not only for beliefs and desires, but also for the way in which the agent
“understands,” “perceives,” or “frames” her situation. See Schick (2003, 61 and passim).
people will be saved. If Program B is adopted, there is a 1/3 probability that 600 people will be
saved, and a 2/3 probability that no one will be saved.
Case 2: There has been an outbreak of a dangerous disease. 600 people are infected. Doctors
can either adopt treatment Program A or treatment Program B, but not both. If Program A is
adopted, 400 people will die. If Program B is adopted, there is a 1/3 probability that on one will
die, and a 2/3 probability that 600 people will die.
There is no factual difference between these two cases. Rather, they simply state the
facts differently: the first case focuses on the number of people who will be saved,
whereas the second one focuses on the number of people who will die. However,
researchers obtained strikingly different results when asking respondents which scen-
ario they would prefer. When presented with Case 1, 72 percent preferred Program
A and 28 percent preferred Program B (sample size was 152). When presented with
Case 2, 22 percent preferred Program A and 78 percent preferred Program B (sample
size was 155).
The researchers demonstrated that this sort of bias is ubiquitous: the way in which
we “frame” or conceive of our situation has a dramatic effect on our deliberation and
choice (for further details, see Kahneman and Tversky 1981). With these points in
mind, we can now explicate type (ii) influence. Our motives can manifest themselves
by influencing the particular descriptions or conceptions that we employ, for the way
in which we frame or conceive of situations depends, in part, on the motives that we
have.
The most obvious way in which motives influence conceptualizations or descrip-
tions is by affecting perceptual saliences: the way in which I describe a situation will be
strongly influenced by facts about which features of my environment are salient to
me.29 Thus, a recent survey of the psychological work on the relationship between
emotion and cognition concludes that
emotion may influence cognition [by] modulating which information in the environment
reaches awareness. At any given time, we are bombarded with sensory input. Only a portion
of this input is available for cognitive awareness . . . A number of psychological studies have
confirmed that attention and awareness may be influenced the emotional content of a stimulus.
(Phelps 2005, 70)
For example, when an agent is hungry, she is more likely to be aware of information
that relates to the presence of food in her environment, and as a result she is more likely
to conceptualize her situation in terms relating to food. Or, to return to the earlier
example, if an agent is jealous, he is more likely to conceptualize his environment in
terms of jealousy.
However, there is evidence that motives have an even more pervasive effect on the
way in which situations are conceptualized. I will cite just two examples. First, it has
been demonstrated that the perceived distance between an agent and a goal increases as
29
Thus, type (ii) influence occurs, in part, through type (i) influence.
130 AC TI O N ’ S F I R S T C O N S T I T U T I V E A I M : AGE N T I A L AC T IV I T Y
the effort associated with walking to the goal is increased. For example, the perceived
distance increases when the agent is wearing a heavy backpack, or when the agent is
tired (Witt et al. 2004). Second, studies have shown that participants’ characterizations
of other individuals are influenced by their moods. For example, subjects who watch a
sad movie and are then asked to characterize a given person as likeable or unlikeable
tend to characterize him as unlikeable, whereas subjects who watch a happy movie
tend to characterize the same individual as likeable (Forgas and Vargas 2004, 357). As
these examples demonstrate, motives exert a significant influence on the ways in which
situations are conceptualized.30
In sum, type (ii) influence occurs when motives affect the starting points of deliber-
ation, by influencing our conceptualizations or descriptions of the case. This brings us
to type (iii) influence: motives can also affect the process of deliberation. In other words,
motives can affect the movement from a description of the situation to a conclusion
about what to do.
There is a variety of work on individuals’ desires to reach specific conclusions about
themselves or others; much of this, in philosophical writing, is placed under the
heading of self-deception. For example, I don’t want to recognize that my lover is
cheating on me, and I selectively process information to that end: I discount evidence
of cheating, and heavily weight evidence of fidelity. Or, I want to view myself in a
positive light, and end up downplaying or ignoring certain aspects of myself.
However, this phenomenon is hardly restricted to self-deception. There is a wealth
of psychological research on the way in which motives influence reasoning and
thinking. It is well known that motives influence judgments by selectively influencing
attention, memory, and association (Bower 1981; Clark and Waddell 1983; Isen 1984,
1987). In general, this happens in two ways. First, as we saw above, motives will “affect
reasoning by affecting which information will be considered in the reasoning process”
(Kunda 1990, 486). Second, motives can influence the reasoning process itself. For
example, motivation impacts “evidence evaluation”: people evaluate information that
supports or contradicts positive self-evaluations in an interesting way, giving more
credence to information that supports a positive self-conception, and less credence to
information that threatens it. Motivation also impacts “information search”: we are
motivated toward “decreased processing and quick acceptance of favorable evidence,
and increased processing and hesitant acceptance of unfavorable evidence” (Molden
and Higgins 2005, 299). There is also a nice body of evidence showing that motives
influence “memory search”—that is, motives influence which memories come to
mind. For example, Sanitioso et al. (1990) induced one group of participants to view
introversion as a desirable character trait, and another group to view extroversion as a
30
For a helpful overview, see Forgas and Vargas (2004), which provides a host of evidence that “mood
[or, more generally, affect] can have both informational and processing effects on cognition. Informational
effects occur because mood influences the content of cognition (what people think). Processing effects occur
because mood influences the process of cognition (how people think)” (Forgas and Vargas 2004, 351).
desirable character trait. When asked to recall memories, the participants in the first
group were significantly more likely to recall memories of their own introversion.
In sum, as the psychologists Molden and Higgins put it in a recent survey of the
literature on motivated thinking, motivation works by “directing people’s cognitive
processes (e.g., their recall, information search, or attributions) in ways that help to
ensure that they reach their desired conclusions” (Molden and Higgins 2005, 297).
Moreover, several studies have shown that increased deliberation and reflection actu-
ally heighten the degrees of these effects (Feshbach and Singer 1957; Forgas 1994).
Much more could be said on the particular ways in which each of these types of
influence occurs. However, this brief overview will suffice for our purposes. The
important point is that a wealth of psychological research supports a point that should
be clear merely from a careful and judicious analysis of the phenomenology of
motivation: motives do not manifest themselves simply as pushes and pulls, whose
urgings we can choose to resist. The effects of motives are far subtler: our motives
manifest themselves as colorings of thought and deliberation. They influence both the
starting points and the process of deliberation itself, and thereby influence the outcome
of deliberation (i.e., the choice). Put simply, motives operate through, rather than
independently of, our rational capacities.31
3.3 The ambiguity of Suspension
The previous two sections argued that motives operate through reflection and deliber-
ation: part of what it is to have a motive is for one’s reflection to be altered in a certain
way. Let’s ask how these claims bear on Suspension:
influence of these motives upon her assessment of the rationality of A-ing.
Kant and Locke make two points here. First, they claim that reflection suspends the
effects of the agent’s motives, as we’ve seen above. Second, they claim that once the
agent has begun reflecting on a motive, she can assess the rationality of acting on it
without influence by the motive.
Let’s start with the first point. I think we can agree that once I have begun reflecting
on a motive, the motive very rarely acts as a brute force compelling me to act. In other
31
Some philosophers have argued for related points. For example, Blackburn (2001) argues on Humean
grounds that traditional Kantian claims about agency are untenable; Risse (2007) adduces Nietzschean
considerations in support of a similar conclusion; and Cuypers (2001) offers a nuanced analysis of the
connections between motivation and reflection. My argument differs from these accounts in several ways.
First, I distinguish the claims that I have labeled Inclination and Suspension, which are often run together;
second, I provide not only philosophical considerations but also empirical evidence of the falsity of
Suspension; third, I chart the particular ways in which Suspension fails. This enables my argument to avoid
some of the difficulties encountered by the authors mentioned above. For example, Kantians can dismiss
Blackburn as caricaturing the Kantian view or as eliding distinctions that Kantians can draw (cf. Korsgaard
2009, 124ff.). No such argument seems applicable to the empirically grounded argument given above.
words, cases of motives operating as overpowering forces—cases of the sort that

Frankfurt describes—are at best rare. However, we have seen that motives can operate
as influences upon reflection itself. Thus, the motive can continue to operate on me
even as I am examining it. The jealous agent, reflecting on her jealousy, will see the
jealousy as warranted, and she will see it as warranted precisely because she has failed to
suspend its influence.
This brings us to the second point. We can also agree that the reflective agent
typically looks for a reason to act on the motive. However, given that motives influ-
ence reflective thought, a problem arises: the reasons that the agent finds will be
products of his motives. For example, when in the grip of jealousy, reflective assess-
ment of one’s jealous motives will typically vindicate these motives, precisely because
the jealousy will manifest itself by inclining the agent to see jealous responses as
warranted by the situation at hand. In other words, motives will affect the agent’s
perception of reasons.
Thus, it is a mistake to think that either the agent is pushed to action by some
overpowering motive, or the agent acts reflectively in a way that is not determined by
his motives. There is a third option: the agent can scrutinize his motives, decide that
there is a reason to act on them, and yet, all the while, be in thrall to some motive. The
effects of the motive needn’t be construed as pushes and pulls that force an agent to act;
rather, the motive moves the agent by influencing the agent’s perception of reasons,
inclining the agent to see action that fulfills the motive as rationally warranted.
Once we recognize that there are two ways of being moved (via force or via
influence), I think we can see that Suspension is false. Grant that our motives rarely
act as brute forces compelling us to act. If that were the only way in which motives
could move us, then (some version of) Suspension would be true, for deliberation often
does suspend this kind of influence. However, there is another way in which motives
can move us: they can pervasively influence the course of reflection itself, in particular
by inclining us to see acting upon the motives as warranted. Reflection does not
suspend this type of influence. On the contrary, reflection is often the vehicle for this
kind of motivation, perpetuating its effects.
Put simply: Suspension will seem true only if we operate with an impoverished
conception of the way in which motives affect reflection. Once we recognize the
different ways in which motivation can operate, we see that Suspension is false.32
32
Note that I am not claiming that reflection never suspends the effects of motives. There may be some
cases in which reflection does suspend the effects of motives. For example, when an agent reflects on weak
and evenly balanced motives, such as the desire to have cereal for breakfast and the desire to have grapefruit
for breakfast, it is entirely possible that reflection actually does suspend these motives. My argument is simply
that Suspension—the claim that reflection either always or typically suspends the effects of motives—is false.
(For the qualified version of Suspension, see section 3.1.)
C O N N E C T I O N B E T W E E N A C T I V I T Y A N D D E L I B E R AT I O N 133
4. Rethinking the connection between agential activity

and deliberation
I began by noting that philosophers have been increasingly drawn to the idea that
deliberative agency is paradigmatically active. There are two ideas at work here. First,
these philosophers believe that deliberation is typically sufficient for agential activity;
excepting a few odd cases, if an agent deliberates, chooses to perform a certain action,
and performs that action, then she manifests agential activity. Second, these philoso-
phers believe that by investigating the structure of deliberative action, we will be able
to understand other types of agential activity.33 In the previous sections, I have asked
why this should be so. I distinguished two claims:
(Inclination) In deliberative action, motives incline without necessitating. The

agent’s motives could be the same, and yet she could choose differently.
influence of these motives upon her assessment of the rationality of A-ing.
If both of these claims were true, then there would be good reason to associate
deliberative action with agential activity. Non-deliberative, unreflective actions
would merely be caused by the agent’s motives, whereas deliberative, reflective actions
would be produced by the agent independently of her motives.
However, I have argued that Suspension is false. Let’s now examine the implications
of this conclusion. There are two questions. First, given that Suspension is false, do the
arguments in favor of Inclination still succeed? Second, given that Suspension is false, is
there any reason for maintaining that deliberative actions are paradigmatically active?
4.1 Should we preserve Inclination?
Let’s begin with the first question. First, notice that there is no logical inconsistency
with denying Suspension and maintaining Inclination: despite the fact that we cannot
suspend the effects of our motives, we might still be able to choose in a way that is not
determined by them. (Consider a simple example: given the above results, intense
hunger is a motive that will affect my reflection. But it does not follow that I am
compelled to satisfy my hunger whenever it strikes intensely.) Second, notice that the
falsity of Suspension does not vitiate the Kantian argument in favor of Inclination. An
agent who recognizes that her reflective thought is pervasively influenced by her
motives will nevertheless, when engaged in deliberation, attempt to determine her
action via choice. Finally, the psychological evidence cited in support of Inclination is
likewise unaffected by the falsity of Suspension: the proponents of the “dual-process
model” of agency discussed in Section 2.2 need not deny that reflection is everywhere
33
As David Velleman puts it, the primary topic in action theory is “how to characterize the stereotypes or
paradigms approximation to which determines the extension of the concept ‘action’. The second is to
characterize the dimensions along which instances can depart from the paradigm, and the contextual variables
that determine how much of a departure is too much” (2004b, 282).
influenced by motives. For these reasons, Inclination remains well supported; the falsity
of Suspension has no bearing on the truth of Inclination.
4.2 Should we preserve the idea that deliberative actions are active?
Let’s now turn to the second question. Given the falsity of Suspension, is there any
reason to maintain that deliberative actions are active? The easiest way to answer that
question is to consider cases in which the agent’s choice determines what she does, and
Inclination is true, but Suspension is false. Consider the following case, which I borrow
from Nomy Arpaly:
I see a piece of cake in the fridge and feel a desire to eat it. But I back up and bring that impulse
into view and now I have a certain distance . . . Is this desire really a reason to act? I consider the
action on its merits and decide that eating the cake is not worth the fat and the calories. I walk
away . . . This could be the inner monologue of a rational, autonomous being, and this is how it is
usually presented. It could equally well be the inner monologue of an individual with severe
anorexia nervosa, weighing eighty-five pounds . . . Imagine such a case: a woman who appears to
herself to be in control of her desires, deciding between them on their merits, but who appears to
her friends (or even to her future self, after having recovered from her anorexia or her irrational
dieting) to be a person who is in fact at the mercy of her desires . . . The anorectic . . . experiences
her psyche in terms of self-control, as if there were something that was her, choosing between her
desires on the basis of their merits, giving her control over herself, while we have good reasons to
believe that unconscious desire or emotion moves her in a manner not characteristic of well-
exercised practical reason. (Arpaly 2004, 17–18)
Presumably, the anorectic agent does have the power to determine her action by
choice: it is her decision not to eat the cake that leads her not to eat it. Moreover, we
can assume that Inclination is true of her: her motives do not necessitate her choice, for
she could have chosen, instead, to eat the cake. However, Suspension is not true of her.
She steps back from her desire to be thin, reflects upon it, and attempts to assess it. She
examines reasons for and against acting on it. She comes to a reasoned conclusion: she
should not eat the cake, for she doesn’t need the extra fat and calories. Yet this is
manifestly not the conclusion that she would have drawn, had she actually managed to
suspend the influence of her desire. Her conclusion is decisively influenced by the very
motive that she takes herself to be suspending and rationally assessing: the desire to be
thin. Her reflective scrutiny of her motive doesn’t enable her to suspend this motive;
on the contrary, it perpetuates the effects of this motive.
Precisely for this reason, it would be perverse to present the anorectic agent as an
example of agential activity. On the contrary, she seems a paradigm of passivity: she
takes herself to be determining her action via choice and rational reflection upon her
motives, but her pathological desire is carrying her reflective thought in its wake.34
34
Here I am assuming that the anorectic does not admit to herself that she is acting on her anorexia;
rather, she attributes her action simply to a desire to be thin. I am also assuming that she disapproves of her
anorexia, so that, if she recognized its manifestations, she would disavow them. Of course, not all anorectics
C O N N E C T I O N B E T W E E N A C T I V I T Y A N D D E L I B E R AT I O N 135
Thus, the anorectic is a good example of how Suspension can fail. Moreover, notice
that if the anorectic agent—who fulfills Inclination, but lacks Suspension—is not
active, then Inclination alone is not a sufficient condition for agential activity. In
other words, she illustrates that given the falsity of Suspension, reflective acts needn’t
be any more active than unreflective ones.
One might worry, however, that the anorectic is an aberrant case, whose condition
is a function of pathology. I think this would be a mistake; the anorectic is simply a
more vivid example of the way in which motives operate in everyday, mundane cases.
As the previous sections have argued, while our motives do not necessitate choice, they
do pervasively influence choice, in ways that the agent often fails to grasp. To illustrate
this point, consider a much more familiar kind of example:
Tom is a manager who must promote one of his employees in order to fill a recent vacancy. The
leading candidates for the job are Candace and Dorothy. As it happens, Tom considers Candace
extremely attractive, whereas he has no such feelings for Dorothy. However, Tom is a judicious
and thoughtful manager, and he knows that perceived attractiveness should have no role in
decisions about promotion. Indeed, he explicitly tells himself that he will not let Candace’s
attractiveness play any part in his decision. Instead, he tells himself, he will make the decision
solely on the basis of Candace and Dorothy’s performances in their current jobs. Carefully
weighing the evidence, Tom ultimately concludes that while both Candace and Dorothy are
admirable employees, Candace seems somewhat more capable, somewhat more attentive, and
somewhat more driven than Dorothy. Consequently, he offers Candace the promotion. So
much for Tom’s thoughts. Let’s suppose, plausibly enough, that Tom’s judgment that Candace is
more capable, more attentive, and more driven is subtly influenced by his attraction to Candace.
Impartial judges, who are not attracted to Candace, would view Dorothy as a marginally superior
candidate.
There is no reason to doubt Tom’s sincerity, in the above example. We can stipulate
that he really does try, he really does struggle, to eliminate any traces of his motives on
his judgment. We can further stipulate that he honestly thinks he succeeds; he thinks he
has suspended the effects of the motive, and has made his decision impartially.
Nonetheless, he fails to do so. The motive (attraction) exerts its influence on Tom’s
reflective thought, even as he tries to distance himself from it.35 Consequently, he
seems to be a passive conduit for forces within.
This kind of case is ubiquitous; a host of psychological studies have demonstrated
that it occurs all the time. Employees who are perceived as more attractive are more
likely to be promoted, to receive better evaluations, to garner higher pay, and
are like this: we could imagine an anorectic who openly admits to herself that she is avoiding the cake because
she is anorectic, and who approves of this action. Arguably, such an anorectic would be active rather than
passive.
35
That is, Tom asks himself whether he should act on his attraction to Candace, decides that he should
not, attempts to suspend the motive, and nonetheless is under the motive’s influence as he reflects. Put
simply, the motive tells him “promote Candace!”, and this is exactly what he does.
136 AC T IO N ’S F I R S T C O N S T I T U T I V E A I M : AGE NT IA L AC TI VI TY
so forth.36 In many such cases, where the agent takes herself to be achieving distance
from a motive that she disavows, but nevertheless is driven by the motive, the agent
will strike us as failing to manifest agential activity.
Indeed, once the agent becomes aware of the above factors, she will typically view
herself as failing to manifest agential activity. An agent who becomes aware that she has
been driven by motives that she would have disavowed will see something amiss in her
action. So both third-person judges and first-personal reflection will find a failure
inherent in these cases.
Thus, we lack good grounds for associating reflective or deliberative agency with
agential activity. Traditional accounts of agential activity need to be rethought.
5. Agential activity as equilibrium

The dominant accounts of agential activity link it to reflective or deliberative action.
However, I have argued that this is a mistake: it relies on a deficient conception of
motivation. Once we operate with a more sophisticated account of motivation,
which treats motives as operating through reflective thought, we see that the activity/
passivity distinction is not the same as the reflective/unreflective or deliberative/non-
deliberative distinction.
If Kant and Locke had been right about Suspension—if deliberation actually did
enable us to suspend our motives, and choose in complete independence of them—
then there would be good reason for singling out deliberative actions as paradigmatic-
ally active. However, given that deliberation typically perpetuates the effects of our
motives, there is no obvious reason to associate deliberative action with agential
activity.
Does this show that the distinction between activity and passivity should be aban-
doned? If being moved by reflective thought isn’t something other than being moved by
motives, should we conclude that there’s no principled reason for attempting to
distinguish between activity and passivity in the production of action?
No. As we saw in Section 1, the idea of agential activity is deeply ingrained in our
thoughts about agency, and is supported by a Kantian transcendental argument about
the nature of agency.
Let’s start with the first point. Consider the following case:
Amy is walking down the street on her way to a meeting. She’s very hungry. However, she is in a
rush, decides that she should hurry on her way, and tries to put the thoughts of food out of her
mind. As she walks along, she doesn’t explicitly consider her hunger; her mind is occupied with
thoughts of what she will say in the meeting. However, as she walks, her attention is drawn to the
36
Marlowe et al. (1996) and a number of other studies reveal that there is a demonstrable bias in
promotion and hiring decisions toward candidates who are judged to be more attractive. Hamermesh and
Parker (2005) discuss a study showing that teachers who are judged more attractive receive higher ratings
from their students on course evaluations.
AG E N T I A L AC T I V I T Y A S E QU I L I B R I U M 137
presence of food; she eyes the restaurants that she strolls by, she finds herself interrupted by
thoughts of food, and so on. As she is about to pass a coffee shop, she suddenly stops and decides
to go in for a snack.
Amy’s hunger manifests itself by drawing her attention to food, by inclining her to
think about food, and so on. Hunger influences reflection by making certain features of
the situation salient, by inclining the agent to give weight to considerations relating to
the acquisition of food, and so on. But in the case of ordinary, mild hunger these effects
are typically unobjectionable, for several reasons. First, and most importantly, hunger
typically attunes one to food only when one actually requires food. Thus, an agent
who is being motivated by her hunger will typically approve of the way in which she is
being motivated. Second, the effects are obvious; it is hardly a surprise to learn that
one’s hunger is leading one to focus on food. Third, the effects are benign; though the
effects of hunger may be distracting, they typically do not interfere with one’s capacity
to reason and to pursue other goals. (Of course, things will be very different in cases of
severe hunger.)
Contrast this with Arpaly’s anorectic. The anorectic case has the same features as our
case of hunger. In the anorectic case, a desire to appear thin exerts itself by affecting
perceptual saliences, judgments, and reflective thoughts; in the hunger case, a desire to
eat exerts itself by affecting perceptual saliences, judgments, and reflective thoughts.
There is no difference in kind between the ways in which motives operate in the
anorectic case and the hunger case; they operate in exactly analogous ways, albeit with
different intensities.
In both cases, then, the agent’s motives channel and guide her reflective thought.
Nonetheless, I think most of us will agree that there is a distinction here: the case of
hunger seems far less problematic than the anorectic case.
Thus, rather than concluding that there is no distinction between the active and the
passive, we might conclude that there is a distinction, but that the distinction is not to
be marked in terms of whether one’s action was caused in a reflective or a non-
reflective manner. Arpaly’s anorectic is highly reflective and deliberate, but seems a
paradigm of passivity. The hungry agent, by contrast, is minimally reflective and
deliberate, yet seems entirely active.
I think these ideas can be developed into a new theory of agential activity. They
seem to indicate that if there is to be a distinction between activity and passivity in the
production of action, we will need something that is more nuanced, something that
comes in degrees, and that takes account of the complex interactions between reflect-
ive thought and affect. The relevant distinction will not be whether the agent suspends
the effects of her motives, for that probably never happens; rather, the relevant
distinction will have something to do with how the agent’s motives affect the agent’s
reflective thoughts.
138 AC T IO N ’S F I R S T C O N S T I T U T I V E A I M : AGE NT IA L AC TI VI TY
5.1 Agential activity as equilibrium

I suggest that the following concept suffices to characterize the relevant difference
between the anorectic and the hungry agent:
(Equilibrium) The agent A’s, and approves of her A-ing. Further knowledge of the
motives that figure in A’s etiology would not undermine her approval of A-ing.
(Disequilibrium) The agent A’s, and currently approves of her A-ing. However, if
she knew more about the motives that figure in A’s etiology, she would no longer
approve of her A-ing.37
The anorectic conceives of her action as resulting from an entirely reasonable desire to
avoid excessive calories, but the more apt description of her action would attribute it to
pathological desires to avoid food, inaccurate self-conceptions, disgust with her own
body, and so on. If the anorectic were to become aware of this, it would likely change
her attitude toward her action. Thus, she is not in equilibrium.38 The hungry agent, on
the other hand, would retain the same attitude toward her action were she to know
more about the motives prompting it. Thus, she is in equilibrium.
A few clarifications are in order. First, the account should be understood as applied
to agents, holding all else constant except giving the agent further information about
the motives figuring in the etiology of the action under consideration. In particular, we
do not want to consider cases in which the agent changes his values. For example,
I would now disapprove of many of the actions that I performed, approvingly, as a
child; but this does not show that the actions were in disequilibrium, for at the time of
action I may have wholeheartedly approved of them.39
Second, notice that the only factor that we are changing here is the amount of
information that the agent has about the etiology of the action. Agents sometimes
37
This account of agency is inspired by Nietzsche; elsewhere (Katsafanas 2011a), I argue that Nietzsche
distinguishes between activity and passivity in roughly this way. I borrow the term “disequilibrium” from
Nozick, who uses it to define a related condition. According to Nozick, an action is in disequilibrium for a
person if “(a) he does (or wants to do) it, yet (b) if he knew the causes of his doing or wanting to do it, this
knowledge would lead him not to do it, or not to want to (or to want not to want to do it, or at least to a
lessening of his want to do it . . . )” (Nozick 1981, 349). Otherwise, the act is in equilibrium. Nozick puts the
notion of equilibrium to different work, suggesting that we might be able to explain goodness in terms of being
in equilibrium and meeting certain additional conditions (Nozick 1981, 350ff.). My task, here, is less ambitious:
I argue that the version of equilibrium defined above offers a characterization of agential activity.
38
Note that I am not claiming that all anorectics are in disequilibrium, just that the anorectic described in
Arpaly’s example is in disequilibrium. It seems to me that some anorectics are in equilibrium—a quick
internet search will reveal a profusion of websites in which well-informed anorectics, who seem to be fully
cognizant of the etiology of their anorexia, nevertheless valorize their anorexia and encourage others to do
the same. On my view, these anorectics are active. See also note 34, above.
39
Equilibrium bears a resemblance to Harry Frankfurt’s notion of wholeheartedness. Roughly, Frank-
furtian wholeheartedness obtains when the agent bears a higher-order attitude of acceptance or approval
toward his lower-order desires (cf. Frankfurt 2004). Equilibrium is more demanding: the approval in question
must be stable in the face of further information about the action’s etiology. For a subtle critique of
Frankfurt’s theory, see Cuypers (2000).
AG E N T I A L AC T IV I T Y A S E QU I L I B R I U M 139
disapprove of a past action not because they learn more about the act’s etiology, but
because they learn more about its consequences. I did not realize that my innocent,
offhand remark would hurt Sarah’s feelings. Now, seeing her upset, I regret the
remark, and wish that I had not made it. But this does not show that my action was
in disequilibrium.
Notice, third, that an agent whose actions are in disequilibrium would not neces-
sarily want to act differently. She might be dissatisfied with her actions not because she
disapproves of what she has done, but because she disapproves of her motives for doing
what she has done. For example, suppose that Sally volunteers in a soup kitchen. She
believes she is volunteering out of a desire to aid the impoverished beneficiaries. Yet a
psychologically adept observer, well acquainted with Sally’s character, would describe
things differently: Sally takes satisfaction in feeling superior to the impoverished recipi-
ents, and her volunteering is in part motivated by this desire. Suppose Sally comes to
realize that one of the desires motivating her action is the desire to feel superior. She
finds this desire reprehensible, and she is no longer able to view her action of
volunteering with approval. Thus, her action is in disequilibrium. However, it
would be inaccurate to say that she wants not to volunteer. Rather, she still wants to
volunteer, but she wants to volunteer out of beneficent motives, rather than self-
serving ones. So, she is in disequilibrium not because she wants to act differently, but
because she wants to act out of different motives.40,41
40
Analogous points apply to more complex cases. Consider the following example: Bill screams at his wife
in anger, and takes this action to be motivated by jealousy. In cool moments, he thinks this jealousy is
unwarranted. He disapproves of screaming out of unwarranted jealousy, so his action is in disequilibrium.
However, unbeknownst to Bill, there are good reasons for his jealousy: though Bill doesn’t admit it to
himself, there are compelling indications that his wife has been unfaithful. Suppose that Bill would approve of
yelling at his wife out of warranted jealousy. What does my model imply about a complex case like this?
I would say that this case is structurally analogous to the above case: the agent is in disequilibrium, because he
wants to act out of different motives (in this case, he wants to act out of warranted jealousy rather than
unwarranted jealousy).
41
Notice that my account implies that an agent could be passive for long stretches of his life. Consider the
following case. Bill meets and becomes strongly attracted to Sarah. Sarah, as it happens, is an ardent and
vociferous vegan, whereas Bill has long regarded veganism as unnecessary and perhaps even vaguely
objectionable. However, Bill now worries that his meat-eating ways will ruin his chances with Sarah. Bill
regards himself as a principled and steadfast person, who will not change his habits merely for the sake of a
potential romance. Thus, he explicitly decides that his desire for a relationship with Sarah is not a good reason
for becoming a vegan. However, reflecting on veganism’s merits, Bill decides on other grounds to become a
vegan. (Perhaps he tells himself that the suffering inflicted on factory-farmed animals is sufficient reason not to
eat them.) Now, assume that, unbeknownst to him, Bill’s romantic desire did influence this decision to such
an extent that, were Bill aware of these facts, he would disapprove of his decision. Bill’s vegan actions are then
in disequilibrium. However, these vegan actions might proceed for many years, structuring large portions of
Bill’s life. Yet, on my account, they will be passive. Is this a problem? Is it odd to think that an agent could be
passive for long stretches of his life? In response, two points are worth emphasizing. First, on my account it
certainly is possible for an agent to be passive in every action that he performs. However—and this is the
second point—such a case would be extremely unusual. In the vegan case, Bill is passive in his larger action
(or project) of being a vegan, but he may be entirely active in the smaller actions that partially constitute this
larger action. For example, when Bill goes to the grocery store to buy vegan food, when he orders tofu at the
restaurant, and so forth, these events, considered as isolate actions, may be in equilibrium, and hence active.
(Thanks to John Brunero for raising this question and suggesting a similar example.)
A final clarification: this account of equilibrium does not associate agential activity
with morally upstanding or praiseworthy activity. An agent could be in equilibrium while
performing an action that is manifestly immoral. For example, a sociopath might inflict
pain on another agent because it pleases him to do so. It is entirely possible that even if
the sociopath knew all of the motives that figure in the etiology of this action, he
would continue to approve of it. Thus, he would be active in performing the action.
With these points in mind, we can see that disequilibrium constitutes a form of
psychic conflict. An agent acts, and approves of his action. However, this approval is
contingent upon his ignorance of the motives that are actually leading him to act. So
there is a conflict between the agent’s attitude toward the action as he takes it to be, and
the agent’s attitude toward the action as it is. Moreover, disequilibrium implies that one
has motives that are influencing one in ways that one would disavow. Thus, there is an
interesting form of conflict between the agent’s reflective and unreflective aspects at
the time of action.
I submit that equilibrium is a necessary condition for agential activity. Equilibrium
seems to offer a characterization of the conditions under which an agent can be said to
be in control of her action. The agent acts, approves of the act, and further knowledge
of the action would not undermine this approval. To speak metaphorically, the agent’s
whole being is behind the action.
5.2 The advantages of this account of agential activity
The proposed account of agential activity as equilibrium has several advantages over
the traditional accounts. First, it does not commit us to the (false) claim that deliber-
ation is capable of suspending the effects of our motives. Second, it does entail that
reflection—or, more precisely, reflective attitudes toward one’s own actions—is partly
determinative of whether one is agentially active. Third, the account of agential
activity as equilibrium offers a correct characterization of certain paradigm cases: it
judges the anorectic agent to be passive, and the hungry agent to be active.
The core idea of this account is that passivity does not involve being moved
unreflectively or independently of reflection. After all, we are moved unreflectively
all the time: relatively few of our actions are preceded or accompanied by anything like
a bout of explicit reflection, and even when we do reflect, this often has the effect of
perpetuating, rather than suspending, our motives. Rather, passivity involves being
moved in a way that conflicts with the reflective attitudes that one would have, if one
recognized how one was being moved.42
42
A central feature of this account is the idea that an individual whose action is in equilibrium cannot act
akratically. If an agent performs an action of which she disapproves, she is not in equilibrium. For this reason,
it would not be inaccurate to characterize my account as maintaining that agential activity is the opposite of
akrasia. Notice that if we do characterize the account in this way, we must classify certain cases of surreptitious
influence by motives—such as Arpaly’s anorectic case and my hiring case, above—as cases of akrasia. Thanks
to Christine Korsgaard for suggesting that my view could be characterized in this way.
AG E N T I A L AC T IV I T Y A S E QU I L I B R I U M 141
5.3 Activity and passivity in reflective thought itself

There is a complication to address. My proposed account of agential activity asks us to
consider whether the agent’s attitude toward the action would be stable if the agent
were aware of additional information about the action’s etiology. Of course, reflecting
on an action’s etiology is itself an action, and is thus susceptible to the forms of influence
by motives that I have discussed above. Consequently, when an agent asks herself
whether her action is in equilibrium, it is possible for this act of equilibrium-assessment
to itself be in disequilibrium.
An example will be helpful. Consider a famous passage from Augustine, in which
Augustine describes his efforts to extirpate his attraction to pride:
Even when I reproach myself for it, the love of praise tempts me. There is temptation in the very
process of self-reproach, for often, by priding himself on his contempt for vainglory, a man is
guilty of even emptier pride. (Confessions, Book X, }38)
Augustine is attempting to avoid the performance of any actions that are motivated
by pride. Thus, suppose that Augustine is engaged in casual conversation with an
acquaintance, and without prompting Augustine begins to speak of his own accom-
plishments. Call this action A. At the time of action, Augustine views himself as
motivated simply by a desire to make conversation. Reflecting on the action at a
later time, though, Augustine recognizes that his A-ing was in fact motivated by pride.
Recognition of this motive for A-ing leads Augustine to disapprove of his A-ing. Thus,
he judges A-ing to be in disequilibrium. Now, let B-ing be Augustine’s higher-order
action of reflecting on and assessing A-ing. Upon further reflection, Augustine notices
that his B-ing is itself motivated by a particular form of pride: the pride that he takes
in chiding himself for falling short of his ideals. Thus, B-ing—that is, Augustine’s
act of assessing his own action and judging it to be in disequilibrium—is itself in
disequilibrium.43
As Augustine illustrates, the very action of trying to determine whether one’s own actions
are in equilibrium can itself be in disequilibrium. This might seem to engender a problem
for my view. Do the equilibrium-assessments themselves have to be in equilibrium in
order for the assessments to be accurate? And if so, won’t this run the risk of launching
us into an infinite regress?
In fact, this is not a problem for my account of activity as equilibrium. It would be a
problem if whether a given (first-order) action, A-ing, were in equilibrium depended
upon whether the (second-order) act of reflecting upon and assessing A-ing were itself
in equilibrium. For, on that view, we would then have to ask whether the (third-order)
act of reflecting upon and assessing the reflecting upon and assessing of A-ing were in
equilibrium, and this would launch us into an infinite regress.
43
Like the soup-kitchen volunteer mentioned above, Augustine does not want to perform a different
action; rather, he wants to perform the same action out of different motives. That is, he still wants to
disapprove of his past prideful action, but he wants this act of disapproval not to be motivated by pride.
However, my account does not depend upon this kind of assessment. It does not
matter whether, like Augustine, the agent actually engages in an episode of reflection
and comes to some assessment of his past action. What matters is simply the status of a
certain counterfactual claim: whether, if the agent were aware of further information
about his motives, he would continue to approve of his action. We are not examining
the agent’s actual thoughts. Instead, we are examining a counterfactual—in which the
agent has more information about the etiology of his action—and asking whether the
agent’s approval of the action then dissipates. In assessing this counterfactual, we take
the closest possible world: we change only the amount of information that the agent has
about the etiology of his action, and we see whether the agent’s approval dissipates or
remains stable.
So, to use the Augustine example: Augustine’s original action of speaking of his own
accomplishments is in disequilibrium, because he would (and in fact does) disapprove
of it given further information about its etiology. His action of reflecting on and
assessing this past action is also in disequilibrium: given further information about its
etiology, Augustine’s approval of this action would (and in fact does) dissipate. But the
question of whether this higher-order action is in disequilibrium has no bearing
whatsoever on the question of whether the original action is in disequilibrium.
With this in mind, we can draw attention to a related point. Notice that whether the
agent’s approval of his action dissipates or remains stable depends on what other
motives the agent has. Put differently, we are not attempting to determine how the
agent would react to his own action in the (impossible) case of not being affected in any
way by his own motives. Take an example. An angry agent may see that his current
action of screaming at Tom is motivated by jealousy. In a cool moment, he would
disapprove of screaming out of jealousy. However, in the midst of his anger, fully
cognizant of why he is screaming, he approves of the action. On my view, this action is
in equilibrium: in the closest world, where all that we have changed is that the agent
has more information about the action’s etiology, he will retain his approval, precisely
because he will still be angry. This seems to be the right result: an agent who clear-
sightedly acts out of jealousy, in full cognizance of what he is doing and why, seems
quite different than an agent whose jealousy surreptitiously guides him in ways that he
would disavow were he cognizant of them.
In sum: higher-order actions, such as the action of assessing one’s actions for
equilibrium, can themselves be in either equilibrium or disequilibrium. In other
words, higher-order actions can be either active or passive. This is no surprise: given
the above account of motivation, we should expect that reflective thought is not
necessarily active. Like overt physical actions, our reflective thoughts are subject to the
vicissitudes of activity and passivity. Yet the question of whether a particular act—be it
higher- or lower-order—is itself in equilibrium is not determined by the agent’s actual
reflective thoughts concerning the act. Consequently, the problem of regress does not
arise.
T H I S AC C O U N T O F AC T I O N Y I E L D S A C O N S T I T U T I V E A I M 143
6. This account of action yields a constitutive aim

We now have an account of agential activity. I promised at the beginning of this
chapter that by focusing on agential activity, we would be able to generate a consti-
tutivist theory. We are now in position to see how.
In Section 2.1, we saw that a deliberating agent aims to determine her action via
choice. We are committed to this aim by the very nature of deliberation: an agent who
deliberates takes herself to be determining her action by choice. She doesn’t view her
reflective deliberations as inert happenings, with no bearing on her forthcoming action;
rather, what it is to deliberate is to attempt to determine one’s actions via choice. If one
didn’t aim to determine one’s actions by choice, then one wouldn’t be engaged in
practical deliberation at all.
But, given this result, aiming to determine one’s action via choice turns out to be a
constitutive aim of reflective action. For, in every instance of reflective action, the
agent aims at determining her action via choice. Moreover, part of what it is for
something to be a token of reflective action is for the agent to aim at determining it via
choice. By the definition of Constitutive Aim, this is just to say that reflective action
constitutively aims at determination via choice.
This, I submit, is a core insight of the Kantian theory of agency. It is integral to
Korsgaard’s account, as explained in the previous chapter. But, if the arguments of this
and the previous chapter are correct, Kantians develop this insight in the wrong way.
They endorse a model of agency that relies on untenable claims about motivation and
agency (as the current chapter has argued), and even setting this aside, the theory fails
on its own terms (as Chapter 4 argued). What we need, then, is a model of agency that
admits an active/passive distinction but does not rely on the discredited claims about
agency. And we have it: I have argued that equilibrium, as defined above, is the correct
analysis of agential activity.
Equilibrium is an analysis of what it is for the agent to determine her action through
choice. Everyone will agree that one way of failing to determine one’s action by choice
is for one’s action not to conform to one’s choice: I decide not to drink at the party, but
end up drinking after all. But the discussion above draws our attention to another way
that one can fail to determine one’s action: one’s choice can be determined by one’s
motives, in a way that one would disavow were one to recognize it. In other words, in
addition to examining the connection between choice and action, we must examine
the connection between the agent and choice.
Accordingly, we have good reason for accepting the claim that reflective action
constitutively aims at agential activity, and that equilibrium is the correct analysis of
agential activity. Equilibrium therefore turns out to be a constitutive aim of reflective
action.
7. Conclusion
This chapter argued that while we need to admit a distinction between activity and
passivity in action, traditional ways of drawing this distinction are misguided. The mere
fact that an act is reflective or deliberative does not entail that it is active, and
unreflective acts are as likely to be active as are reflective ones. Accordingly, the
traditional Kantian/Lockean model of agential activity must be rethought. Nonethe-
less, there is something correct in that model: reflective or deliberative action does aim
at a form of activity. In particular, I have argued that we should understand agential
activity as equilibrium. An agent’s A-ing is in equilibrium iff she approves of her A-ing,
and this approval is stable in the face of further information about A’s etiology. I argued
that this account of agential activity satisfies the desiderata developed in the prior
sections. Moreover, I have argued that agential activity, so construed, is a constitutive
aim of reflective action. In the next chapter, we will begin considering the normative
implications of this conclusion.
6
Action’s Second Constitutive
Aim: Power
The previous chapter argued that an agent actively A’s when she approves of her A-ing,
and this attitude is stable in the face of further information about A-ing’s etiology.
I argued that agential activity, so construed, is a constitutive aim of action.
Of course, Chapters 3 and 4 argued that not every constitutive aim yields substantive
normative results. There, I showed that even if Velleman were right that action
constitutively aimed at self-knowledge, and even if Korsgaard were right that action’s
constitutive standard was self-constitution, these results would be of limited interest.
For they would not enable us to generate any substantive conclusions about what there
is reason to do. We need to ask whether the same problem might apply to the idea that
agential activity is a constitutive aim of action. How much normative content does this
aim generate? Can we use it to generate substantive normative conclusions?
Once we pose these questions, the prospects for our constitutivist theory might look
rather grim. After all, if the constitutive aim of action is agential activity, then actions
are successful to the extent that they are active. But this simply means that an action
would be defective if (a) the agent did not approve of it, or (b) the agent’s approval
would dissipate if she knew more about the action’s etiology.
This standard does rule out certain types of actions: cases of akrasia, actions that are
done in ignorance of their having been motivated by desires that the agent condemns,
and so forth. However, it is crucial to notice that this standard places no constraints on
the basis of the agent’s approval. The approval must be stable in the face of further
information about the action’s history; but other than that, anything goes. So a socio-
path with no regard for others could, presumably, be active with regard to acts such as
murder and thievery: if he approves of his acts of murder and thievery, and this attitude
would be stable in the face of further information about his motives, then he is active.
More generally, we can see that what agential activity prohibits will depend on facts
about what attitudes and values the agent embraces. For example, if an agent disvalues
murder and thievery, then (provided that he is consistent) he would be passive with
respect to these acts; if an agent approves of them, then he would be active with respect
to them. Whether the agent’s action meets the constitutive standard, then, would
depend on which values the agent embraces. This seems problematic, for the notion of
agential activity does not seem to give us a way of assessing the values that it relies upon.
146 A C T I O N ’ S S E C O N D C O N S T I T U T I V E A I M : P OW E R
Thus, other than a general injunction to know thyself (in order to avoid being in
condition [b]), the standard doesn’t seem to generate much normative content.
As a result, it might seem that while our account of activity and passivity illuminates
the notion of agency, it does not help us to understand practical reason. For it gives us a
severely attenuated account of what there is reason to do. Moreover, it fails to yield
universal reasons—reasons that would apply to all agents, regardless of their contingent
motivational states.
Notice, however, that if we could combine the notion of agential activity—which
requires us to assess our actions—with some standard enabling us to assess our values
and attitudes, then the theory would generate substantive and potentially universal
results. So we should ask whether there is a way of assessing the values and attitudes that
one deploys in deliberation.
It is at this point that we must turn to Nietzsche’s works. I am going to argue that by
extracting some ideas from Nietzsche on value and agency, we can show that action has
a second constitutive aim. This second constitutive aim enables us to assess values and
attitudes. When combined with the first constitutive aim, it produces a successful
version of constitutivism.
The second constitutive aim has a counterintuitive nature: it is what Nietzsche calls
“the will to power.” This evocative phrase is easily misinterpreted; it sounds like a
claim about domination, tyranny, and the like. But it is not. As I will argue below, by
claiming that we will power Nietzsche means that we strive not only to bring about
determinate ends, but also to encounter and overcome resistances in the pursuit of
these ends. Indeed, Nietzsche means that, without quite realizing it, we select deter-
minate ends partly in terms of how much resistance they enable us to encounter and
overcome. This surprising and counterintuitive claim will be the subject of this chapter.
I begin, in Section 1, by discussing an apparent tension in Nietzsche’s ethical
thought, a tension which should be familiar even to casual readers of Nietzsche’s
texts: Nietzsche endorses certain values while seeming to deprive himself of any
grounds for this endorsement. In particular, Nietzsche routinely criticizes modern
morality, religion, and contemporary society in general for undermining will to
power; his most common form of objection to an evaluative claim is that it is
inconsistent with power. In short, Nietzsche assesses all other values in terms of
power. But it is hard to see why power should enjoy this privileged status: why
shouldn’t it, too, come up for assessment and possible rejection? This is particularly
pressing given Nietzsche’s claim that there are no objective facts about what is valuable.
The following sections show that we can answer this question by treating power as a
constitutive aim of action. Section 2 begins this task by explaining what Nietzsche
means by “will to power.” I show that Nietzsche grounds power’s normative status in
its relation to willing: power is the “essence” of willing, in the sense that every episode
of willing aims at power. Translating this into contemporary terminology: power is the
constitutive aim of willing. Accordingly, power has a privileged normative status. It is a
value to which we are committed merely in virtue of acting.
A N A P PA R E N T T E N S I O N I N N I E T Z S C H E ’ S E T H I C A L T H O U G H T 147
With the doctrine of will to power explicated, Section 3 asks whether the doctrine is
true. I examine Nietzsche’s arguments for the claim that willing constitutively aims at
power. I suggest that Nietzsche relies on a two-part argument, which incorporates
elements of his drive psychology and claims about the nature of human satisfaction, to
establish that willing does, in fact, constitutively aim at power. Section 4 considers
potential objections to this argument.
1. An apparent tension in Nietzsche’s ethical thought

1.1 Nietzsche gives power a privileged normative status
We typically accept a wide variety of normative and evaluative claims, such as “you
should not lie,” “compassion is good,” and “egalitarianism is valuable.” These claims
purport to have a certain authority over us. They purport to be principles that should
constrain our actions, telling us to perform some and refrain from others. But what
justifies this claim to authority?
Much of Nietzsche’s writing is concerned with this question. Nietzsche insists that
we must study and critically assess our normative claims, asking what compels us to
obey them:
Your judgment ‘this is right’ has a prehistory in your drives, inclinations, aversions, experiences,
and lack of experience: you must ask ‘how did it arise?’ and then ‘what is really driving me to listen
to it?’ (GS 335)
Nietzsche pushes this kind of questioning quite far: he famously wants us to scrutinize
all of our values by engaging in a “revaluation of all values.” As he puts it, “we need a
critique of moral values, for once the value of these values must itself be called into question”
(GM Preface 6).
Nietzsche’s critiques of values take several forms. First, Nietzsche sometimes con-
ducts genealogical investigations of values, revealing the ugly, disagreeable ways in
which certain values arose (GM I–III). Second, he sometimes uncovers hidden tensions
in our sets of values, as when he claims that egalitarianism conflicts with will to power
(TI IX.38; cf. D 163, BGE 62, A 5). Third, he sometimes brings into view the psychic
costs of embracing a value, as when he claims that valuing pity is bad both for the
person doing the pitying and the person being pitied (Z II.3, GS 338).1 In a sense, then,
the mechanics of Nietzsche’s revaluative judgments are clear enough: we acquire a
deep and comprehensive understanding of the value’s history, its connections to other
values, and its effects, and we then assess the value.
1
This list covers some of Nietzsche most frequent modes of critique, but it is not exhaustive. Nietzsche
has other, less direct forms of critique. For example, he will often argue against an ethical theory or a
particular valuation by showing that it relies on false assumptions about agency. See Chapter 1 for a discussion
of these issues. See also Leiter 2002 and 2010 for helpful overviews of Nietzsche’s modes of critique.
However, the final step of this process—the assessment itself—is rather puzzling.
Once we have amassed a set of facts about the value’s history, connections, and effects,
how do we determine which of these facts count as good reasons for accepting or
rejecting the value?
Nietzsche makes it clear that it is not the value’s distasteful origin, as such, that
matters:
The inquiry into the origin of our evaluations and tables of the good is in absolutely no way identical
with a critique of them, as is so often believed: even though the insight into some pudendo origo
certainly brings with it a feeling of diminution in the value of the thing that originated thus and
prepares the way to a critical mood and attitude toward it. (KSA 10:24[31])
Here, Nietzsche claims that acquiring information about a value’s history serves as a
preparation for critically assessing the value. So the mere fact that a value has a disagree-
able history has no direct bearing on the question of whether the value should be
accepted.2
A similar problem arises with respect to the discovery of internal inconsistencies in
our set of values. Nietzsche claims that compassion conflicts with power (Z II.3, GS
338). Suppose he is right. We might respond to this discovery in any number of ways:
we might reject compassion; we might reject power; we might simply live with the fact
that our values cannot all be jointly realized to the fullest extent, and strike a com-
promise. Indeed, Nietzsche explicitly rejects the idea that inconsistency as such is
problematic: he praises those who embrace contradictory values.3
Nor does the mere fact that a value generates psychic costs impugn its authority.
Consider an example: Nietzsche often points out that the valuation of knowledge
generates immense psychic tensions by undermining our most cherished beliefs and
values. As he puts it, “truth has had to be fought for every step of the way, almost
everything else dear to our hearts has had to be sacrificed to it . . . The service of truth is
the hardest service” (A 50).4 However, Nietzsche nonetheless values the pursuit of
2
For an argument against this claim, see Loeb (1995). Loeb raises two very interesting points. First, he
claims that Nietzsche’s genealogies do bear on the status of values: in particular, the genealogies show that our
values have a “plebian ancestry,” and this “proves their disvalue from an aristocratic standpoint” (Loeb 1995,
127). Second, Loeb claims that the “genetic fallacy” charge—that is, the idea that the origins of a value do not
bear on its justificatory status—is itself an evaluative judgment, bound up with a particular moral system. In
particular, Loeb claims that the aristocratic value system would reject the genetic fallacy charge. While Loeb’s
reading is intriguing, I find it ultimately unconvincing. First, it is not clear why contemporary readers should be
moved by the fact that the aristocratic standpoint would regard our values as plebian. Nietzsche makes it quite
clear, after all, that the aristocratic evaluative system is not our evaluative system. Second, I submit that there is
a quite general distinction, not bound up with any particular evaluative system, between the question whether
to continue believing that X and the question how my belief in X came about. The genetic fallacy charge simply
applies this general distinction to the case of values. For discussions of the role of genealogy, see Reginster
(1997), Leiter (2002), Owen (2007), Wallace (2007), and Katsafanas (2011d).
3
See HH 618, BGE 212, BGE 284, GM III.12, KSA 11:26[119].
4
For similar claims, see D 154, 424, 429; GS P 4, 76, 110–13, 344; BGE 4, 24–6, 30, 34, 39, 152, 229;
GM III.13, III.23–7; A 54–5.
knowledge. He writes, “How much truth does a spirit endure, how much does it dare?
More and more that became for me the real measure of value” (EH P 3).5
Thus, Nietzsche’s insistence that we revalue our values raises a question: what,
exactly, are we supposed to be looking for when we make these critical assessments
of values? If disagreeable histories, inconsistencies, and psychic costs do not by them-
selves impugn values, what does?
In fact, Nietzsche provides a clear and explicit answer to this question: we are
supposed to examine the relationship between a given value and will to power. Thus,
Nietzsche declares that the “principle” of revaluation is “will to power” (KSA 12:2
[131]). Or, as he elsewhere puts it: “What is good? Everything that heightens in human
beings the feeling of power, the will to power, power itself” (A 2). Apparently, then,
Nietzsche wants us to investigate our values to determine how they relate to will to
power. If a value conflicts with power, it is to be rejected; if it harmonizes with power,
it is to be accepted.6 The genealogies and other investigations are designed, at least in
part, to uncover these facts about the relationship between power and other values.
Nietzsche pursues this strategy throughout his texts. To choose just a few examples:
Nietzsche complains that the effects of “liberal institutions” are “known well enough:
they undermine the will to power” (TI IX.38). He writes, “well-being as you
understand it—that is no goal, that seems to us an end, a state that soon makes man
ridiculous and contemptible” (BGE 225). What “has been called morality” will
“deprive existence of its great character” (EH IV.4). And he warns that “our weak,
unmanly social concepts of good and evil and their tremendous ascendancy over body
and soul have finally weakened all bodies and souls and snapped the self-reliant,
independent, unprejudiced men, the pillars of a strong civilization” (D 163; cf. BGE
62, A 5). In all of these passages, and many others, Nietzsche implicitly or explicitly
examines the connection between a given value and power.
In fact, Nietzsche’s reliance on will to power becomes even clearer when we notice
that he analyzes a cluster of terms including health, flourishing, and life in terms of power.
For example, he opens the Genealogy by asking “What if a symptom of regression were
inherent in the ‘good’, likewise a danger, a seduction, a poison, a narcotic, through
which the present was possibly living at the expense of the future?” (GM P 6). Making it
clear that these claims about health and flourishing are to be interpreted in terms of will
to power, he continues: “So that precisely morality would be to blame if the highest
power and splendor actually possible to the type man were never in fact attained? So
that precisely morality were the danger of dangers?” (GM P 6).7 As the reference to
5
Compare BGE 227, which claims that honesty is “our virtue, the only one left to us.” See also D 479; GS
88, 107, 123, 283, 301, 319, 324; GM I.1, III.12, III.24; A 50; EH IV.3. For an insightful discussion, see
Anderson (2005).
6
For the moment, I am being deliberately vague about what it means to say that a value “conflicts with”
or “harmonizes with” power. I analyze the notion of conflict in the next chapter.
7
Christopher Janaway points out that we might fairly summarize the Genealogy as arguing that “morality’s
various phenomena are explained as ways in which human beings, like all animals, strive to discharge their
“power” in this passage indicates, when Nietzsche writes about the effects of morality
on human flourishing, he is concerned about its effects on will to power.
Nietzsche’s most frequent euphemism for power, though, is life. The appeal to life is
absolutely pervasive in Nietzsche’s works: he constantly assesses values in terms of
whether they are “harmful to life,” represent “declining life”, a “decrease” in life, and
so on. To give just a few examples, he writes, “every individual may be scrutinized to
see whether he represents the ascending or the descending line of life” (TI IX.33). He
tells us that modern morality is “the formula of decadence” (TI IX.35); it is “hostile to
life” (GM III.11); it “negates life” (CW Preface). Christianity and its associated mores
are “the denial of the will to life become religion” (EH III.CW.2). Christian morality
has “made an ideal of whatever contradicts the instinct of the strong life” (A 5). As a
result, in Christian morality “value judgments have been stood on their heads,” for
“whatever is most harmful to life is called ‘true’; whatever elevates it, enhances, affirms,
justifies it, and makes it triumphant, is called ‘false’ ” (A 9). To drive the point home, he
tells us that
The lightning bolt of truth struck precisely what was highest so far . . . Everything that has
hitherto been called ‘truth’ has been recognized as the most harmful, insidious, and subterranean
form of lie; the holy pretext of ‘improving’ mankind, as the ruse for sucking the blood of life
itself. Morality as vampirism. Whoever uncovers morality also uncovers the disvalue of all values
that are and have been believed; he no longer sees anything venerable in the most venerated
types of man, even in those pronounced holy . . . Finally—this is what is most terrible of all—the
concept of the good man signifies that one sides with all that is weak, sick, failure, suffering of
itself . . . an ideal is fabricated from the contradiction against the proud and well-turned out
human being who says Yes, who is sure of the future, who guarantees the future—and he is now
called evil.—And all this was believed, as morality!—Ecrasez l’infame! (EH IV. 8)
Life—whatever that means—can flourish or decline, be healthy or sick, increase or

regress, be strong or weak. And modern morality is presented as “false,” “infamous,”
“harmful,” and so on precisely because it causes life to decline, be sick, regress, be
decadent.8
Of course, Nietzsche is well aware that there are different ways of conceiving of life.
He critiques Darwin’s claim that life aims at self-preservation (BGE 13), Spencer’s
interpretation of Darwin on life (KSA 11:35[34]), the Stoic interpretation of life (BGE
9), the Christian interpretation of life (GM I, A 8–20, A 47–50), and so on.9 He argues,
power and maximize their feelings of power under the exigencies of their own characters and externally
imposed constraints” ( Janaway 2007, 145).
8
For additional examples, and to acquire a sense of just how pervasive this mode of critique is in
Nietzsche’s corpus, see Z II.12; BGE 2, 13, 44, 293; GM III; TI II (entire), V.1, V.4, V.5, X.4–5; A 2,
5–6, 63; EH IV.7–9; KSA 12:5[71].
9
For discussions of Nietzsche’s critiques of Darwin and Spencer, see Moore (2002) and Richardson
(2004).
instead, that “the essence of life” is simply “its will to power” (GM II.12).10 He tells us
that “life itself” is a striving for “power” (A 6). He speaks of “a world whose essence is
will to power” (BGE 186), and asserts that “the will to power” is “the will of life”
(BGE 259); “life simply is will to power” (BGE 259). Thus, Nietzsche offers a fair
summary of his project in the following passage:
Life itself is to my mind the instinct for growth, for continuance [Dauer], for accumulation of
force [Häufung von Kräften], for power; where the will to power is lacking there is decline. It is my
contention that all the supreme values of mankind lack this will—that the values which are
symptomatic of decline, nihilistic values, are lording it under the holiest names. (A 6)
Thus, when Nietzsche evaluates a value or individual in terms of “life,” life should be
understood as a shorthand for will to power.11 This is explicit in the final sentence of
the following passage:
In its measure of strength every age also possesses a measure for what virtues are permitted and
forbidden to it. Either it has the virtues of ascending life: then it will resist from the profoundest
depths the virtues of declining life. Or the age itself represents declining life: then it also requires
the virtues of decline, then it hates everything that justifies itself solely out of abundance, out of
the overflowing riches of strength . . . In the narrower sphere of so-called moral values one
cannot find a greater contrast than that between a master morality and the morality of Christian
value concepts: the latter developed on soil that was morbid through and through (the Gospels
present us with precisely the same physiological types that Dostoevsky’s novels describe), master
morality (‘Roman,’ ‘pagan,’ ‘classical,’ ‘Renaissance’) is, conversely, the sign language of what
has turned out well, of ascending life, of the will to power as the principle of life. (CW Epilogue)
Although the analysis of life in terms of will to power is perfectly clear from the
published works cited above, Nietzsche’s notebooks make it even more explicit:
What are our valuations and tables of moral values really worth? What results from their rule? For
whom? With regard to what?—Answer: for life. But what is life? Here a new, more definite
version of the concept ‘life’ is needed. My formula for it is: life is will to power. (KSA 12:2[190]/
WLN 95–6)
There is nothing in life that has value except the degree of power—assuming, precisely, that life
itself is the will to power. (KSA 12:5[71]/WLN 119).
10
In Zarathustra, he gives this a more imagistic expression: “where there is life is there also will: not will to
life but—thus I teach you—will to power” (Z II.12).
11
For a sustained and insightful investigation of these claims, see Hussain (2011) and Richardson
(forthcoming). Hussain writes that for Nietzsche “talk of the will to power is clearly meant as a shorthand,
a statement of the fundamental tendency, a tendency that is essential to life, towards expansion, domination,
growth, accumulation of force, and power” (Hussain 2011, 151). Schacht recognizes this point, writing “Life,
as [Nietzsche] construes it, is ‘will to power’ in various forms—an array of processes all of which are
‘developments and ramifications’ of this basic tendency” (Schacht 1983, 367). Schacht continues: Nietzsche
“takes ‘life’ in this world to be the sole locus of value, and its preservation, flourishing, and above all its
enhancement to be ultimately decisive for determinations of value” (Schacht 1983, 359). Geuss, too, writes
“There is little doubt that ‘Life’ . . . in Nietzsche does seem to function as a criterion for evaluating moralities”
(Geuss 1997, 10).
So there can be no doubt that when Nietzsche claims that a given value undermines
flourishing, health, or life, he means that it undermines will to power.
1.2 Is power’s privileged normative status compatible with Nietzsche’s other views?
Thus, on Nietzsche’s account, we revalue a value by inquiring into the relationship
between the value and power. If the value conflicts with power, it is to be rejected; if it
promotes power, it is to be accepted. However, this strategy presupposes that power
has a privileged normative status. Usually, when we discover inconsistencies between two
values A and B, we have three options: rejecting A, rejecting B, or striking a com-
promise. But with power, Nietzsche maintains that things are different: the only
possible response to a discovery of conflict is to reject the value that conflicts with
power. In this sense, power has a privileged normative status: while all other values can
come up for review and possible rejection, power cannot.
However, the claim that power has a privileged normative status seems to be in
tension with another aspect of Nietzsche’s view—that there are no objective values,
Whatever has value in our world now does not have value in itself, according to its nature—
nature is always value-less, but has been given value at some time, as a present—and it was we who
gave and bestowed it. (GS 301)
There is nothing good, nothing beautiful, nothing sublime, nothing evil in itself, but . . . there are
states of soul in which we impose such words upon things external to and within us . . . (D 210)
In these passages, Nietzsche claims that all values arise from human activities. If there
were no human beings, there would be no values. Presumably for this reason,
Nietzsche denies that there can be any objective facts about what is valuable:
There are altogether no moral facts. Moral judgments agree with religious ones in believing in
realities where there are no realities. (TI VII.1)
There are no moral phenomena at all, but only a moral interpretation of phenomena. (BGE 108)
So Nietzsche maintains that all values are created by human activities, and that, as a
result, there are no objective facts about what has value.
This denial of objectivity seems to be in tension with Nietzsche’s insistence that
power has a privileged normative status. After all, the claim that there are no objective
values suggests that that there are no constraints on what can be valued. If all values are
created, then there seems to be no reason for preferring will to power to any other
value.
In short, Nietzsche accepts the following three claims about value:
(1) Power has a privileged normative status.
(2) There are no objective values, or there are no objective facts about what is
valuable.
(3) All values are created by human activities.
Unfortunately, these three claims seem to be in tension with one another. If there are
no objective values, and all values are created, why should power enjoy a privileged
status?
It is important to note that Nietzsche is not committed to (1)–(3) merely in virtue of
a stray, passing remark here and there. Rather, the entire gist of Nietzsche’s thinking
points in this direction. The last section gave some indication of just how pervasive (1)
is in Nietzsche’s works. His commitment to (2) and (3) is at least as obvious. To cite just
a few examples, (2) and (3) are central themes in the Genealogy, which asks “under what
conditions” we “devise[d] these value judgments” (GM P 5, emphasis added). More-
over, Nietzsche repeatedly calls for “new philosophers” who are “strong and original
enough to provide the stimuli for opposite valuations and to revalue and invert ‘eternal
values’” (BGE 203). In BGE 211 he says that true philosophers “create values”; in GS
335 he enjoins us to “limit ourselves to the purification of our opinions and value
judgments and to the creation of tables of what is good that are new and all our own.”12 Thus,
even casual readers can see that (1)–(3) reach right to the heart of Nietzsche’s work;
these are abiding, central commitments.
So we have an interpretive problem, which has bedeviled generations of commen-
tators. The structure of the possible solutions should be clear enough. First, we could
conclude that Nietzsche is simply confused, and ends up endorsing inconsistent
propositions. This is not very plausible—these aren’t arcane or deeply hidden inconsist-
encies, of the sort that might escape a philosopher’s notice; the tensions are palpable.
Second, we could try to read Nietzsche in a way that avoids committing him to one
of the claims: we could deny that he gives power a privileged status, or try to read him
as claiming that power is objectively valuable, or some such. Unfortunately, this
approach isn’t very plausible, as each claim is amply supported by a range of textual
evidence. Nietzsche’s commitment to (2) and (3), in particular, is undeniable, and the
prior section gave ample proof of Nietzsche’s commitment to (1).
So what is the solution? The remaining possibility is that the tension between (1)–(3)
is merely apparent. In the following sections, I will argue for this possibility. I show that
Nietzsche has a way of reconciling the claim that there are no objective values with the
claim that power has a privileged status. In particular, he grounds power’s privileged
status in facts about philosophical psychology. To see this, we will need to explicate the
concept of will to power. I turn to that task in Section 2. Before proceeding to that
stage, though, I pause to consider whether we might respond to the puzzle by
complicating the way in which we understand Nietzsche’s commitment to claim (1).
1.3 Might Nietzsche regard his privileging of power as unjustifiable?
As I mentioned above, we could respond to this interpretive puzzle by attempting to
show that Nietzsche actually rejects one of claims (1)–(3). With that in mind, consider
12
Additional relevant passages are D 3, 210, 453; GS 116, 143, 296, 301, 345, 353, 380–2; Z I.15, II.12;
BGE 62, 212, 260; GM (the Preface and First Essay are especially relevant); TI Preface; A 2–3.
claim (1). Here we should draw a distinction. The material in Section 1.1 establishes
that Nietzsche’s ethical critiques rely on the idea that power has a privileged normative
status. But there are two ways of interpreting this fact: on the one hand, Nietzsche
might think we can justify the claim that power has a privileged normative status. On
the other hand, Nietzsche might give power a privileged normative status while
maintaining that this stance cannot be justified. Call this the no-justification reading.
It admits the obvious point that Nietzsche assesses values in terms of power; it claims,
however, that Nietzsche does not think there is any way of justifying the normative
privileging of power. Thus, if someone wanted to value the reduction of power instead,
there would be no objection. As Brian Leiter, who argues in favor of the no-justifica-
tion interpretation, puts it, “at bottom, [Nietzsche] has nothing to say to those readers
who don’t share his evaluative tastes” (2000, 290). In effect, this reduces Nietzsche to
nothing more than a skilled rhetorician, attempting to sway us to his side.
If the no-justification interpretation were correct, then Nietzsche would endorse
only a limited version of (1), which we might state as follows: Nietzsche himself gives
power a privileged normative status, but does not regard this privileging of power as
justified. Thus, there would be no tension between claims (1)–(3).
I think Nietzsche denies the no-justification view: he not only prefers power, but
maintains that we can justify the claim that power should be preferred. In support of
this interpretation, let me offer three considerations.
First, consider the urgency and importance that Nietzsche attributes to revaluation.
His books are full of exhortations to scrutinize our values, apocalyptic warnings about
the threat of impending nihilism and the collapse of traditional valuations, and so forth.
On the no-justification reading, Nietzsche labors over these writings and warnings
while believing that there is ultimately no reason for him to do so; it’s just a contingent
fact about him that he prefers valuations couched in terms of power rather than, say,
compassion. I find this exceptionally difficult to believe. If preferences and evaluations
are ultimately arbitrary, why all the fuss about investigating them?
Now, a proponent of the no-justification view might respond by pointing out that
people sometimes do have exceptionally strong preferences that they regard as unjusti-
fiable. We could try to explain the urgency Nietzsche attributes to revaluation in that
way; people sometimes devote themselves to a love, a cause, a career, without thinking
that there is any rational justification for so doing. But notice that in this particular
context, such an approach would be decidedly odd. For Nietzsche isn’t just pursuing
some personal project, throwing himself passionately into some unjustifiable task;
rather, he is trying to reach others, to convert others to this project, and insisting that
they will be converted if they devote sufficient attention to their own valuations. It is
hard to believe that this project would be psychologically sustainable if Nietzsche
viewed it as groundless. After all, why would the Nietzsche who insists that we
scrutinize and critique our values and preferences not apply this kind of critique to
himself? Why wouldn’t he question the status of this allegedly unjustifiable preference
for power? It’s hard to imagine the psychological state of an agent who writes
thousands of pages insisting that we diagnose our own values while viewing this very
valuation of diagnosis as nothing more than an unjustifiable, arbitrary preference.
In short: the first problem with the no-justification reading is that it’s hard to believe
that Nietzsche himself, who is so critical about the unexamined, undefended evaluative
claims made by others, would rest his own ethical critiques and exhortations on an
undefended and indefensible evaluative claim. This brings us to a second, related
problem: the no-justification interpretation does not seem textually adequate. For
example, consider section 9 of the Antichrist. There, Nietzsche tells us that embracing
Christian morality necessitates “closing one’s eyes . . . to the sight of incurable false-
hood,” having a “faulty perspective,” and being unable “to respect reality at any point
or even to let it get a word in.” He writes,
Wherever the theologians’ instinct extends, value judgments have been stood on their heads and
the concepts of “true” and “false” are of necessity reversed: whatever is most harmful to life is
called “true”; whatever elevates it, enhances, affirms, justifies it, makes it triumphant, is called
“false.” (A 9)
This hardly sounds like a mere expression of personal preference. The concepts of truth
and falsity are consistently invoked: according to Nietzsche, the Christian moralist is
not just someone with a different set of preferences; rather, he is someone who closes
his eyes to truth and embraces falsehood in order to preserve his value judgments. In
ignoring facts about what is life-enhancing, the Christian moralist makes a mistake. He
is, as Nietzsche elsewhere writes, corrupted by “mendaciousness,” dependent on
“faith,” and unable critically to reflect on his own values (EH IV.1; see also BGE 44,
227; GS 335; HH I.226). For “Christianity grew up on this false soil, where every
nature, every natural value, every reality ran counter to the deepest instincts of the
ruling class” (A 27). The Nietzsche who writes “[Zarathustra’s] doctrine and his alone
has truthfulness as the highest virtue” does not seem to limit himself to the airing of
personal preferences (EH IV.3, emphasis added). Indeed, he writes, “the truth speaks
out from me—But my truth is terrible, because lies have been called truth so far.—
Revaluation of all values: that is my formula for an act of humanity’s highest self-
examination” (EH IV.1). Thus, when Nietzsche quite bluntly writes, “What is
good? Everything that heightens in human beings the feeling of power, the will to
power, power itself ” (A 2), it’s hard to read this as an assertion that Nietzsche regards as
unjustified.13
13
Some of these passages can be interpreted in line with the no-justification reading if we take Nietzsche
to be critiquing the factual presuppositions upon which certain valuations are based. For example, we might
read the critique of the Christian moralist as follows: both the moralist and Nietzsche happen to prefer values
that are life-enhancing; but the moralist is confused about which particular values actually do enhance life;
thus, the dispute is factual rather than evaluative. Leiter (2002, 159–61) attempts to read many of these
passages in this manner. I agree with Leiter in a limited sense: if we had good independent reasons for reading
Nietzsche as endorsing a no-justification view, then we could interpret these passages in ways that were
consistent with that view. However, I think the more natural reading of the passages is the one I suggest
above: saying that value judgments have been “stood on their heads,” that the Christian moralist is
Finally, there is a third and perhaps more decisive problem with the no-justification
reading. Consider the Hegelian/Kierkegaardian point that I explained in Chapter 1,
Section 1.3: in order for our goals to inspire real allegiance, we need to see these goals
as having more authority than mere whims. We need to see something as non-
arbitrarily structuring and constraining our choices. I suggested that Nietzsche accepts
this picture of evaluation: unless evaluations can be justified in a way that does not
collapse the distinction between norm and whim, the very practice of evaluation will
be untenable. But if we read Nietzsche as treating power as, in effect, a mere whim, this
distinction will indeed be collapsed.
For these reasons, I interpret Nietzsche as maintaining a strong version of claim (1).
In other words, Nietzsche maintains both that power has a privileged normative status
and that this claim can be justified; it is not merely an expression of personal preference.
2. Will to power as a constitutive aim

We have a puzzle: the claim that power has a privileged normative status sits uneasily
with Nietzsche’s claims that there are no objective facts about what is valuable and that
all values are created by human activities. I promised, above, that we could resolve this
tension by interpreting Nietzsche as grounding power’s privileged normative status in
facts about the nature of agency. To see how this works, let’s begin by examining the
nature of will to power.
2.1 What is will to power?
What does Nietzsche mean by “will to power”? It is important not to be misled by
the surface connotations of the term “power.” In ordinary discourse, the claim that
people will power would suggest that people strive to dominate, tyrannize, and
subjugate others. This is not what Nietzsche has in mind. Power is a term of art, for
Nietzsche; he gives it a special sense.
In order to introduce this special sense, it is helpful to start with a mélange of
characteristic quotations on power. Nietzsche describes will to power in language that
seems deliberately vague; he associates power with a family of terms, such as “giving
form,” “expanding,” “imprinting,” “overcoming,” “mastering,” and “shaping.”14 He
writes that will to power is “the will’s forward thrust and again and again becoming
master over that which stands in its way” (KSA 13:11[75]/WLN 213). Moreover,
Nietzsche does not attribute any specific end to those who will power: he claims that
we can will power in the pursuit of a diverse range of activities, such as artistic creation,
the pursuit of knowledge, asceticism, and so on (cf. GM II.17–18 and passim).15
“mendacious,” that Christian values run “counter” to “natural value[s],” that “lies have been called truths so
far,” and so on, is most straightforwardly read as the claim that Christian moralists are wrong about their values.
14
GM II.18, GS 349, BGE 259, Z II.12, KSA 13:11[75].
15
For discussions of this idea, see for example Muller-Lauter (1971), Stegmaier (1994), and Aydin (2007).
W I L L T O P OW E R A S A C O N S T I T U T I V E A I M 157
In order to see exactly what will to power is, we will need to determine what these
characterizations of will to power have in common. Although Nietzsche’s descriptions
tend to be rather elliptical, he does repeatedly and insistently emphasize two points
about will to power.
First, Nietzsche claims that will to power can never enjoy permanent satisfaction,
but instead involves perpetual striving. Nietzsche often makes this point by contrasting
the will to power with the wish to abide in a certain state, or the desire to preserve
oneself:
The wish to preserve oneself is the symptom of a condition of distress, of a limitation of the really
fundamental instinct of life which aims at the expansion of power, and in so doing frequently risks
and even sacrifices self-preservation. (GS 349)
Physiologists should think before putting down the instinct of self-preservation as the cardinal
instinct of an organic being. A living thing seeks above all to discharge its strength—life itself is will
to power; self-preservation is only one of the indirect and most frequent results. (BGE 13)16
Rather than seeking a placid state of achievement, will to power inclines us perpetually
to seek more. Many passages phrase this point in terms of self-overcoming: whenever
some end is achieved, the agent immediately wishes to go beyond. BGE 73 tells us that
“whoever reaches his ideal transcends it eo ipso.” BGE 27 speaks of the “law of the
necessity of ‘self-overcoming’ in the nature of life.” In Zarathustra, Nietzsche puts the
point more poetically: “And life itself told this secret to me: ‘See’, it said, ‘I am that
which must always overcome itself ’ ” (Z II.12).17 In these passages, Nietzsche empha-
sizes that will to power involves perpetual striving.18
Second, as Bernard Reginster (2006) has persuasively argued, will to power mani-
fests itself as a particular form of striving: striving for resistances or obstacles. Nietzsche
makes this clear in his notebooks:
The will to power can manifest itself only against resistances; therefore it seeks that which resists
it . . . (KSA 12:9[151]/WLN 165)
The will is never satisfied unless it has opponents and resistance. (KSA 13:11[75]/WLN 213)
16
Compare the following passage: “A condition once achieved would seem to be obliged to preserve
itself—Spinoza’s law of ‘self-preservation’ ought really to put a stop to change: but this law is false, the
opposite is true. It can be shown most clearly that every living thing does everything it can not to preserve
itself but to become more—” (KSA 13:14[121]).
17
BGE 257 associates the “enhancement of the type ‘man’ ” with “continual self-overcoming.” See also
GS 363; Z I.15, II.2; BGE 10, 175, 259; GM II.11; TI IX.38; EH I.8.
18
Alexander Nehamas notes that will to power involves perpetual striving. As he puts it, “willing as an
activity does not have an aim that is distinct from it; if it can be said to aim at anything at all, that can only be
its own continuation. Willing is an activity that tends to perpetuate itself, and this tendency to the
perpetuation of activity . . . is what Nietzsche tries to describe by the obscure and often misleading term
‘will to power’ ” (1985, 79). Heidegger seems to be making a similar point when he writes that “will to power
is will to will” (1979, vol. I, p. 37). Although my own interpretation of will to power differs from those of
Nehamas and Heidegger, I think they are quite right to draw attention to the fact that will to power involves
perpetual striving.
What man wants . . . is an increase in power; out of that will man seeks resistance, needs
something to oppose him . . . (KSA 13:14[174]/WLN 264)
Analogous passages are present in the published works:

[Strength or power is] a desire to overcome, a desire to throw down, a desire to become master, a
thirst for enemies and resistances and triumphs. (GM I.13)
The highest type of free man should be sought where the highest resistance is constantly
overcome . . . (TI IX.38)
A tablet of the good hangs over every people. Behold, it is the tablet of their overcomings;
behold, it is the voice of their will to power. Praiseworthy is whatever seems difficult to a
people . . . (Z I.15)19
When Nietzsche refers to “resistances,” he means impediments or challenges to one’s

ends. For example, the ascetic manifests will to power by overcoming his body’s
resistance to suffering; the artist overcomes the difficulties inherent in turning a blank
canvas into a painting; the scientist overcomes the obstacles and challenges inherent in
her quest for understanding. In the passages quoted above, Nietzsche makes it clear that
willing power involves actively seeking these resistances, in order to overcome them. As
he puts it in Zarathustra,
That I must be struggle and a becoming and an end and an opposition to ends—alas, whoever
guesses what is my will should also guess on what crooked paths it must proceed. Whatever I create
and however much I love it—soon I must oppose it and my love; thus my will wills it . . . Only
where there is life is there also will: not will to life but—thus I teach you—will to power. (Z II.12)
The will to power doesn’t aim at any definite end, but merely at the overcoming of
resistances, difficulties, or obstacles.
Thus, the will to power manifests itself as the aim of seeking challenges and resist-
ances that impede the pursuit of one’s ends.20 Of course, one doesn’t want these
challenges or resistances to serve as permanent impediments to one’s ends; rather, one
wants to overcome the impediments. As Nietzsche puts it, will to power is “the will’s
forward thrust and again and again becoming master over that which stands in its way”
(KSA 13:11[75]/WLN 213). For example, the ascetic doesn’t want to face his resist-
ance to pain and then freeze in the face of it; rather, he wants to overcome this
resistance, by managing to inflict the pain on himself. So, too, the scientist doesn’t want
to take on some difficult problem and find herself unable to solve it; rather, she wants
19
GM II.11 emphasizes that will to power involves overcoming resistance; as Nietzsche puts it, “life
functions essentially, that is in its basic functions, by injuring, assaulting, exploiting . . . ” BGE 259 makes the
same points.
20
This might sound like a claim about instrumental efficacy—it might sound as if willing power is just
willing to overcome whatever obstacles happen to lie between you and your end. As the passages cited above
make clear, Nietzsche means something much stronger: to will power is actively to seek resistances, in order to
overcome them.
to solve the problem. This is why Nietzsche above mentions the “crooked paths” on
which the will proceeds: we simultaneously desire to achieve various ends and to
experience resistance to the achievement of these ends.
In short, Nietzsche seems to identify willing power with the activity of perpetually
seeking and overcoming resistance to one’s ends. I therefore conclude that, as Bernard
Reginster puts it, “will to power, in the last analysis, is a will to the very activity of
overcoming resistance” (Reginster 2006, 127). In other words, to will power is perpetually
to seek to encounter and overcome resistance in the pursuit of some end.21 Accord-
ingly, when Nietzsche tells us that will to power is simply “the will’s forward thrust and
again and again becoming master over that which stands in its way” (KSA 13:11[75]/
WLN 213), when he refers to “giving form,” “shaping,” and so on, he means to bring
out the central feature of will to power: to will power is to aim at the activity of overcoming
resistances to ends.22
2.2 Will to power as a claim about the essence of willing
Now we know what it is to will power: an agent wills power by seeking to encounter
and overcome resistances. However, there is another important component to
Nietzsche’s claims about will to power: he presents his will to power thesis as an
elucidation of the essential nature of willing. Nietzsche makes this point in a number of
passages, repeatedly claiming that will to power is the “essence” [Wesen, Essenz] of
willing. To choose just four passages, he writes,
The genuinely basic drive of life [Lebens-Grundtriebes] . . . aims at the expansion of power . . . the will
to power . . . is just the will of life [Wille des Lebens]. (GS 349)
All ‘purposes,’ ‘aims,’ ‘meaning’ are only modes of expression and metamorphoses of one will
that is inherent in all events: the will to power. To have purposes, aims, intentions, willing in
general, is the same thing as willing to be stronger, willing to grow—and, in addition, willing the
means to this. (KSA 13:11[96]/WLN 217)
Everything that happens out of intentions can be reduced to the intention of increasing power. (KSA
12:2[88]/WLN 76)
21
One might object to this analysis on the grounds that Nietzsche often treats will to power as a capacity
rather than an activity (thanks to Scott Jenkins for raising this point). In response, I doubt Nietzsche would
countenance any strict distinction between activities and capacities: as GM I.13 suggests, having a capacity for
X is analyzed in terms of actually X-ing. In particular, if willing power is aiming to encounter and overcome
resistance, then it certainly makes sense to talk of individual’s differing capacities for willing power, but this
talk of capacities will be spelled out in terms of the resistances that the individuals actually do encounter and
overcome.
22
Bernard Reginster has offered convincing arguments for this interpretation of will to power; my
analysis of will to power relies on his work. I have not provided enough textual evidence to convince a
skeptic that this reading is correct; Reginster himself does so (Reginster 2006). While I am in agreement with
Reginster on the nature of will to power, the following sections develop an account of will to power’s
normative authority that differs from Reginster’s view. For another insightful analysis of will to power, see
Gerhardt (1996).
Where there is life is there also will: not will to life but—thus I teach you—will to power.
(Z II.12)23
How are we to interpret the claim that will to power is the essence of willing?
A passage from Heidegger is extremely helpful on this point. Heidegger writes,
The expression ‘will to power’ does not mean that, in accord with the usual view, will is a kind of
desiring that has power as its goal rather than happiness or pleasure. True, in many passages
Nietzsche speaks in that fashion, in order to make himself provisionally understood. But when he
makes will’s goal power instead of happiness, pleasure, or the unhinging of the will, he changes
not only the goal of will but the essential definition of will itself. In the strict sense of the
Nietzschean conception of will, power can never be pre-established as will’s goal, as though
power were something that could first be posited outside the will . . . The expression ‘to power’
therefore never means some sort of appendage to will. Rather, it comprises an elucidation of the
essence of will itself. (Heidegger 1979, vol. I, 42)
In this passage, Heidegger notes that the will to power thesis is a description of the
essential nature of willing. Yet Heidegger warns us not to interpret Nietzsche in the
same way that we interpret philosophers such as Bentham and Mill, who held that
happiness or pleasure are the final ends of action. This point will be worth dwelling on,
for a moment.
It would be natural to assume that when Nietzsche claims that will to power is the
essence of life or of willing, he means that the final end of each goal-directed act is
power. In other words, there is an instrumental relation between power and all other
ends. Consider, by analogy, the form of psychological hedonism endorsed by Bentham
and Mill. Psychological hedonists claim that pleasure is the final end of each goal-
directed act. Thus, if we consider any act—the pursuit of knowledge, the pursuit of
friendship, the pursuit of a dish of ice cream—the psychological hedonist will claim
that these goals are pursued simply as means to the attainment of pleasure. Analogously,
it would be natural to read Nietzsche as claiming that all goals are pursued for the sake
of power. On this reading, Nietzsche is simply making a substitution into Mill’s theory,
replacing every occurrence of “happiness” or “pleasure” with “power.” So whereas
Mill argues that we pursue friendship as a means to pleasure, Nietzsche would be
arguing that we pursue friendship as a means to power.
However, Heidegger argues that this reading is mistaken; Nietzsche is making a
subtler and more interesting point. For notice that power is not something that, strictly
speaking, would be intelligible apart from willing. Happiness or pleasure can be
conceived independently of willing: we can understand what it is to be happy or
23
A few more examples: “What man wants, what every smallest part of a living organism wants, is an
increase in power” (KSA 13:14[174]/WLN 264). “All driving force is will to power” (KSA 13:14[121]/
WLN 256). “Striving is nothing other than striving after power” (KSA 13:14[81]). “Life, as the form of being
that is best known to us, is specially a will to the accumulation of force: this is the lever of all the processes of
life [ . . . ] Life [ . . . ]: strives for a maximum feeling of power: is essentially a striving for more power: striving is
nothing other than striving for power” (KSA 13:14[82]/WLN 248; punctuation is Nietzsche’s).
pleased without presupposing the concept of willing. We cannot, however, understand

what it is to encounter and overcome resistance apart from the concept of willing. The
very concept of resistance is unintelligible except in relation to a determinate end:
resistance is always resistance to the achievement of some end. Accordingly, an agent
who does not have any ends—an agent who is not engaged in willing—cannot face
any resistance. It follows that will to power can only manifest itself in the pursuit of
some determinate end: in order to seek resistance at all, we must also seek something
other than power.
Once we keep this point in mind, it becomes difficult to see what the instrumentalist
conception of power could even mean. The will to power is the will to encounter and
overcome resistance. But this will cannot manifest itself as the blank aim of seeking
resistance, for resistance is only intelligible in connection to some other, more deter-
minate end. As John Richardson puts it, “power, as something willed by every drive,
‘lacks content,’ requiring a contingent filling out from some given case” (Richardson
1996, 24). In order to pursue resistance at all, I must already be engaged in the pursuit
of some other, more determinate end. For example, I might will to encounter and
overcome resistance in the pursuit of knowledge; or I might will to encounter
and overcome resistance in the pursuit of friendship. It would be distorting, at best,
to describe this relationship by saying that I pursue friendship as a means to encounter-
ing and overcoming resistances, for the relevant forms of resistance are not intelligible
apart from the determinate, first-order end of friendship. So, while the instrumentalist
conception suggests that I decide to pursue friendship because I want power and see
friendship as a means to power, this seems inaccurate. Rather, I find myself already
pursuing friendship, already under the influence of motives that incline me toward that
end. In willing power, I pursue that end in a certain manner: I will the end, and also
will to encounter and overcome resistance in the course of pursuing the end.
Thus, to say that a person wills power is not to describe an instrumental relation; it is
not to say that we select various goals as means to power. Rather, the will to power
thesis describes a formal or structural relation between two ends. John Richardson
highlights this point, noting that the claim that we will power is not a claim about
what we will; it is a claim about how we will (Richardson 1996, 21). Whenever a person
wills an end, this episode of willing has a certain structure. It consists not only in the aim
of achieving some end, but also in the aim of encountering and overcoming resistance
in the pursuit of that end. This is why, as Heidegger notes, claims about will to power
are not claims about “some sort of appendage to will”; they are “an elucidation of the
essence of will itself.”24
24
This is not to deny that some of Nietzsche’s phrasings suggest an instrumentalist interpretation of will to
power. For example, when Nietzsche writes that agents strive for the “maximum feeling of power
[Machtgefühl]” (GM III.7), it’s natural to read him as claiming that agents perform various actions in order
to achieve a separate end, power. Nevertheless, as I’ve explained above, attention to what Nietzsche means
by will to power shows that this instrumentalist conception makes little sense. As so often happens with
Given that the will to power is a formal relation, which describes the structure of
willing, it becomes easier to see what Nietzsche means when he claims that the essence
of willing is will to power. He means that power is a formal aim present in each
instance of willing. Whenever an agent wills an end, the agent aims not only at
attaining that end, but also at encountering and overcoming resistance in the pursuit
of that end. This is why Nietzsche says that “all ‘purposes,’ ‘aims,’ ‘meaning’ are only
modes of expression and metamorphoses of one will that is inherent in all events: the
will to power” (KSA 13:11[96]; emphasis added). He is not claiming that every goal is a
means to power; rather, he is claiming that whenever we will any goal at all, we express
will to power by also willing resistance to that goal.
2.3 Resolving the interpretive problem
Of course, the claim that we actively seek obstacles and resistances is highly counter-
intuitive. I will address that fact in a moment, by asking whether Nietzsche’s claims
about willing are defensible. First, though, I want to say a bit about what the conse-
quences of Nietzsche’s claims about power would be.
What is interesting about Nietzsche’s remarks on will to power is that they seem
designed to show that power is the constitutive aim of willing. Suppose that one
wanted to offer a constitutivist argument about willing. The first step would be to show
that there is some aim that is essentially involved in each instance of willing. Notice that
this is exactly what Nietzsche’s arguments concerning the will to power are designed to
establish. As we just saw, Nietzsche contends that each instance of willing aims at
power. Moreover, the will to power doctrine is a claim about the essential nature of
willing: it is a description of the form or structure that every episode of willing manifests.
But, by the definition of Constitutive Aim, this is just to say that power is the
constitutive aim of willing.25
Suppose Nietzsche’s arguments succeed in establishing that willing constitutively
aims at power. If the constitutivist argument form were valid, then Nietzsche would be
entitled to conclude that power has a privileged normative status. In drawing that
conclusion, Nietzsche would not have to rely on the idea that power is an objective
value. Rather, the argument would rely simply on the idea that insofar as an agent wills
an end, the agent is committed to treating power as a standard of success for willing.
Again, this seems to be exactly what Nietzsche does conclude about power.26
Nietzsche denies that there are objective values, but treats power as the one standard
Nietzsche’s texts, the reading that initially seems most natural turns out to be problematic; Nietzsche intends
something subtler.
25
Chapter 1 defined constitutive aim as follows: Let A be a type of attitude or event. Let G be a goal. Then
A constitutively aims at G iff (i) each token of A aims at G, and (ii) aiming at G is part of what constitutes an
attitude or event as a token of A.
26
Consider the following remark from the notebooks: “Replacement of morality by the will to our goal,
and consequently to the means of gaining it. Replacement of the categorical imperative by the natural imperative”
(KSA 12:9[27]/WLN 146).
of evaluation that readily meets challenges to its authority. And we can now see why.
Nietzsche is grounding power’s privileged evaluative status in an incapacity: it is the one
aim that we cannot give up, insofar as we are engaged in willing.
Surprisingly, then, Nietzsche’s claims about will to power and revaluation seem to
be linked by a constitutivist argument. The premises, the argument form, and the
conclusion are all just what we would expect, if Nietzsche were a constitutivist.
Moreover, recall that we began examining Nietzsche’s notion of will to power
because Nietzsche embraces a triplet of seemingly inconsistent claims about value:
(1) Power has a privileged normative status.
(2) There are no objective values, or there are no objective facts about what is
valuable.
(3) All values are created by human activities.
Notice that we can render these three claims consistent by interpreting power as the
constitutive aim of action. If power is the constitutive aim of action, then it has a
privileged normative status: it is the one standard that is intrinsic to willing. Yet power
is not an objective value; rather, it arises from the fact that our actions have a certain
structure. In that sense, the value of power is created by us.27
2.4 A remark on strategy
Before proceeding, it will be useful to clarify my goals. Below, I will be arguing that we
can take some central themes from Nietzsche’s work and interpret them along con-
stitutivist lines. It is important to be clear that I am not arguing that a few stray remarks,
here and there, evoke constitutivism. That would be of limited interest: Nietzsche’s
texts are so rich and disparate that one could find a few passages that, when stripped of
their context, would seem to support almost any ethical view. My strategy is entirely
different: above, I argued that there are three central themes present throughout
Nietzsche’s works that seem deeply and obviously inconsistent, but that are revealed
to be entirely consistent if we assume that Nietzsche is pursuing a constitutivist strategy.
This counts strongly in favor of interpreting Nietzsche as a constitutivist.
Nonetheless, decisively establishing that Nietzsche is in fact a constitutivist would
require a great deal of exegetical work. We would have to examine passages that seem
to create trouble for this reading; we would have to show that alternative interpret-
ations of Nietzsche’s ethical theory are less successful; we would have to show that
Nietzsche’s understanding of reflective agency is compatible with an endorsement of a
constitutivist position; we would have to show that an acceptance of constitutivism is
compatible with Nietzsche’s general skepticism about ethical theory; and much else
besides. I offer a few remarks on these potential objections in the Appendix, and
27
That is, we can say that power’s normative status is “created” in the sense that its normativity arises
merely from our inescapably aiming at it. In the next chapter, I will show that there is an even stronger sense
in which Nietzsche’s claim that values are created is true: we have reason to adopt a variety of values not
themselves derived from will to power, and there is room for free variation here.
interested readers should turn there after reading the remainder of this chapter.
However, offering a full-fledged defense of my interpretation of Nietzsche would
require a level of detailed textual work that would be out of place in the current
volume; indeed, it would require a book-length project of its own.28 My primary goal
here is to defend a new version of constitutivism, rather than to establish that Nietzsche
himself would have endorsed each detail of this version of constitutivism. Accordingly,
skeptical readers can take the constitutivist interpretation as a rational reconstruction of
central themes in Nietzsche, a way of rendering consistent some deeply puzzling
features of his texts, a way of resolving a generations-long interpretive puzzle. I will
not attempt to convince a skeptic that Nietzsche understood himself to be offering
exactly the argument that I present below. (In this respect, my strategy parallels that of
many contemporary Kantians, Humeans, and Aristotelians, who often aspire to show
that their Kant-, Hume-, and Aristotle-inspired theories are successful, rather than that
the details of these theories were actually endorsed by the respective philosophers.)
3. A Nietzschean argument for constitutivism

By interpreting Nietzschean will to power as a constitutive aim, we can resolve the
interpretive problem: Nietzsche’s three claims about value are consistent. Values aren’t
out there in the world, but arise from us; yet we are inescapably committed to one
value, power.
So now we know that the Nietzschean view is coherent. The question is whether
the view is true. Is it true that willing constitutively aims at power?
The hardest part of any constitutivist argument is showing that action really does
have a constitutive aim. That is doubly difficult here, for Nietzsche singles out a
decidedly counterintuitive aim: how could it possibly be true that in each instance of
action, we aim at encountering and overcoming resistances?
We are inclined to think of that the essential function of willing is to bring something
about. On this conception of willing, the paradigmatic case of willing has the following
form: I desire some end X, I see that I could get X by doing Y, so I will to do
Y. Willing aims merely at effecting a change in the world, so that the world conforms
to my desires.
As we saw in Chapter 1, John Stuart Mill endorsed this conception of willing. He
writes,
All action is for the sake of some end, and rules of action, it seems natural to suppose, must take
their whole color and character from the end to which they are subservient. (Mill 2002, 2)
28
I offer defenses of my interpretation of Nietzsche in Katsafanas (2011a, 2011b, 2011d, and forthcom-
ing a). I think the most significant challenge in attributing this constitutivist argument to Nietzsche is
reconciling it with his account of agency and moral psychology. In a book manuscript in progress (The
Nietzschean Self: Agency and the Unconscious), I defend an interpretation of Nietzsche on agency and moral
psychology that is compatible with a constitutivist argument.
A NIETZSCHEAN ARGUMENT FOR CONSTITUTIVISM 165
So, the point of action is to bring about some desired goal; the rules of action, the
standards of success for action, pertain solely to how well the action brings about this
goal.
This is certainly a natural conception of willing, and it makes Nietzsche’s claims
about will to power seem patently absurd. First, Nietzsche’s will to power thesis
doesn’t focus on bringing anything about at all. It focuses simply on seeking and
surmounting obstacles or resistances. This is highly counterintuitive. Don’t we simply
seek to achieve goals, and view overcoming resistance as a necessary yet regrettable
condition of achieving these goals? Wouldn’t we prefer to avoid resistance if we could?
Moreover, if the point of willing is to bring something about, then why would we
seek resistance to our own willing? After all, an action is successful to the extent that it
brings about its goal. But will to power makes it less likely that we will achieve our
goals, by making these goals more difficult to achieve. For this reason, will to power, so
far from being the essence of willing, seems to be a perversion of willing: it is self-
conflicted, self-defeating, defective willing.
So, at any rate, it appears. If Nietzsche’s will to power thesis is going to be at all
plausible, he will need to offer some argument for it.
As a preliminary step, it is important to note that Nietzsche is well aware that his
claim will strike most readers as counterintuitive. Nietzsche is not trying to elucidate
our ordinary conception of willing; rather, he is attempting to reveal the true structure
of willing, which he believes has been misunderstood. This is why, in BGE 19,
Nietzsche bemoans our tendency to treat willing “as if it were the best known thing
in the world.”29
With that in mind, let’s reconstruct Nietzsche’s argument for the claim that we will
power. The argument has three stages. First, Nietzsche makes a series of conceptual
claims about the nature of a certain kind of motivational state, the drive. Second, given
the structure of drives, it turns out that any drive-motivated action will in fact have the
constitutive aim of encountering and overcoming resistance. Third, Nietzsche argues
for an empirical claim, namely that all human actions are drive-motivated activities. If
this is right, then it turns out that human action has the constitutive aim of encounter-
ing and overcoming resistance.
29
It is also important to keep in mind that the will to power typically manifests itself in ways that the agent
will not recognize. Recall that in Chapter 3, I emphasized that constitutive aims need not be conscious
desires; they can, indeed, be subpersonal aims. When Nietzsche claims that we will power, he is certainly not
claiming that we all have conscious desires for power; rather, as I will argue below, Nietzsche is claiming that
we all aim at power, often without recognizing it. Relevant here is Janaway’s discussion of the distinction
between Schopenhauer’s term ‘will’ and his term ‘desire.’ Janaway notes that for Schopenhauer, “ ‘will’ is not
a psychological term, and is certainly not equivalent to ‘desire’; it denotes a tendency to end-directed
behavior throughout nature, only one of whose manifestations . . . is the psychological phenomenon of
desire. Schopenhauer would say that desires are but one specialized instance of the will in nature, and indeed
he remarks as misguided the attempt to see every kind of will throughout nature as a kind of desiring. Even in
the human case, the will to life does not manifest itself primarily as a desire for life . . . So when Nietzsche asks
us to replace ‘will to life’ with ‘will to power’, he may count on our complicity with the notion that some
natural tendency more primitive, less psychological than desiring is meant” ( Janaway 2007, 157).
3.1 Step one: the nature of drives

To begin, we need to see what drives are. Elsewhere, I have argued that drives have
two central features. First, drives are dispositions that induce affective orientations.
Second, drives do not dispose the agent to bring about any determinate end, but
instead dispose the agent to engage in particular forms of activity. While I won’t have
space to offer a full defense of this reading here, I will reconstruct the key points.30
First, Nietzschean drives are dispositions that induce affective orientations in agents.
Nietzsche often describes drives in agential language:
Anyone who considers the basic drives of man to see to what extent they may have been at
play . . . will find that all of them have done philosophy at some time—and that every single one
of them would like only too well to represent just itself as the ultimate purpose of existence and
the legitimate master of all the other drives. For every drive wants to be master—and it attempts to
philosophize in that spirit. (BGE 6)
It is our needs that interpret the world; our drives and their For and Against. Every drive is a kind
of lust to rule; each one has its perspective that it would like to compel all the other drives to
accept as a norm. (KSA 12:7[60])
These passages suggest, to some readers, that Nietzsche treats drives as homunculi.31 For
example, Clark and Dudrick argue that drives “exhibit agency of a sort,” and are
“homunculi” or “proto-persons” (2009, 264–5). Similarly, Thiele attributes a robust
form of agency to drives, including even the idea that drives have “political relations”
with one another (1990, 57). He claims that each drive “has its will to dominate and exploit
its competitors . . . the ruling drive(s) provides its own agenda and worldview . . . The
individual . . . is a battleground of competing drives, each with its own perspective” (57–8).
These readings can appear quite natural. After all, Nietzsche seems to attribute to
drives the capacity for reasoning, adopting perspectives, and so on, all of which would
require that drives are independent agents. However, I have elsewhere argued that it is
a mistake to interpret Nietzsche’s language in this way. Drives are always embodied in
agents. When Nietzsche claims that drives adopt perspectives, philosophize, and so on,
he is referring to the way in which a drive, operating through an agent, can affect the
agent’s perspective, philosophizing, and so forth.
To see this, it helps to consider Schopenhauer’s discussion of drives. In a wonderful
discussion of the human reproductive drive, Schopenhauer writes that this drive leads
us to pursue sexual partners not by blindly impelling us to this end, but by fostering
a distorted orientation toward the world. The reproductive drive “creates illusions
[Illusionen schafft ]” (WWR, vol. II, 566) or generates a “delusion [Wahn]” (WWR,
vol. II, 541).32 As Schopenhauer explains:
30
For a full explanation and defense of these claims, see Katsafanas (forthcoming a).
31
For interpretations of this sort, see Thiele (1990), Poellner (1995), and Clark and Dudrick (2009).
32
I use WWR to refer to The World as Will and Representation.
Here then, as in the case of all drive, truth assumes the form of delusion, in order to act on the
will. [Also nimmt hier, wie bei allem Instinkt, die Wahrheit die Gestalt des Wahnes an, um auf den Willen
zu wirken.] It is a voluptuous delusion which leads a man to believe that he will find greater
pleasure in the arms of a woman whose beauty appeals to him than in those of any other, or
which, exclusively directed to a particular individual, firmly convinces him that her possession will
afford him boundless happiness . . . The character of drive is here so completely present, namely
an action as though in accordance with the conception of an end and yet entirely without such a
conception, that whoever is urged by that delusion often abhors it and would like to prevent the
end, procreation, which alone guides it . . . (WWR, vol. II, 540)
Schopenhauer claims that although the human reproductive drive aims at reproduc-
tion, when we are in the grip of this drive, we typically do not believe that we are
pursuing reproduction. We believe we are pursuing happiness, or pleasure, or posses-
sion of a particular individual. Schopenhauer claims that this belief—or, as he puts it,
this delusion—is produced by the drive itself. In other words, the reproductive drive
manifests itself by leading us to conceive of our potential sexual partners as supremely
alluring, capable of providing us with great happiness and pleasure. The reproductive
drive moves us not by generating a blind urge or disposition to copulate, but by
producing desires and other emotions, by influencing the way in which we perceive
potential partners, and so on.
Thus, Schopenhauer writes that “in all sexual love, instinct holds the reins, and
creates illusions [bei aller Geschlechtsliebe der Instinkt die Zügel führt und Illusionen schafft]”
(WWR, vol. II, 566). But the phenomenon is not restricted to the sexual: Schopen-
hauer believes that all drives work in this fashion. Accordingly, he claims that animals
acting on drive “are urged not so much by an objective, correct apprehension, as by
subjective representations which stimulate the desire . . . and that accordingly they are
urged by a certain delusion . . . ” (WWR, vol. II, 541).
So the drive manifests itself by inducing, in the agent, a distorted perspective, a
delusion. Nietzschean drives operate in just the same way; thus, Nietzsche will speak of
affects and drives as “coloring,” “gilding,” “lighting,” and “staining” the world. These
terms suggest that affects and drives highlight or even alter aspects of an experience (see,
for example, GS 7, 139, 152, 301). Drives produce selective orientations toward the
world—orientations that, in some cases, are distorting enough to qualify as illusions.
This is what Nietzsche means when he claims that drives adopt perspectives: drives
influence the perspectives through which the agent experiences the world.
In sum, drives can be understood as dispositions that generate affectively charged
orientations, or perspectives. This brings us to the second feature of drives: Nietzsche
emphasizes that drives continuously seek expression. He speaks of the “ebb and flood”
of our drives, their “play and counterplay among one another,” their “growth and
nourishment” (D 119, BGE 6, and passim). Drives are almost inevitably associated with
active forces, impulsions, and pressures seeking discharge.
When Nietzsche writes of drives in this way, he has the following point in mind:
drives arise independently of external stimuli, and, once they have become active, they
will seek discharge. The fact that drives are active and do not arise in response to
external stimuli creates a problem. In many cases, a drive will be active in conditions
that do not provide the agent with appropriate objects with which to satisfy the drive.
Just as we can be hungry when there are no opportunities to eat, we can be angry when
there are no occasions for anger. For example, suppose the aggressive drive is active in a
situation in which the individual has not been threatened or provoked. Nietzsche tells
us that the drive will seek outlets, seek objects on which to vent itself.
Schopenhauer’s discussion of the sexual drive illustrates this phenomenon: when the
drive is active, it leads individuals to conceive of their potential partners as supremely
alluring, regardless of whether they would see these people in the same way in a cool
moment. Just so, the aggressive drive will lead individuals to perceive their surround-
ings as warranting aggression: thus, the driver who cuts in front of one is perceived as
deliberately provoking, or the inattentive cashier is seen as personally vindictive. In
these ways, the drive leads the agent to see the situation as warranting the drive’s
characteristic form of activity.
In order to express this point, it is helpful to draw on some terminology from Freud.
Freud distinguishes between the aim (Ziel) and the object (Objekt) of a drive.33 While
Nietzsche doesn’t employ this terminology, the idea is present in his work.34 The
drive’s aim is the relatively constant end of the drive, in terms of which it is distin-
guished from other drives. Drives aim at their characteristic forms of activity: the
aggressive drive aims at manifesting aggressive activity, the sexual drive aims at mani-
festing sexual activity, and so on. In order to express this activity, the drive needs to find
some object: the aggressive drive might vent itself on another driver, or a cashier, or a
participant in an athletic game. In other words, we can distinguish between what the
drive seeks (i.e., the manifestation of some characteristic form of activity) and how the
drive expresses this aim (i.e., by finding some object upon which to vent its form of
activity).
3.2 Step two: drive-motivated actions aim at power
So drives are motivational states that aim at their own expression, and take various
objects merely as chance occasions for expression. The aggressive drive does not seek
any particular object; it merely motivates the agent to engage in aggressive activity. Of
course, in order to engage in aggressive activity, one needs to find someone or
something to be aggressive toward. But the object upon which the drive is vented is
comparatively unimportant: what the drive seeks is simply expression. In other words,
what the drive motivates one to seek is not aggression toward some particular person or
object; rather, the drive motivates one to seek the expression of aggressive activity.
(Indeed, as we saw above, Nietzsche maintains that drives often manifest themselves by
33
See “Instincts and their Vicissitudes,” in Vol. 14 of the Standard Edition.
34
I defend this claim in Katsafanas (forthcoming a).
distorting the agent’s view of the environment, so that the agent perceives the situation
as an appropriate venue for expressing the drive’s aim.)
As a result, an activity that is motivated by a drive does not aim to attain some object
that would put an end to the activity. Many desire-motivated actions do aim solely at
bringing about some end. When I am motivated by a desire to end a headache, I seek
an end—having an aspirin—that brings the course of action to a stop. When I am
motivated by a desire to get to my office, I perform an action—walking—that is
designed to accomplish this goal. However, drive-motivated activities are quite differ-
ent. The agent who is being motivated by the aggressive drive isn’t seeking to achieve
any particular end, but is seeking merely to express aggressive activity. Activity
motivated by a drive aims not at the achievement of a determinate end, but at the
performance of the activity itself.
So drives aim at expression, in the sense that they aren’t satisfied by the attainment of
any one determinate object; rather, they want continuous attainment of objects. We
can mark this feature of drives by saying that drives induce the agent to engage in
process-directed actions, rather than goal-directed actions. A goal-directed act is an act
that aims solely at the attainment of some definite goal; thus, the act would cease once
its object is achieved. A process-directed act is an act that aims not merely at some goal,
but also at the performance of a characteristic form of activity. In process-directed
activities, the attainment of the goal would not necessarily bring the process to an
end.35
It will be easiest to illustrate this point with an example. Take the activity of running
a marathon. Marathon running has two important features. First, the goal seems
unimportant if divorced from the process: there is nothing particularly valuable
about being in the state of having travelled twenty-six miles from one’s starting
point. But, second, the goal acquires importance when it is considered as part of the
process. Running a marathon requires strenuous exertion, the overcoming of great
resistance, and the experience of sometimes intense pain. But marathon runners
typically do not view these aspects of running as objectionable; on the contrary, part
of the point of running a marathon is that one encounters these resistances and
obstacles. In the usual case it is not that the runner values the state of having run
twenty-six miles, and views the pain as a necessary, but regrettable, aspect of running.
Rather, the runner values the whole activity of encountering obstacles and holding
35
The distinction between state- and process-directed actions is roughly analogous to Aristotle’s distinc-
tion between production (kinesis) and activity (energeia). Activities are events that are complete at every
moment, whereas productions are events that are never complete at a moment, but must always complete
themselves over some period of time (Metaphysics 1048b). Aristotle offers a grammatical criterion for
distinguishing the two types of event: if the present progressive description of the event implies the past
perfect description of the event, then the event is an activity. Otherwise, it is a production. For example,
seeing is an activity: if I am seeing, then I have seen. But building a chair is a production, not an activity: if
I am building a chair, it does not follow that I have built a chair (after all, I may be only halfway done). Using
this terminology, my process-directed actions are roughly analogous to Aristotle’s activities, whereas my goal-
directed actions are roughly analogous to Aristotle’s productions.
herself to a course of action despite the pain involved in doing so. This is why the
runner chooses to run twenty-six miles, instead of twenty-six feet; the latter would be
too easy, would not be challenging. The runner views the marathon running as
valuable partly because it requires encountering and overcoming resistances and
obstacles. Thus, in the normal case, it would be distorting to view a marathon runner
as aiming solely to have run twenty-six miles. That goal is valued only as a part of the
whole process of running.36
With this in mind, consider again Nietzsche’s will to power thesis. The will to
power thesis claims that whenever we act, we aim not only to achieve some determin-
ate end, but also to encounter and overcome resistances. For a goal-directed act, this
claim would be absurd: if I am seeking merely to bring about some end, then it would
be perverse to will resistance to that end. But for a process-directed act, things are
crucially different. Process-directed acts typically do involve goals that need to be
achieved, but these goals would be unimportant if divorced from the process. What
matters is the process, and engaging in the process requires finding objects upon which
to direct the process. In this sense, manifesting the process requires perpetually seeking
new obstacles, resistances, or challenges, upon which the process can be directed. (For
example, if the aggressive drive is to be manifest, it must continuously find new objects
for aggression.) Thus, drive-motivated actions do in fact aim at encountering and
overcoming resistance.
So, while a purely goal-directed act would seek to minimize resistances, a process-
directed activity involves an active desire to encounter and overcome resistances.37 For
if you aim to engage in some process, you must seek the objects upon which the
process can be directed. This is why Nietzsche describes the will to power as “a will to
overcome, a will that has in itself no end . . . a processus in infinitum, an active determin-
ing” (KSA 12:9[91]/WLN 155).38 Or, as he elsewhere puts it, “the will to power can
manifest itself only against resistances; therefore it seeks that which resists it” (KSA 12:9
[151]/WLN 165).
The link between the will to power thesis and the drive psychology should now be
clear. Will to power is not an independent drive, but a description of the form that all drive-
motivated actions take. This is why Heidegger was entirely correct to write that “power
36
This is not to deny that attainment of the goal is valued; a marathon runner who collapses a mile before
the finish line will feel defeated. My point, however, is that attainment of the goal is not the agent’s sole aim.
To clarify, consider another example: a writer who wants to author a poem. This process-directed act
certainly has a goal: the completion of a poem. But the writer doesn’t aim solely at the attainment of the goal.
For example, he would not be satisfied by being presented with a poem that someone else had written. The
writer wants the whole process of writing.
37
Resistances arise from two directions, so to speak. There are external resistances, arising from the fact
that the world is recalcitrant. But there are also internal resistances, arising from competition amongst our
motivational states, our drives. If each drive seeks expression, and if there are limits to the number of drives
that can be expressed at any one moment, then in order to express itself a drive will have to “overpower”
competing drives.
38
See also the many passages associating will to power with self-overcoming, cited above.
can never be pre-established as will’s goal, as though power were something that could
first be posited outside the will . . . The expression ‘to power’ therefore never means
some sort of appendage to will. Rather, it comprises an elucidation of the essence of
will itself.” To say that we will power is to say that we are motivated by drives.
In sum, the will to power thesis describes the structure of drive-motivated actions. If
an action is drive-motivated, then it aims at power. If, as Nietzsche seems to maintain,
all actions are drive-motivated, then all actions aim at power.
This entails that will to power is the constitutive aim of drive-motivated actions. For
drives are motivational states that aim at their own continuous expression, and aiming
at continuous expression entails aiming to encounter resistances to overcome. The
conclusion is simple: drive-motivated activities aim at encountering and overcoming
resistance. This is part of what makes an activity qualify as drive-motivated. By the
definition of Constitutive Aim, this is just to say that drive-motivated activities have the
constitutive aim of overcoming resistance.
3.3 Step three: all human actions are drive-motivated
We have just investigated the structure of a certain kind of motivational state, the drive.
We have seen that any drive-motivated activity will have the constitutive aim of
overcoming resistance. Now we are faced with the obvious question: are human
actions drive-motivated?
In fact, Nietzsche argues that all human actions are drive-motivated. If this claim is
true, then it follows that human action has the constitutive aim of overcoming resist-
ance. In this section, I will ask whether the claim is true.
Nietzsche has several different kinds of argument for the claim that human actions
are drive-motivated.39 His most familiar form of argument is this: he will take a central
type of human action which appears to have nothing to do with drives, which appears
to be conditioned merely by the agent’s self-conscious thought, and show that, upon
closer scrutiny, the form of activity is motivated by some drive. For example, he argues
that the pursuit of knowledge is motivated by what he calls the ascetic drive (cf. GM
III). Taken together, these various exercises might amount to an inductive proof that
human actions are drive-motivated.40 But I won’t be examining these arguments here.
For I think Nietzsche has a deeper argument, which focuses on the nature of satisfaction
(or happiness).41 The argument claims, in essence, that the nature of human satisfaction
indicates that human actions are drive-motivated.
39
Nietzsche argues that all actions—indeed, all life—manifests will to power. However, to avoid
introducing more complexity into an already complex topic, I will focus on the case of human action.
40
Janaway notes that will to power plays the principal explanatory role in the Genealogy. As he puts it,
“throughout the three treatises morality is explained by psychological mechanisms which are diverse manifest-
ations of will to power: ressentiment, internalized cruelty, and the conscious adoption of self-denial as an ideal”
( Janaway 2007, 144).
41
For an interesting discussion of Nietzsche and Schopenhauer on happiness, see Reginster (2006).
The first claim in Nietzsche’s argument is:
(A) There is no state such that being in that state provides lasting satisfaction.
This claim is in direct opposition to a philosophical thesis that came to prominence in
the work of the eighteenth- and nineteenth-century utilitarians, and that continues to
exert influence in contemporary philosophical work. I will put the thesis in an
unfamiliar way, and then translate it into more familiar language. Here is the unfamiliar
way: Bentham and his followers thought that there were states such that human beings
seek to enter into and abide in those states.
Here is a more familiar way of putting the point. Consider psychological hedonism.
Psychological hedonists assume that human beings seek pleasure, where pleasure is
typically conceived as an experiential state. Bentham wrote, “Nature has placed
mankind under the governance of two sovereign masters, pain and pleasure. It is for
them alone to determine . . . what we shall do” (Bentham 2007, 1).
Many philosophers reject psychological hedonism. However, psychological hedon-
ism contains a thesis that is accepted even by those who allow things other than
pleasure to serve as goals. This is the assumption that human beings seek solely to
achieve goals, or to attain certain states. This assumption was explicit in eighteenth-
and nineteenth-century philosophical works: we find it in Bentham, in Mill, and in
Schopenhauer. For example, Schopenhauer writes, “we call [the will’s] hindrance
through an obstacle placed between it and its temporary goal, suffering [Leiden]; its
attainment of the goal, on the other hand, we call satisfaction, well-being, happiness”
(WWR vol. I, 309). Thus, Schopenhauer concludes that complete happiness would
require “a final satisfaction of the will, after which no fresh willing would occur”
(WWR vol. I, 360).
The claim, here, is that human beings seek to achieve and abide in certain states, such
as the state of experiencing pleasure or the state of having one’s desires fulfilled. But
this very assumption—that there are states such that being in those states provides
satisfaction—came under attack by Schopenhauer. Schopenhauer argued that the
attainment of a goal provides, at most, a temporary satisfaction: “with the satisfaction
[of a desire], the desire and therefore the pleasure cease” (WWR vol. I, 319). For if
pleasure is the state of desire-fulfillment, and desires dissipate once they are fulfilled,
then pleasure is inescapably fleeting. Thus, there can be no state the occupation of
which provides lasting satisfaction.
In other words: Schopenhauer argued for claim (A). And Schopenhauer famously
concluded, from (A), that happiness is impossible. As he put it, “life swings to and fro
like a pendulum between pain and boredom, and these two are in fact its ultimate
constituents” (WWR vol. I, 312).42
42
Here I am passing over one important component of Schopenhauer’s argument, namely the claim that
when we have no desires to pursue, we experience a distinctive type of pain: boredom. It is the most
Like Schopenhauer, Nietzsche argues for (A). But he draws a very different conclu-
sion from the truth of (A). If we follow Schopenhauer and the early utilitarians in
conceiving of happiness as an experiential state, then we may be led to Schopenhauer’s
conclusion. But, as Nietzsche puts it in one of his characteristically pithy remarks,
“Man does not pursue happiness [Glück]—only the Englishman does that” (TI II.12).43
The English utilitarians’ conception of happiness is defective, Nietzsche thinks. The
defect is not the familiar one, not the fact that they conceive of happiness as an
experiential state. Rather, the defect is that they conceive of happiness as a state.44
Nietzsche rejects the state-based conception of happiness. He argues that happiness
is not a state at all; happiness obtains when we are engaged in efficacious pursuit of a
goal, not when we attain the goal. As he puts it,
It is not the fulfillment [Befriedigung] of the will that causes pleasure [Lust] (I want to fight this
superficial theory—the absurd psychological counterfeiting of the nearest things—), but rather
the will’s forward thrust and again and again becoming master over that which stands in its way.
The feeling of pleasure lies precisely in the non-fulfillment [Unbefriedigung] of the will, in the fact
that the will is never satisfied unless it has opponents and resistance. (KSA 13:11[75])45
Or, more succinctly:

What is happiness?—The feeling that power increases—that a resistance is overcome. Not
contentedness, but more power; not peace at all, but war . . . (A 2)
We think we are satisfied by stasis, by being in the state of having attained some goal; but
in fact we are satisfied by successful activity, by actively seeking and attaining goals.46
So whereas Schopenhauer concludes, from (A), that satisfaction is impossible,
Nietzsche concludes, from (A), that satisfaction is processual. That is,
(B) There are processes such that engaging in those processes provides lasting
satisfaction.
astonishing fact about intelligent animals that we experience boredom, and Schopenhauer is one of the few
philosophers to have investigated the consequences of this fact.
43
See also NCW Epilogue 2, where Nietzsche writes, “How repulsive pleasure is now, that crude, musty,
brown pleasure as it is understood by those who like pleasure, our ‘educated’ people, our rich people, and our
rulers!”
44
Nietzsche’s objection to the utilitarians is twofold: first, he rejects their conception of happiness;
second, he rejects the idea that happiness (no matter how it is conceived) should serve as a criterion of
evaluation for actions. Here, I am concerned only with the first point. See BGE 225 for a discussion of the
second point.
45
Compare the following passage: “Human beings do not seek pleasure [Lust] and avoid displeasure
[Unlust] . . . What human beings want . . . is an increase of power . . . driven by that will they seek resistance,
they need something that opposes . . . ” (KSA 13:14[174]/WLN 264).
46
Perhaps it is relevant that Dostoevsky, whom Nietzsche referred to as “the only psychologist from
whom I had something to learn” (TI IX.45), offered the same view: “Man is a frivolous and incongruous
creature, and perhaps, like a chess player, loves the process of the game, not the end of it. And who knows
(there is no saying with certainty), perhaps the only goal on earth to which mankind is striving lies in this
incessant process of attaining, in other words, in life itself, and not in the thing to be attained” (Dostoevsky
2006, Section IX).
In other words, there are processes such that engaging in these processes provides
satisfaction that lasts as long as the process lasts.
Nietzsche argues for (B) by engaging in psychological diagnoses. He develops
detailed and highly insightful investigations of the vicissitudes of our actions. Some
of these investigations show Nietzsche at his best and most psychologically acute. Here,
I will mention just one of these investigations: Nietzsche’s analysis of creative actions.
Nietzsche believes that creative actions provide the best illustration of (B)’s truth.
Creativity is, Nietzsche emphasizes, a central and paradigmatic case of human activity.
Moreover, he claims that creative activity enables a “new happiness” (GS Preface),
constituting “the great redemption from all suffering, and life’s growing light” (Z II.2).
Creative activity generates lasting satisfaction. But what is creative activity? Import-
antly, creativity is not directed toward the attainment of definite ends, but toward the
active confronting of new challenges. As Bernard Reginster puts it,
Nietzsche emphasizes the case of artistic creation because artists are not simply inventive
individuals who overcome limitations or difficulties only when they have to, but they actively
look for them because they value creative activity itself . . . The creative individual . . . deliberately
seeks to confront and break boundaries, to expand the domain of human experience, to
overcome limitations hitherto unchallenged, or to vanquish resistance perhaps once thought
unassailable. (Reginster 2006, 191–2)
In other words, artistic creation does not involve seeking to abide in some determinate
state, but instead requires actively seeking resistances. Thus, a central case of satisfying
human activity illustrates that satisfaction obtains when we are engaged in active pursuit
of goals, rather than when we attain goals.
Using analyses of this form, Nietzsche argues that (B) is true.47 However, there is also
a shorter route to the conclusion that (B) is true: we can appeal to contemporary
empirical work on human psychology.48 For recent psychological research strongly
supports (B). There is persuasive evidence indicating that human beings are most
satisfied when engaged in activities that provide them with challenges that are neither
too easy nor beyond their capacities:
In all the activities people in our study reported engaging in, enjoyment comes at a very specific
point: whenever the opportunities for action perceived by the individual are equal to his or her
capabilities . . . [For example,] a piece of music that is too simple relative to one’s listening skills
will be boring, while music that is too complex will be frustrating. Enjoyment appears at the
47
For Nietzsche’s arguments in favor of (B), the following passages are especially relevant: D 60, Z Preface
3, Z II.2, Z III.5, Z IV.13, BGE 200, BGE 228, GM I.10, GM III.17, A 1–3.
48
A version of (B) is defended by Aristotle in Book X of the Nicomachean Ethics. There, Aristotle argues
that pleasure is something that “completes” an activity, something that occurs together with the performing
of activities. Thus, rather than being something that is intelligible as a separable sensation, something that
could occur apart from the performance of activities, pleasure is said to be inextricably bound to the
performance of activities.
boundary between boredom and anxiety, when the challenges are just balanced with the
person’s capacity to act. (Csikszentmihalyi 1990, 52)
Indeed, the most satisfying situations were found to be those in which “all of a person’s
relevant skills are needed to cope with the challenges of a situation” (Csikszentmihalyi
1990, 53).49
This point can be illustrated with a mundane example from everyday life: consider
the ubiquity of games. Take crossword puzzles. These puzzles present players with
challenges that, though surmountable, require a good deal of effort. An agent who
adopts the end of solving a crossword puzzle also seems to aim at encountering and
overcoming resistance in the pursuit of a solution. For example, a skilled crossword
puzzle player won’t derive much enjoyment from solving a puzzle aimed at beginners.
Part of the point of playing a game is that it is challenging, that it requires ingenuity and
skill, that it presents obstacles to be overcome.50 One could draw similar lessons from
examinations of other activities: athletes want to play teams that present challenges, not
to play groups of amateurs; scientists want to discover new truths, not to go over truths
already discovered.
Suppose we accept the arguments in favor of claims (A) and (B). These claims
indicate that the conditions of human satisfaction have a decidedly odd structure: we
desire to achieve certain ends, and are satisfied when pursuing these ends, but not when
the ends have been attained. Thus, as Oscar Wilde put it, “in this world there are only
two tragedies. One is not getting what one wants, and the other is getting it. The last is
much the worst; the last is a real tragedy!” (Wilde 1998, 37). This is paradoxical; why
should it be so? What is the point of pursuing an end, if achieving the end produces at
most a fleeting satisfaction? Why would we take satisfaction in this process?
Well, Nietzsche’s work provides an explanation: claim (B) indicates that we aim not
only to achieve ends, but also simply to be active. We aim to engage in the process of
encountering and overcoming resistances in the pursuit of ends. Activities such as
marathon running and puzzle solving are exemplary: they involve aiming to attain a
state that will provide at most a temporary satisfaction, a state that can seem valueless.
Who cares whether I have run twenty-six miles, or whether I have filled in the blanks
in a crossword puzzle? The point of these activities isn’t the end—it is the process, the
activity itself. And Nietzsche has a straightforward explanation for this otherwise
49
For introductions to the psychological work on satisfaction, see for example Bradburn (1969), Deci and
Ryan (1985) and Csikszentmihalyi (1988, 1990).
50
Suits (2005) argues that games have two essential features: (1) they have an identifiable goal (such as
checkmating one’s opponent or filling in the blanks in a crossword), and (2) they have rules forbidding the
easiest or most efficient way of achieving that goal (such as moving one’s pawn as many spaces forward as one likes,
or looking up the answers to a crossword). We might summarize this second point by saying that the essential
feature of games is that they involve deliberately imposing obstacles or resistances on oneself. Suits further
argues that the best life would be a life devoted solely to the playing of games. In other words, the best life
would be the life devoted solely to the deliberate impositions of obstacles on oneself, so that one can
overcome them. The connection with Nietzsche’s thought is striking.
puzzling feature of human agency: the explanation is simply that we are motivated by
drives. For recall that drives aim at their own expression, at manifesting their charac-
teristic form of activity, and take objects only as chance occasions for this expression.
The attainment of objects doesn’t satisfy the drive; only the performance of the drive’s
characteristic form of activity satisfies the drive. If our actions were drive-motivated,
(A) and (B) would follow. If our actions were not drive-motivated, (A) and (B) would
be puzzling. Therefore, Nietzsche argues, the best explanation for (A) and (B) is that
our actions are drive-motivated.
Of course, this falls short of a deductive proof that our actions are drive-motivated. It
just gives us some reason for believing that they are. As I mentioned above, Nietzsche
elsewhere offers other arguments, which bolster his case for the claim that our actions
are drive-motivated. But for now, I want to rest content with a modest claim: we have
some reason for believing that human actions are drive-motivated.
3.4 Summary of the argument
We now have the outline of the Nietzschean argument before us. In essence, it can be
summarized in just three claims. First, drives are motivational states that aim at their
own expression, and take various objects merely as chance occasions for expression.
Second, drive-motivated actions constitutively aim at encountering and overcoming
resistance. Third, all human actions are drive-motivated. It follows that all human
actions inescapably aim at encountering and overcoming resistance. Or, to translate
these claims back into Nietzsche’s terminology: all human action manifests will to
power. Power is a constitutive aim of action.
4. Objections and clarifications

The Nietzschean claim that action constitutively aims at encountering and overcoming
resistance is bound to be surprising and counterintuitive. Part of the counterintuitive
nature of Nietzsche’s claim rests on the fact that it is easily confused with some
superficially similar, but implausible, views. In this section, I will clarify and defend
the view.
4.1 Agents aim at determinate forms of resistance, rather than resistance in general
It is important to notice that the will to power thesis does not imply that we are
motivated to seek all forms of resistance. Rather, it implies that we are motivated to
seek the forms of resistance that give the motivating drives an opportunity for expression.
To bring this point out, consider the difference between the following two claims:
(A) Whenever we act, we aim to encounter and overcome resistance of any and all
kinds.
(B) Whenever we act, we aim to encounter and overcome resistances that are
related to the activity that we are performing.
O B J E C T I O N S A N D C L A R I F I C AT I O N S 177
According to claim (A), we seek resistance as such. If this claim were true, it would
have two decidedly odd implications. First, we would be motivated to perform a
number of activities that generate immense amounts of resistance, but seem pointless or
indeed even insane. For example, sticking one’s hand in a fire, or hacking off one’s
own limbs, would generate enormous amounts of resistance. Thus, if claim (A) were
true then every agent would be motivated to perform these actions. But that is
obviously false.
Second, according to (A) we have an aim that can be fulfilled by introducing any
kind of difficulty into an activity. For example, suppose an agent is engaged in the
process of writing an essay. The agent could fulfill aim (A) by interrupting her writing
with loud distracting music, or blindfolding herself, or in other ways introducing
difficulties. Again, it seems absurd to claim that agents aim at these kinds of resistances.
Claim (B) avoids these bizarre implications. First, few agents are engaged in the
activity of inflicting suffering on themselves, so few agents will be motivated to place
their hands in fire or chop off their limbs (of course, Nietzsche provides interesting
discussions of the way that some agents—ascetics—do seek these kinds of actions).
Second, aim (B) would not be fulfilled by introducing superfluous difficulties or
unrelated distractions into one’s acts. For these reasons, (B) is more plausible than (A).
Fortunately, Nietzsche endorses claim (B) rather than claim (A). The reason for this
should be clear from the prior section: claim (B) follows from Nietzsche’s conception
of drives, whereas claim (A) does not. Recall the argument for Nietzsche’s will to
power thesis: our actions are motivated by drives; drives aim at continuous expression
of their characteristic forms of activity; thus, drives motivate the agent continuously to
seek resistances upon which to vent their form of activity. The particular types of
resistance that the agent is motivated to seek will depend upon the drive in question.
For example, an agent who is in the grip of an aggressive drive will seek to engage in
aggressive activity, and so will seek those objects that afford opportunities for aggressive
activity. In this sense, the drive-motivated action of manifesting aggressive activity
seeks resistances. That is, the activity does not aim at putting itself to an end, but rather
aims at continuous expression, and therefore aims at the conditions of continuous
expression: encountering and overcoming obstacles or resistances.
It follows that the drive-motivated action seeks resistances of a particular form: those
that seem, to the agent, to present appropriate objects for expression of the form of
activity.51 Thus, the agent who is being motivated by an aggressive drive does not, as
(A) would suggest, seek just any kind of resistance. Rather, as (B) suggests, the agent
51
Section 3.1 mentioned that the objects that the drive seeks needn’t actually be appropriate targets for
the form of activity; if nothing in the agent’s environment warrants aggressive responses, the drive will
typically distort the agent’s perception of the environment, so that some object seems to the agent to be
worthy of aggression. So the agent will vent his activity upon objects that seem to him to be appropriate
targets of aggressive activity. For extended discussions of this point, see Katsafanas (forthcoming a, forthcom-
ing b).
seeks resistances that he views as appropriate objects upon which to continue manifest-
ing aggressive activity.52
There is also a second reason for accepting (B) and rejecting (A): interpretation (A)
would make a mess of the claim that will to power is a constitutive aim. (A) amounts to
an instrumentalist interpretation of power, treating it as a goal that we strive to fulfill.
But, as Section 2.2 discussed, will to power should not be interpreted in this manner.
Power is not a goal that we find ourselves with and strive to fulfill; rather, power is the
constitutive aim of action, which arises whenever we pursue any end at all and modifies
the way in which we pursue that end. Consider an analogy. Many philosophers would
agree that the instrumental principle, which tells us to take the means to our ends, is a
constitutive aim of action. However, the instrumental principle does not itself function
in an instrumentalist fashion: agents do not simply find themselves with a blank aim of
taking means to ends, and then go about looking for opportunities to fulfill this aim.
Rather, an agent decides to pursue some end, and pursuit of that end commits the
agent to acting in accordance with the instrumental principle.
Just so with Nietzsche’s claims about encountering and overcoming resistance. It’s
not quite right to say that I go about looking for resistances to overcome. Rather, I go
about performing drive-motivated activities, and it turns out that part of what it is to
perform these activities is to look for resistances to these activities. So agents do not
have a naked aim of encountering and overcoming just any form of resistance. Rather,
agents aim at performing activities of certain kinds, and in performing these activities
they aim at encountering and overcoming the forms of resistance that are related to the
activity in question.
In sum, Nietzsche’s theory looks extremely counterintuitive, and indeed patently
absurd, if we interpret him as saying that we all have one true aim—encountering and
overcoming resistance—that is equally well fulfilled by all types of resistances (claim
[A], above). For, on this interpretation, agents would simply go about looking to
maximize resistance of all kinds. This would generate incredible results: we would all
have reason to hack off our arms and so forth. Fortunately, Nietzsche does not accept
(A). Rather, he argues for claim (B): whenever we engage in a particular form of
activity, we aim to encounter and overcome resistances that are related to the activity.
4.2 Do all actions aim at resistance?
The previous section argued that agents do not pursue resistance as such, but rather
pursue the forms of resistance that are related to their end in acting. Thus, Nietzsche’s
view can avoid the absurd conclusion that we seek all forms of resistance. But there is
52
Reginster provides a helpful discussion of this point. He distinguishes intrinsic and extrinsic resistance.
Intrinsic resistance has two features: “First, it is pertinent insofar as it is created by the specific requirements of
the end one pursues. Second, it is essential, insofar as it is resistance anyone who engages in the pursuit of this
end would have to confront . . . ” (Reginster 2006, 179). Extrinsic resistance lacks one or both of these
features. Reginster argues that will to power aims at intrinsic, rather than extrinsic, resistance.
O B J E C T I O N S A N D C L A R I F I C AT I O N S 179
another problem that the view must face: we sometimes have ends that could be
pursued in ways that generate resistance, and yet we decline to pursue this resistance.
For example, consider a perfectly ordinary action: moving a pen across a page, in
order to produce a written sentence. It would be absurd to claim that agents aim at
encountering resistance in the moving of their pens. So we have an apparent problem:
if moving the pen is an action, then there are actions that do not aim at encountering
and overcoming resistance.
I have deliberately chosen an action that suggests an obvious solution: moving a pen
in order to write a sentence is typically a part of or means to some larger action, such as
writing a paper. Only a very odd agent would aim at encountering and overcoming
resistance in writing a sentence, but many agents do aim at encountering and over-
coming resistances—such as intellectual challenges—in writing papers. So Nietzsche
could preserve his claim that all actions aim at encountering and overcoming resistance
by treating writing a sentence as a part of the larger action, writing a paper. The action as a
whole does have the aim of encountering and overcoming resistance. So we were
simply looking for resistance in the wrong place: the agent isn’t aiming at encountering
and overcoming resistance to his pen movements, but he is aiming at encountering and
overcoming intellectual challenges.
In general, when some action appears not to manifest the constitutive aim, we must
be careful to determine whether a more expansive view of the action reveals that it
does, in fact, manifest the aim. This point is independent of the particular constitutivist
theory that we embrace. Take a much simpler example: an agent who is engaged in a
game of chess. Earlier, I argued that chess is governed by the constitutive aim of
checkmating one’s opponent. However, if we consider particular moves in isolation, it
may not be obvious that they are governed by this aim. For example, suppose a player
puts his queen into a position where it can be taken by his opponent’s pawn. This
move, considered in isolation, would not seem to be regulated by the aim of check-
mate; indeed, it would seem to make checkmate far less likely. However, if this move is
simply one step in a series of moves, it could of course be part of some strategy that does
aim at checkmate.
In fact, this point is not even unique to constitutive aims; it applies to all aims.
Consider a person who is walking down the street with the aim of getting to his office.
To do so, the person must perform a number of simpler actions, such as putting one
foot in front of the other. If we isolated one of these foot movements, we would not be
able to discover that it was regulated with the aim of getting to the office. But, taking a
broader view of the whole series of movements, the aim reveals itself.
I suggest that Nietzsche can block purported counterexamples to the will to power
thesis in an analogous way. Confronted with some action that appears not to aim at
power, Nietzsche can treat it as a part of a larger action that does aim at power.
Of course, if this strategy is to be convincing, then Nietzsche will need a principled
way of determining when one action counts as a part of some larger action. In the
absence of such an account, the strategy would seem ad hoc. Although Nietzsche never
explicitly addresses this topic, I think his texts reveal an implicit commitment to a
straightforward view: an action, A-ing, is a part of a larger action, B-ing, iff A-ing and
B-ing have a common causal history. For example, writing a sentence is a part of writing a
paragraph because these actions are both motivated by the same desire: to write a paper.
Or, moving my leg is a part of crossing the street because moving my leg is regulated by the
same aim as crossing the street.
I believe Nietzsche relies on this view for two reasons. First, as I argue in Katsafanas
(forthcoming a), Nietzsche claims that desires are products of drives. So, for any action
that appears to be merely desire-motivated, Nietzsche believes that we can find some
drive that produces the desire. For example, if I find myself with a desire to read a book,
this may be the product of an intellectual drive. This implies that many apparently
discrete actions are actually connected: they are done in the service of a common aim.
Second, recall that drives admit an aim/object distinction: the same drive can
manifest its characteristic form of activity on quite different objects. So, for example,
the aggressive drive might vent itself in athletic endeavors, in the making of humiliating
remarks, in overtly aggressive behavior, and so forth. These seemingly disparate
behaviors are unified by a common causal source. Again, we have reason for classifying
apparently discrete actions as parts of one broader action. By examining the causal
history of actions, we can find the “long logic” in a person’s actions (KSA 11:34[96]).
Nietzsche’s investigations of these deep causal histories give him reason to individuate
actions in terms of their common causal antecedents.
In sum, apparent counterexamples to the will to power thesis can be dealt with in
the following way. In order to generate a counterexample to will to power, we would
need to find a discrete action that does not aim at resistance. However, if Nietzsche is
correct in arguing that all desires are products of drives, then no such actions exist. For if
the causal history of this action includes a drive, then it would not be a discrete action;
it would be part of a larger action that did aim at resistance.
4.3 Do all actions aim at resistance? Further apparent counterexamples
In the prior section, I argued that many actions that appear not to manifest the aim of
power do, in fact, manifest that aim, when they are viewed as parts of larger actions.
However, not all actions will admit of this treatment. When I have a casual chat with a
friend, or watch a sitcom on the television, it seems implausible to regard these actions
as parts of larger actions that aim at encountering and overcoming resistance. On the
contrary, they seem to be discrete actions that do not aim at encountering and
overcoming resistance. So again, we seem to have a counterexample to Nietzsche’s
view. However, Nietzsche does have a response to these sorts of examples, which I will
explain below.
In order to determine whether an action aims at resistance, we must first determine
what kinds of resistance the action-type in question affords. Section 4.1 pointed out
that Nietzsche rejects the claim (A) that we seek every kind of resistance, and instead
maintains (B) that when we are performing a type of activity, we seek resistances that
CONCLUSION 181
are related to that activity. This has an interesting result. Suppose an agent performs a
particular type of action that does not afford many opportunities for overcoming
resistances, such as having a casual conversation with a friend. Precisely because there
are few related forms of resistance, the agent may already be encountering all the
resistance that the action-type affords. The agent is not seeking more resistance
precisely because there is no more resistance to seek.
To be sure, the agent could try to change the activity so that it would present such
opportunities; rather than a causal conversation, the agent might attempt to start an
argument or a debate. But these would be different activities. Insofar as the agent
merely seeks to have a casual conversation, the agent will have few (if any) opportun-
ities for overcoming resistance. If we accepted (A), the action would be a counter-
example to the will to power thesis. But if we accept (B), there is nothing wrong with
the conversation.53
Consider another potential counterexample: when I loaf on the couch and watch a
lowbrow sitcom on television, it may seem that I am not encountering and overcom-
ing any resistance. However, there are resistances here, albeit of the most minimal sort:
one must attend to the program, one must support oneself on the couch, one must
resist competing desires that incline one to perform other actions, and so on (after all,
loafing is marginally more demanding than non-action events such as sleeping). While
different types of activities generate different degrees of resistance (marathoning is far
more difficult than watching television), every activity generates at least some modicum
of resistance. After all, acting is shaping a recalcitrant world: part of what it is to act is to
effect a change in the world, and effecting a change in the world requires overcoming
resistance. (For remarks to this effect, see GM II.12, GM II.18, and KSA 13:11[111].)
5. Conclusion
Interpreting power as the constitutive aim of action resolves a longstanding interpretive
puzzle in Nietzsche scholarship, by enabling us to see how Nietzsche can consistently
53
This conclusion might seem problematic, for in a typical case the person’s activity will encompass many
different action descriptions, and these different descriptions may generate different claims about what is
“related” to the person’s activity. For example, in a given case the descriptions “having a casual conversation
with a friend” and “talking about philosophy” might both be appropriate. Yet, depending on which
description we choose, we might be able to discern different forms of resistance. Thus, one might object
that in the case I have described there are resistances that the agent could be, but is not, pursuing. In response
to this objection, Nietzsche could maintain that some of the action descriptions are more apt or more
revealing than others. There are, after all, important differences between aiming primarily to have a casual
conversation, aiming primarily to talk about philosophy, and aiming jointly at both of these goals. When we
examine a particular action, it will not be easy to determine which of these descriptions is most appropriate,
and in any case this may change over time, as the action progresses. However, if we can settle on an
appropriate description at a given time, the claims made above in the text will hold. (This might be even
clearer if we consider a simpler and more isolated action, such as the action of watching a sitcom on
television.) Accordingly, I take this to be an epistemological point about the difficulties inherent in describing
actions, rather than an objection that undermines the will to power thesis. (Thanks to Richard Moran for
raising this objection.)
maintain the following three claims: (1) power has a privileged normative status, (2)
there are no objective values, and (3) values are created. According to the interpretation
that I have advanced, power is not an objective value, in the sense that it would not
have value independently of a particular feature of human activities. Rather, we are
committed to valuing power merely in virtue of acting, because power is the consti-
tutive aim of action. Thus, power has a privileged normative status. Moreover, there is
a sense in which the value of power is created by human activity: the structure of our
own actions commits us to valuing power. Nietzsche’s three claims about power and
value are therefore consistent.
I have argued that once the structure of Nietzsche’s will to power theory is rendered
explicit, the theory becomes compelling. The theory is grounded in facts about the
nature of human motivation, facts that are not only philosophically plausible, but are
also supported by empirical research on human agency. Thus, Nietzschean consider-
ations provide us with a successful argument for the claim that action constitutively
aims at power.54
However, we have yet to investigate the particular normative consequences gener-
ated by this claim. The next two chapters take up this task. Chapter 7 looks more
closely at the structure of Nietzschean constitutivism, explaining how we use the
theory to generate normative conclusions. Chapter 8 asks which particular normative
conclusions are entailed by the theory.
54
The Appendix considers objections to my interpretation of Nietzsche. Readers with interpretive
concerns may wish to turn there before proceeding to Chapter 7.
7
The Structure of Nietzschean
Constitutivism
The previous chapter defended a Nietzschean version of constitutivism, according to

which will to power is a constitutive aim of action. To say that agents will power is to
say that they aim to encounter and overcome resistance in the course of pursuing other,
more determinate ends. Nietzsche’s drive psychology establishes that all actions have
this aim. If we accept Nietzsche’s claims that all actions are drive-motivated and that
drive-motivated actions aim at encountering and overcoming resistance, it follows that
action has a constitutive aim of encountering and overcoming resistance. However, we
have not yet examined the way in which the Nietzschean will to power thesis
generates an ethical theory. How, exactly, does the constitutive aim of encountering
and overcoming resistance generate normative results? That is the topic of this chapter.
Section 1 begins by discussing the way in which the constitutive aim generates norm-
ative conclusions about what we have reason to do and what we have reason to value.
Section 2 continues this discussion, examining in more depth the relationship between
will to power and other values. Section 3 asks whether the constitutive aim of power is
differentially realizable. Section 4 introduces a complication. As Chapter 5 argued that
agential activity is a constitutive aim of action, the Nietzschean theory that I defend is
bipartite: it posits two constitutive aims, agential activity and power. Section 3 explores
the way in which these aims interact.1 There, I argue that the bipartite structure of this
Nietzschean constitutivism enables it to overcome both an objection inspired by
Schopenhauer and the “why bother?” problem discussed in Chapter 2.
1. Introduction to the normative consequences

of the Nietzschean theory
The previous chapter introduced Nietzsche’s will to power thesis and defended it from
various counterexamples. The next task is to examine the way in which the view
1
Analogously, Korsgaard sometimes describes her theory as positing two constitutive standards. She writes,
“the hypothetical and categorical imperatives are constitutive principles of volition and action” (Korsgaard 2009,
81). Likewise, she claims that “a good action is one that constitutes its agent as an autonomous and efficacious
cause of her own movements” (2009, xii). In other words, autonomy (acting on the categorical imperative) and
efficacy (acting on the hypothetical imperative) are jointly constitutive of action.
184 T H E S T RU C T U R E O F N I E T Z S C H E A N C O N S T I T U T I V I S M
generates normative and evaluative consequences. From the claim that action consti-
tutively aims at power, we can derive two kinds of consequences: claims about what
we have reason to do and about what we have reason to value. I address these points in
turn.
1.1 How the account generates claims about what we have reason to do
Suppose Nietzsche is correct that all actions aim at power. Is there a way of moving
from this claim about our aims to a normative claim about what we have reason to do?
There is, provided that we grant the assumption that aims engender standards of success.
More precisely: if you have an aim, you have a (pro tanto) reason to fulfill it.2 As
I pointed out in Chapters 1 and 2, this claim, which I called “Success,” is relatively
uncontroversial; even the most minimal accounts of practical reason, including most
variants of the Humean account, accept it.3
If actions aim at encountering and overcoming resistance, and if aims are reason-
providing, then it follows that we have reason to seek those actions that afford resist-
ance. For example, if I aim to write a philosophical essay, I have reason to engage in a
challenging version of that task, by choosing a topic that presents me with difficulties.
(In a moment I will introduce a complication: the degree of resistance encountered and
overcome is not the only standard of success.)
In order to apply this prescription to a determinate case, we must take account of
two facts. First, which actions an agent has reason to perform depends on facts about
her capacities and psychological makeup. Most straightforwardly, an individual has
reason to pursue those activities that afford resistances that are difficult to overcome,
but not so difficult as to be impossible for her to overcome. After all, part of one’s aim is
to overcome the resistance that one encounters. Thus, if Sally is a brilliant novelist, and
Bill can barely string together a coherent sentence, then Sally has reason to pursue the
task of writing a great novel, whereas Bill does not.
Second, the results depend on facts about one’s environment. The particular courses
of action that constitute the highest realizations of will to power differ according to the
circumstances in which an agent finds himself. Nietzsche notes that during some
historical epochs, the opportunities for the highest expressions of will to power involve
distasteful events. In ancient societies, Nietzsche likes to remind us, the greatest
expressions of will to power occurred in acts of physical violence, subjugations of
one’s enemies, and the like. It is important not to sanitize Nietzsche: as he was well
aware, it is an implication of his will to power thesis that, given certain pre-societal
2
A pro tanto reason is a reason that has some weight, but nonetheless may be outweighed by other
reasons. For example, if I aim to get to my office within ten minutes, and if doing so requires driving at ninety
miles per hour, I have a pro tanto reason to drive at this speed. Nevertheless, this reason is outweighed by
reasons provided by my other aims, such as my aims of driving safely and avoiding potential harm to others.
3
See Robertson (2011) for a persuasive argument that Nietzsche relies on a related principle.
I N T RO D U C T I O N T O T H E N O R M AT I V E C O N S E Q U E N C E S 185
conditions, the flourishing human being will be the one who engages in acts of
oppression and subjugation that would today be regarded as horrific.4
However, Nietzsche’s writings do not urge contemporary individuals to engage in
these kinds of activities. Rather, Nietzsche enjoins us to undertake activities such as
artistic creation, the pursuit of self-understanding, and the willingness to subject oneself
to suffering in order to achieve goals.5 For this reason, the figures whom Nietzsche most
often presents as ethical exemplars are Goethe, Beethoven, and—rather immodestly—
Nietzsche himself (see, for example, TI IX.49, BGE 245, and EH IV). So Nietzsche
seems to believe that in our historical circumstances, his theory recommends intellec-
tual, artistic, and creative endeavors, rather than brute assertion of physical might or
dominance.6
To see why Nietzsche presents his theory as entailing these results, we must turn to a
more general point. Consider the claim that we have reason to seek resistances and
obstacles. An objection will have occurred to many readers: if this were the only
constraint on our actions, then murder, oppression, torture, and the like would be
exemplary actions. After all, those actions typically engender a great deal of resistance
and a large number of obstacles. Does the Nietzschean theory entail these results?
Upon closer inspection, then, will to power doesn’t seem to be a good candidate for
the only normative standard that informs our deliberations. This point must be put
carefully: the problem is not that if will to power were the sole normative standard, it
would generate results that conflict with our current moral intuitions. As Chapter 1
indicated, our intuitions have no authoritative status. The real problem is that if will to
power were the only normative standard, then we would all have reason to perform
have equal reason to perform actions that generate equal resistance. This seems not
only implausible, but also counter to the results that Nietzsche claims for his theory.
For Nietzsche certainly doesn’t recommend those actions; he doesn’t think we select
actions merely in terms of how much resistance they provide. Rather, the degree of
resistance afforded by an action seems to be one important factor among many others.
In order to understand this point, we will need to examine a second aspect of the
Nietzschean view.
1.2 How the account generates claims about what we have reason to value
Values pervade our world. Nietzsche claims that our experience of the world is value-
laden: “There is no doubt that all sense perceptions are wholly permeated with value-
4
See, for example, GM II, where Nietzsche claims that ancient individuals had strong desires to vent
cruelty on others. Given these desires, the will to power will incline these agents toward horrific acts. But
note also that we do not have these desires; as Nietzsche likes to put it, the aggressive instincts of modern
individuals have been “tamed” and redirected.
5
For the pursuit of self-understanding, see GS 11, 109–10, 335; A 54; EH IV.1, IV.3. For calls to artistic
creation, see HH II.174; GS 78, 107, 299; GM III.25; TI IX.8–9. For the connections between suffering and
greatness, see D 146; GS 13; Z Preface 3, II.3; BGE 39, 202, 225, 260; A 2, 7.
6
For helpful discussions of this point, see Richardson (1996, Chapter 3).
judgments” (KSA 12:2[95]; cf. GS 114, D 119). So, too, deliberation involves the
deployment of values; an agent who is deliberating about what to do will make her
decision, in part, by considering her own values. Nietzsche clearly does not think that
power is the only value influencing perception and deliberation: “the extent of moral
evaluations: they play a part in almost every sense impression. Our world is colored by
them” (KSA 10:24[15]; cf. GS 7, 139, 152, 301; BGE 186). Nor does Nietzsche think
that it should be: on the contrary, he enjoins us to create new values.7
Power is one value, yet is not the only value. This raises the question of what
relationship these values have with one another. Interestingly, as we noted in
Chapter 6, Nietzsche frequently presents will to power as a “principle of revaluation.”
He argues that will to power generates a standard according to which we can assess not
only actions, but also values.8 Thus, I take it that Nietzsche operates with the following
picture: facts about which actions we have reason to perform are determined not only
by the degree of resistance that the potential actions afford, but also by facts about the
relationship between potential actions and our values; however, these values must in
some way be vetted by the standard of will to power. In the following sections, I will
elaborate upon and defend this reading.
Let’s begin by asking what a value is. For Nietzsche, values arise through valuings:
“only through valuing is there value” (Z I.15).9 In other words, to say that X is a value
is just to say that we value X. Valuing X involves having a certain kind of affective
response to X and making certain kinds of judgments about X. The particular kinds of
affective responses and judgments will depend on the type of object or state of affairs
that is being valued. Take compassion, for example: valuing compassion involves
judging that one has reason to act compassionately, praising compassionate behavior,
striving to be compassionate, having feelings of guilt when one fails to be compassion-
ate, and so forth. Thus, when Nietzsche speaks of compassion as one of our values, he
simply means that we make certain kinds of judgments about compassion (e.g., that
7
For example, Nietzsche seeks “new philosophers” who are “strong and original enough to provide the
stimuli for opposite valuations and to revalue and invert ‘eternal values’ ” (BGE 203). He explains that these
new philosophers will “create values” (BGE 211). See also D 453, GS 301, GS 335, GS 382, BGE 211–12,
and BGE 260. Of course, creating values involves more than simply declaring “X is valuable.” For a few
examples of what is involved, see HC’s discussion of the Greek valuation of the contest; GM I’s discussion of
the ancient nobility’s valuation of beauty, struggle, and power; GM III’s discussion of the modern valuation
of truth; and A’s discussion of Jesus’ valuation of love. In all of these cases, creating a value involves fostering
attitudes of reverence toward the thing in question, organizing social institutions and cultural practices so as to
promote this reverence, and so on.
8
I discussed some of the relevant passages in Chapter 6, Section 1. Recall, for example, Antichrist 6: “Life
itself is to my mind the instinct for growth, for continuance [Dauer], for accumulation of force [Häufung von
Kräften], for power; where the will to power is lacking there is decline. It is my contention that all the supreme
values of mankind lack this will—that the values which are symptomatic of decline, nihilistic values, are lording
it under the holiest names.” See also BGE 2, 210–11; GM P 3, P 6, I.17n, III.27; TI IX.35, IX.38; A 2, 6, 9;
EH IV.1, IV.4; KSA 12:2[131], 12:2[190]. There are inklings of this view as early as Daybreak; see D 146, 262,
271, 360, 507.
9
Some relevant passages are D 3, 210, 453; GS 143, 296, 301, 335, 345, 353, 380–2; Z I.15; BGE 62, 212,
260; GM (especially essay I). Richardson emphasizes this point in Chapter 2 of his 2004.
there is reason to be compassionate), that we have certain kinds of affects directed at

compassionate behavior, and so on.10
These spare remarks about value are sufficient for our purposes. The only point that
will matter is this: an agent’s values constrain and influence her behavior, in part by
influencing her judgments about reasons.11 Accordingly, the agent’s values will interact
with will to power. The will to power generates claims about which acts there is reason
to perform; so, too, do the agent’s values. Given that claims about reasons arise from
these disparate sources, there is a potential for conflict.12
For example, consider someone who values a form of complacency. This individual
believes that it is valuable to be content with what one already has: one should not seek
further accomplishments. This value clearly conflicts with will to power. As the prior
chapter explained, will to power commits us to aiming at resistances and challenges. So
we have a straightforward conflict: valuing complacency involves judging that there is
reason not to confront challenges; valuing power involves judging that there is reason
to confront challenges. If an agent harbors both of these values, he will be in a state of
conflict, endorsing contradictory propositions about how to act.
What does this tell us about the value of complacency? It is clear enough that, in
presenting will to power as the standard of revaluation, Nietzsche wants us to reject any
value that generates conflicts of this form. I take it that this is what Nietzsche has in
mind when he writes that the “standard by which the value of moral evaluation is to be
determined” is “will to power” (KSA 12:2[131]). In this passage, and a number of
other sections, Nietzsche suggests that any value that conflicts with will to power
should be rejected.13
However, as we saw in Chapter 6, this approach does raise a question. Suppose it is
true that the reasons generated by valuing complacency conflict with the reasons
arising from will to power. We could eliminate this conflict in two ways: we could
cease to value either complacency or power. So why does Nietzsche enjoin us to reject
complacency rather than power?
10
Poellner (2007) and Clark and Dudrick (2009) provide nuanced analyses of the particular forms of
affects and types of judgments that constitute valuings. In Katsafanas (forthcoming b), I argue that Nietzsche
interprets values as drive-induced affective orientations of which the agent does not disapprove. For present
purposes, though, we can work with an ordinary understanding of values; the details will not be relevant.
11
This point is brought out in a recent discussion from T. M. Scanlon: “To value something is to take
oneself to have reasons for holding certain positive attitudes toward it and for acting in certain ways with
regard to it. Exactly what these reasons are, and what actions and attitudes they support, will be different in
different cases” (2000, 95). As Scanlon points out, an agent’s values influence the reasons that the agent takes
herself to have. (Scanlon himself argues that we can identify the agent’s values with facts about what the agent
takes herself to have reason to do.)
12
In Chapter 5 I argued that motives strongly and pervasively influence reflective judgments about
reasons, but need not determine these judgments. Accordingly, the above forms of conflict can arise when
the agent’s reflective judgments are out of accord with the tendencies induced by her affects. Nietzsche
recognizes this point, encouraging us to examine cases in which this discrepancy arises. For a few examples,
see D 18, D 103, GS 335, and GM Preface.
13
For additional examples, see the passages discussed above in Chapter 6, Section 1.
Here, the Nietzschean version of constitutivism can rely on the inescapability of the
constitutive aim. The values arising from our agential nature cannot be reassessed or
altered; as the previous chapter argued, we are inescapably committed to aiming at
power, and we therefore have inescapable reasons for performing those actions that
involve manifestations of power. In that sense, we are committed to valuing power.
However, Nietzsche maintains that other values can be reassessed and altered: there are
many different, mutually incompatible sets of values that we could embrace. For
example, we could—and many do—regard complacency as not valuable, or even as
disvaluable. So the Nietzschean point is simple: when there is a conflict between the
constitutive aim and some other value, the only way in which we can alleviate the
conflict is by modifying the other value.14
We are now in a position to see an important difference between the Nietzschean
theory and more familiar ethical views. The Nietzschean theory does not hold that in
order for a value to be justified or legitimate, it must be derived from the will to power.
To elucidate this point, it will be helpful to contrast the Nietzschean view with two of
the most prominent ethical views, Kantianism and Utilitarianism. These theories have
a foundationalist structure: they are committed to the idea that there is one source of
ethical norms. Kantians claim that we can extract all ethical content from one principle,
the Categorical Imperative. Likewise, Utilitarians believe that we can derive all ethical
content from a principle enjoining us to maximize aggregate utility. On these views,
normative claims are legitimate only insofar as they follow from the respective foun-
dational principle. For example, consider a normative claim such as “lying is wrong.”
On a Kantian view, this claim will be true if and only if it follows from the Categorical
Imperative; on a Utilitarian view, it will be true only if acceptance of it maximizes
aggregate utility.15
If the Nietzschean view were intended to function in an analogous fashion, then all
legitimate normative claims would be derived from facts about the will to power.
However, this approach would give us a set of norms so severely attenuated as to defy
description as an ethic. For it should be clear enough that we are not going to be able to
derive claims such as “lying is wrong” and “murder is wrong” from facts about will to
power; there is no way of moving from the idea that we aim to encounter and
overcome resistance to the idea that we should not lie, or that we should not murder.
On the contrary, lying and murdering are ways—possibly quite good ways—of willing
power.
14
I do not mean to suggest that changing one’s values is always easy. Changing one’s values is typically a
gradual, aggregative process, and can involve great struggles. Consider the hackneyed example of a formerly
religious individual, who used to believe that sex was disvaluable, but now regards it as valuable. We can
easily imagine that in order genuinely to rid himself of the old evaluation, he will have to struggle with
residual feelings of guilt, shame, and so forth.
15
I am being deliberately vague here, as the details will vary depending upon the particular version of
utilitarianism that we embrace.
Fortunately, we need not interpret Nietzsche’s will to power doctrine as a founda-

tional principle from which we derive all other normative claims. Rather, as this
section has explained, will to power is intended to serve as a “principle of revaluation.”
That is, the will to power generates a standard in terms of which we are to assess all
other values. So the Nietzschean view grounds one normative principle in facts about
our agential nature, and uses this principle not to derive, but to assess, the other values that
we embrace.16
In sum, the Nietzschean theory holds that values are legitimate insofar as they do not
generate conflicts with will to power.17 This allows us to see why the Nietzschean view
does not simply generate a crude injunction to maximize resistance. The degree of
resistance that an action affords is one standard of assessment, but not the only standard.
In addition to evaluating potential actions with regard to the degree of resistance they
afford, we should evaluate them with other values that we embrace.
This explains why the Nietzschean theory does not entail that we all have reason to
perform every activity that generates great resistance, such as murdering, sticking our
hands in fires, and so forth. These pursuits would conflict with the other values that
most of us embrace. For example, while murdering would engender resistances to
overcome, it would conflict with the value that most individuals place on human life.18
While sticking one’s hand in the fire would generate great resistance, it would conflict
with the value that we tend to place on self-preservation. On the other hand, activities
such as the pursuit of intellectual endeavors would not conflict with, and in some cases
will be supported by, our other values. Thus, individuals tend to have more reason to
perform these activities.
16
I take it that this is what Nietzsche has in mind when he claims that the will to power is a “principle of
revaluation” for other values: it is not a source from which we derive other values, but a standard by means of
which we assess them. However, Nietzsche is not completely explicit on this point, and establishing that
the texts support my interpretation of Nietzsche’s will to power thesis would take us too far afield. For a
defense of my interpretation of Nietzsche on this point, see my “The Problem of Normative Authority in
Kant, Hegel, and Nietzsche.”
17
If we are committed to willing power, shouldn’t we welcome the resistance that obtains when we
embrace values that conflict with will to power? That is, wouldn’t embracing conflicting values be one way
of willing power? If so, this would undermine Nietzsche’s view. However, I think we can see that the answer
is no. First, Nietzsche does not claim that we have reason to seek all forms of resistance; rather, we have reason
to seek the forms of resistance that are related to the activities we are pursuing. Second, when I say that values
conflict with will to power, I do not mean that they engender resistances. Rather, I mean that they generate
contradictory claims about what there is reason to do. This isn’t a conflict that one can overcome or
eliminate, anymore than one could overcome or eliminate the contradictory claims generated by believing
both that 1 + 1 = 2 and that 1 + 1 = 3.
18
Of course, Nietzsche wants us to scrutinize and reassess all of our values, including our disvaluation of
murder. These valuations can’t be held stable; but I am here assuming that they will withstand critical
scrutiny. Consider Nietzsche’s remark: “It goes without saying that I do not deny—unless I am a fool—that
many actions called immoral ought to be avoided and resisted, or that many actions called moral ought to be
done and encouraged—but I think the one should be encouraged and the other avoided for other reasons than
hitherto” (D 103).
1.3 Summary of the normative structure of the Nietzschean theory

The previous sections have argued that the will to power does not serve as the only
legitimate value. Although will to power has a privileged normative status, it is one
value among many others. Thus, the normative implications of the Nietzschean view
depend upon what other values one embraces. An act such as committing a murder is
typically quite challenging, and in that respect performing the act would fulfill the
agent’s aim of encountering and overcoming resistance. But, of course, the fact that
murder is one way of fulfilling the will to power does not imply that we should
murder.19 After all, we accept the following normative claims: murder is wrong,
human life is valuable, gratuitous infliction of suffering upon others is wrong, etc.
These normative claims entail that murder is disvaluable. So, on the Nietzschean view,
we would have to ask whether there are other acts that produce just as much resistance
as murder, but do not violate our evaluative commitments. And, as Nietzsche makes
clear, there are. Nietzsche’s favorite examples of such acts are, as we have seen, artistic
and intellectual endeavors. Thus, in one of his earliest works, Nietzsche writes that only
“the philosophers, artists, and saints” are “truly human beings and no-longer-animals”
(UM II.5); in one of his final works, he tells us that “the most spiritual men” are “the
strongest ones” (A 57). Nietzsche emphasizes that these “most spiritual men” are
engaged in difficult personal projects and challenges, rather than overt aggression:
The most spiritual men, as the strongest, find their happiness where others would find their
destruction: in the labyrinth, in hardness against themselves and others, in experiments. Their joy
is self-conquest: asceticism becomes in them nature, need, and instinct. Difficult tasks are a
privilege to them; to play with burdens that crush others, a recreation. Knowledge—a form of
asceticism. They are the most venerable kind of man: that does not preclude their being the most
cheerful and the kindliest. (A 57)
So Nietzsche presents his exemplars as engaged in intellectual rather than aggressive

activities.20
In sum, then, the Nietzschean theory requires that we assess potential acts along two
dimensions:
(1) The extent to which the act would present opportunities for encountering and
overcoming resistance.
(2) Whether the act is permitted, recommended, or forbidden by the other values
that we embrace.
19
It is worth noting that Nietzsche sometimes makes this point completely explicit. For example, consider
the passage cited in the prior note. On Nietzsche’s view, we have misunderstood the basis of ethical claims;
we thought they had deeper foundations than they actually do. But this does not entail that, in every case in
which traditional morality tells us not to A, Nietzsche thinks that A-ing is permissible.
20
So far, I have only noted this point, rather than explained it. Chapter 8 will explain why these individuals
are Nietzsche’s exemplars.
U S I N G W I L L T O P OW E R T O A S S E S S VA L U E S 191
Moreover, as Section 1.2 pointed out, we must also assess our other values in terms of
power. So we need a third dimension:
(3) The extent to which the values cited in (2) are compatible with the will to
power.
Only by answering these three questions can we generate normative results.
So the Nietzschean ethical theory functions in the following way. First, an agent
undertakes a revaluation of values, assessing her values in light of power. She sheds
some values and embraces others. She then uses this new set of values, together with
the valuation of power, to determine what she should do.21
As a result, Nietzsche’s ethic is always historically situated: it tells us how to go on, not
how to begin. In other words, Nietzschean revaluation always begins in some deter-
minate cultural setting. An agent starts with an array of accepted values. The agent
diagnoses these values in terms of will to power, abandoning some and preserving
others. Where the agent ends up will depend, in part, on where she began.22
2. Using will to power to assess values

We now have the outline of Nietzschean constitutivism before us. But some aspects of
the theory stand in need of clarification. In this section, I address three points. In
Section 2.1, I explain what, exactly, it is for a value to conflict with will to power.
Section 2.2 addresses a complication, arguing that the reasons generated by the consti-
tutive aim are pro tanto reasons rather than all-things-considered reasons. Section 2.3
reflects on why the Nietzschean theory treats will to power as employed in conjunc-
tion with other values, rather than in independence from other values.
2.1 What is it for a value to conflict with will to power?
In presenting the structure of Nietzschean constitutivism, I have frequently claimed
that various values “conflict with” will to power. This raises two questions. First, what
21
Notice that this interpretation of Nietzsche’s ethical theory is consistent with the model of agency
offered in Chapter 5. There, I argued that motives operate through, rather than independently of, reflection.
In short, the claim that I labeled “Suspension” is false: deliberation does not suspend the effects of our
motivational states. It follows that reflective assessment of motives and values will be influenced by these
motives; there will be no case in which the agent steps back from and achieves independence from the effects
of her motives. Nonetheless, I argued that Inclination is true: motives do not determine choice. While the
reflective assessment of values will be influenced by motives, these motives will not determine the outcome
of the reflection. Accordingly, there is space for the kind of critique that Nietzsche envisions.
22
This is clear from the structure of Nietzsche’s texts. He does not propose new values out of the blue; his
texts always begin with a critical investigation of extant values. Thus, Nietzsche writes that philosophers’ task
is to apply “the knife vivisectionally to the chest of the very virtues of their time,” revealing the hypocrisy,
contradictions, hidden motives, and defunct ideals at the heart of their society’s way of life (BGE 212).
Nietzschean ethical inquiry takes the form of an investigation of our cultural practices, expectations, and
institutions, bringing to light their implicit principles, motives, and ideals.
exactly is it for a value to conflict with will to power? Second, what kinds of conflicts
are possible?
When I say that a value conflicts with will to power, I mean that having this value
entails taking there to be reason to perform certain actions which will to power entails
that there is reason not to perform. For example, valuing complacency entails taking
there to be reason to avoid resistance, whereas the will to power thesis entails that there
is reason not to avoid resistance (and that, on the contrary, there is reason to seek it). In
this sense, valuing complacency conflicts with will to power. In order to have a
shorthand way of referring to this kind of conflict, I will simply say that value in
question conflicts with will to power.
One of the central themes in Nietzsche’s work is the idea that Judeo-Christian values
come into pervasive, systematic conflict with the will to power, in a way that the values
of classical societies did not. For example, Nietzsche argues that the value that the
ancient Greeks placed upon agonal struggle was exemplary, in that it induced individ-
uals to take upon themselves great challenges; whereas the value that the Judeo-
Christian system places upon modesty and humility is paradigmatically in conflict
with will to power, for it induces individuals not to struggle to distinguish themselves.23
Consider some of the predominant values within our culture—values such as
compassion, charity, equality, pleasure, the absence of suffering, altruism, peacefulness,
and the like. Nietzsche argues that valuing these things conflicts with will to power. In
the Preface to the Genealogy, Nietzsche asks,
What if a symptom of regression were inherent in the ‘good’, likewise a danger, a seduction, a
poison, a narcotic, through which the present was possibly living at the expense of the future? . . . So
that precisely morality would be to blame if the highest power and splendor actually possible to the
type man were never in fact attained? So that precisely morality were the danger of dangers? (GM
Preface 6)
Nietzsche’s answer to those questions is clear enough: conforming to these values does,
indeed, conflict with the will to power. Nietzsche complains that the effects of “liberal
institutions” are “known well enough: they undermine the will to power” (TI IX.38).
He writes, “well-being as you understand it—that is no goal, that seems to us an end, a
state that soon makes man ridiculous and contemptible” (BGE 225). And he warns that
“our weak, unmanly social concepts of good and evil and their tremendous ascendancy
over body and soul have finally weakened all bodies and souls and snapped the self-
reliant, independent, unprejudiced men, the pillars of a strong civilization” (D 163; cf.
BGE 62, A 5).
Nietzsche’s complaint can be put simply: we believe that equality, altruism, pleasure,
and compassion are valuable. But, upon investigating these values, we find that these
values conflict with the will to power. As the will to power has a privileged evaluative
status, we should abandon our commitment to the values that conflict with it. Thus,
23
For discussions of this point, see HC, TI II, and A.
many of our traditional values must be jettisoned. Nietzsche puts the point in charac-
teristically apocalyptic terms: as we uncover the true nature of these traditional values,
“morality will gradually perish now: this is the great spectacle in a hundred acts reserved
for the next two centuries—the most terrible, most questionable, and perhaps also the
most hopeful of all spectacles” (GM III.27).
So we can see that Nietzsche describes his evaluative project as follows: we
should rid ourselves of values that conflict with will to power. With this guiding
remark in mind, it will be instructive to distinguish two ways in which a particular
value can conflict with will to power. (Nietzsche himself does not draw this distinction
explicitly.)
First, there can be an internal conflict between willing power and adopting some
value V, in the sense that willing power and adopting V is self-contradictory. That is,
valuing V entails taking there to be reason to A, whereas willing power entails taking
there to be reason not to A. Second, there can be an external conflict between willing
power and adopting V. An external conflict arises when, although adopting V and
willing power is not self-contradictory, adopting V does undermine one’s capacity to
will power, either in particular instances or in general. In other words, valuing V entails
taking there to be reason to A, where A-ing ultimately reduces one’s capacity to will
power.24
Let’s start with a case of internal conflict. Consider the negative value that is typically
accorded to suffering. Nietzsche argues that disvaluing suffering conflicts with will to
power.25 To will power is, in part, to aim at encountering resistance in the pursuit of
one’s ends. But encountering resistance causes displeasure. Indeed, some philosophers
have gone so far as to define suffering in terms of encountering resistance: as Schopen-
hauer writes, “we call [the will’s] hindrance through an obstacle placed between it and
its temporary goal, suffering” (WWR vol. I, 309). Thus, in willing power we will a
certain kind of displeasure or suffering. Nietzsche writes,
Human beings do not seek pleasure and avoid displeasure . . . What human beings want . . . is an
increase of power; driven by that will they seek resistance, they need something that opposes it—
Displeasure, as an obstacle to their will to power, is therefore a normal fact . . . human beings do
not avoid it, they are rather in continual need of it. (KSA 13:14[174]/WLN 264)
24
As a result, internal conflicts will tend to be global, whereas external conflicts can be more localized;
what generates an external conflict for a person with a given array of character traits and abilities may not
generate such a conflict for another person with different character traits and abilities.
25
For example, he writes, “The discipline of suffering, of great suffering—do you not know that only this
discipline has created all enhancements of man so far? That tension of the soul in unhappiness which cultivates
its strength, its shudders face to face with great ruin, its inventiveness and courage in enduring, persevering,
interpreting, and exploiting suffering, and whatever has been granted to it of profundity, secret, mask, spirit,
cunning, greatness—was it not granted to it through suffering, through the discipline of great suffering?”
(BGE 225). See also D 18, 146; GS 13, 338; Z Preface 3, II.3; BGE 39, 202, 229, 260; A 2, 7, 57. Bernard
Reginster explores this topic at length in his 2006; the discussion above is indebted to his insightful
explication of Nietzsche’s thoughts on suffering.
In this passage, Nietzsche notes that his will to power doctrine implies that we seek a
certain kind of displeasure, the displeasure attendant upon having one’s will opposed.
With this in mind, we can see why an internal contradiction arises in those who will
power and disvalue suffering. If one views suffering as disvaluable, then one takes there
to be reason to avoid suffering. However, part of what it is to will power is to aim at a
certain form of suffering, namely resistance (cf. KSA 13:11[77], GS 56). So, willing
power entails taking there to be reason to seek certain forms of suffering.26
In short, willing power is, in part, valuing suffering. This is why the view that
suffering is bad conflicts with the will to power. The claim that suffering is bad entails
that suffering is to be avoided. But the will to power doctrine claims that in each
instance of action, we aim at a certain kind of suffering. Consequently, an agent who
disvalues suffering is in a conflicted, self-contradictory state: he is simultaneously
attempting to avoid all forms of suffering and seeking out a particular type of
suffering.27,28
Now let’s turn to a case of external conflict between a value and will to power.
A value could conflict with will to power if, given background circumstances including
facts about the agent and her environment, having the value undermines the agent’s
capacity to will power (i.e., to encounter and overcome resistances). In other words:
the value entails that we have reason to perform some action A. Willing power does
not commit us to the claim that we have reason not to A. Thus, there is no internal
26
Notice that Nietzsche is not simply belaboring the truism that suffering or struggle is often instrumen-
tally necessary in order to bring about a desired goal. Rather, Nietzsche is claiming that part of what I aim at is
suffering. A homely example may make the contrast clearer. Everyone can agree that suffering is sometimes
instrumentally valuable: going to the dentist requires a bit of suffering, but the long-term benefits outweigh
the short-term costs. In that sense, the suffering induced by the dentist is instrumentally valuable. Contrast this
with a case in which the suffering is more intimately related to the activity in question, such as the activity of
running a marathon. As we noted in Chapter 6, the runner typically views marathon running as valuable
partly because it requires encountering and overcoming pain. (Of course, this is not to say that the runner
aims solely at pain. The runner could increase the pain of a marathon by, say, putting sharp objects in his
shoes. My point is rather that part of the runner’s attraction to marathon running is that marathon running is
painful and challenging.) In the dentist case, suffering is viewed as a regrettable, but necessary, means to a
valued end. In the marathon case, part of what is valued is suffering. Nietzsche’s doctrine of will to power
entails that suffering is sought in this second sense.
27
As Reginster puts it, “if . . . we take power—the overcoming of resistance—to be a value, then we can
see easily how it can be the principle behind a revaluation of suffering. Indeed, if we value the overcoming of
resistance, then we must also value the resistance that is an ingredient of it. Since suffering is defined by
resistance, we must also value suffering” (Reginster 2006, 177). Thus, “from the standpoint of the ethics of
power, suffering cannot be coherently deplored . . . ” (Reginster 2006, 233).
28
Thus, Nietzsche complains, “You want, if possible—and there is no more insane ‘if possible’—to
abolish suffering. And we? It really seems that we would rather have it higher and worse than ever” (BGE
225). Here, Nietzsche claims that whereas modern morality wants to abolish suffering, he would like to
preserve and heighten it. However, we might limit Nietzsche’s claims: there is an important difference
between the suffering associated with achievement, and the suffering associated with conditions such as
disease or hunger. Nietzsche’s arguments show that it would be contradictory to disvalue the first type of
suffering. However, they do not seem to show that it would be contradictory to disvalue the second type
(though Nietzsche does sometimes argue that the second type of suffering is instrumentally valuable, as a spur
to achievement, or as one more obstacle to overcome).
conflict. However, A-ing ultimately reduces our capacity to will power in the per-
formance of other actions.
Take one of Nietzsche’s favorite examples: asceticism. Suppose I am trying to decide
whether to value ascetic practices. There is no internal contradiction between asceticism
and will to power: as Nietzsche makes clear, ascetics typically manifest a high degree of
will to power by encountering and overcoming resistances to inflicting suffering on
themselves (cf. BGE 230). However, ascetic practices of the sort that Nietzsche rejects
enervate the person, lessening his capacity to pursue other ends.29 In particular, these
practices reduce one’s ability to pursue a variety of ends that require physical or mental
exertion. In that sense, asceticism inhibits one’s ability to encounter and overcome
resistance in the pursuit of one’s ends. Valuing asceticism will therefore conflict with
one’s capacity to overcome difficult resistances. Insofar as one aims at encountering and
overcoming as much resistance as possible, then, one will have reason not to value
asceticism.30
In quite different ways, then, disvaluing suffering and valuing asceticism both
conflict with the will to power. Disvaluing suffering requires avoiding will to power;
valuing asceticism involves manifesting high degrees of will to power, but in a way that
ultimately undermines one’s capacity to pursue power.31 Recognition of this point
should lead us to revise these valuations: we should not disvalue suffering, and we
should not value asceticism. This is what Nietzsche means when he says that the will to
power should serve as a principle of revaluation for other values.
One final remark is in order. Determining the extent to which a given value
conflicts with will to power is no easy task. The cases of suffering and asceticism are
straightforward; the conflicts between these values and power arise in fairly obvious
ways. But in other cases, the conflicts will be far subtler. In fact, it is often impossible to
read conflicts off the surface content of the value. For example, Nietzsche repeatedly
claims that valuing compassion conflicts with will to power, but there is certainly
29
Nietzsche rejects only certain forms of asceticism, while commending others. For example, Nietzsche
endorses forms of asceticism that consist of letting go of material and romantic interests in order to devote
oneself fully to a creative process (cf. GM III.8). However, he rejects the form of asceticism that require
inflicting suffering on oneself for its own sake (GM III.11–21).
30
The will to power doctrine also bears on asceticism in a second, less direct way. In the third essay of the
Genealogy, Nietzsche argues that asceticism arose as a response to the idea that meaningless suffering is
disvaluable. Asceticism gives suffering a meaning: crudely, we suffer in order to redeem ourselves before God.
We saw, above, that the will to power hypothesis implies that suffering is neither meaningless nor disvaluable.
Its meaning and value lies in its relationship to will to power. Thus, the will to power hypothesis undermines
the original motivation for accepting asceticism. And, of course, there is a third problem with asceticism:
Nietzsche claims that it is typically motivated by false beliefs about religion, an afterlife, otherworldly rewards,
and so forth.
31
Perhaps an analogy will help to illustrate the contrast. Suppose we are hedonists. Some actions would
conflict with hedonism internally, in the sense that they generate pains. Others would conflict with hedonism
externally, in the sense that they generate pleasure in the short term but pain in the long term (consider
smoking, taking certain kinds of drugs, and so forth).
nothing in the bare idea of compassion that would lead one to think that this is so.32
Instead, we must examine the way in which the value functions in society, the way in
which it interacts with other values, the projects it leads people to embrace, and so on.
This is one of the tasks of Nietzschean genealogy; I will examine it in the next chapter.
2.2 Why not jettison all values that are not derived from will to power?
The prior section examined the ways in which values can conflict with the will to
power. Even a casual reading of Nietzsche’s texts reveals that he believes these conflicts
are pervasive; most of our values are, to some extent, in conflict with will to power.
But one might raise a question at this point: if conflicts between will to power and
other values are so common, why not simply get rid of most other values? For example,
we might keep the various values that are directly derived from will to power, such as
the values of struggle and difficulty, while rejecting all values that cannot be so derived.
This dramatic proposal would leave us with a meager set of values, to be sure, but it
would fully eliminate conflicts between will to power and other values.
Nietzsche rejects this idea; he does not regard it as a realistic possibility. His response
involves a consideration of what is necessary in order for agents to inhabit a society. In
Daybreak and the Second Essay of the Genealogy, Nietzsche argues that in order to form
a society, human beings need to adopt some set of norms regulating conduct and
interpersonal relationships (cf. D 16 and GM II.1–3). In particular, human beings
must adopt norms that restrain and redirect aggressive tendencies, as well as norms
that render human beings “regular and calculable” (GM II.2). Put simply, Nietzsche
claims that in order to transition from a state of nature to a society, some set of norms
must come to be regarded as authoritative.33
The basic idea that Nietzsche is relying upon—that our existence in society demands
the adoption of norms governing interpersonal relationships—is familiar from Hobbes,
Rousseau, and others. But Nietzsche departs from the way in which this idea has
sometimes been developed: he argues that there are no substantive constraints on the
content of these norms.34 In other words, if we examine different sets of social norms,
32
EH I.4 says that “compassion is called a virtue only among decadents,” so that “the overcoming of
compassion I count among the noble virtues.” Analogously, A 7 claims that “compassion . . . preserves what is
ripe for destruction; it defends those who have been disinherited and condemned by life.” It “crosses those
instincts which aim at the preservation of life and at the enhancement of its value.” Moreover, “compassion
stands opposed to the tonic affects, which heighten the energy of the feeling of life: it has a depressing effect.
One loses strength, when one is compassionate” (A 7). See also D 134; GS 13, 338; BGE 225, 260, 269; GM
P 5.
33
Clark argues for a related point, writing that in GM II Nietzsche is working with the idea of a “social
contract” (Clark 1994, 28).
34
For example, consider a norm forbidding murder. This would seem to be a good candidate for a norm
that our social nature commits us to respecting. However, Nietzsche suggests that this norm could be
restricted in its application: it would function just as well if one restricted it to members of one’s immediate
social group. Thus, in GM II.16–17, Nietzsche considers early forms of society, in which a conquering group
exercises some restraints on its aggressive activities with respect to other members of the group, but vents
these aggressive activities on all others whom it encounters. Here, Nietzsche suggests that a society—at least
there is no external evaluative standard—except that provided by will to power—by

means of which we can say that one set is better or more correct than another set.35
Thus, as Nietzsche writes in Daybreak, there is a “mighty proposition with which
civilization begins: any custom is better than no custom” (D 16; cf. KSA 12:7[4]). Our
existence in society simply demands that we adopt some set of norms regulating
interpersonal relations, but it does not dictate the content of these norms.36 We can
express this point by saying that, just as a norm arises from our agential nature (the will
to power), so too a demand for norms arises from our social nature.
This is why we cannot simply jettison all values but will to power and those values
directly derived from it: we need to adopt additional values, in order to exist in
society.37 With these points in mind, we can see how the will to power doctrine
generates a “principle of revaluation” for all other values. We need to embrace certain
norms and values that regulate our interpersonal relationships. Certain sets of interper-
sonal norms come into greater conflict with will to power than do other sets of
interpersonal norms. Nietzsche thinks that these conflicts are pervasive and ineradic-
able; although different sets of values conflict to greater and lesser extents with the
values arising from our agential nature, there is no set of values that would completely
eliminate conflict. Our agential natures are always to some extent in conflict with our
social natures.38 In order to manage this conflict, we ought to embrace the sets of values
that conflict as little as possible with will to power.
2.3 Does the constitutive aim generate pro tanto reasons or all-things-considered reasons?
Suppose the arguments in the prior sections establish that the constitutive aim generates
reasons. Still, an objection might arise. The reasons generated by the constitutive aim
are merely pro tanto reasons; thus, they could be overridden by the reasons generated
by other aims. But, in presenting will to power as a standard of revaluation, doesn’t the
in a primitive form—can function just as well with the norm “don’t murder members of the group” as with
the norm “don’t murder.” Of course, philosophers who ground norms in our social nature argue that there is
more reason to adopt the universalized version of the norm. Nietzsche rejects this idea: the universalized
norm might be necessary for certain types of society, but not for society as such (compare BGE 259). No
doubt there are many ways of resisting Nietzsche’s conclusion. However, I will not explore them here, as the
point I wish to make is simpler: our social nature commits us to adopting some norms.
35
This is why, throughout his works, Nietzsche is at pains to emphasize that in early societies, the social
norms—so far from being rationally justifiable—are forcibly imposed on a servant class by a conquering class.
36
In an interesting study, Tamsin Shaw examines the way in which a version of this point is at work in
Nietzsche’s political thought. As she puts it, “Nietzsche believes that the kinds of certainty and agreement
necessary to forge political unity or establish de facto political authority are not likely to be achieved through
rational reflection” (2007, 144). On the contrary, Nietzsche believes that “reflection breaks down agree-
ment” (2007, 146). For reflection reveals that putative agreement has no deep foundations; it rests on nothing
more than convention. This generates a problem: although the modern state “requires normative consensus
in order to rule,” Nietzsche argues that “uncoerced consensus [is] impossible” (2007, 3).
37
See GS 76, where Nietzsche claims that “man’s greatest labor so far has been to reach agreement about
very many things and to lay down a law of agreement—regardless of whether these things are true or false. This
is the discipline of the head which has preserved humanity . . . ”
38
This is one of the themes of the Second Essay of the Genealogy.
Nietzschean view assume that the reasons generated by will to power should be
overriding in cases of conflict?
In fact, the Nietzschean theory is only committed to the claim that the constitutive
aim generates pro tanto reasons.39 In this respect, the constitutive aim is on par with all
other aims. Although this fact might seem to vitiate Nietzsche’s normative conclusions,
I will argue that it does not.
Let’s return to the chess example from Chapter 1. Chess has the constitutive aim of
checkmate. However, in Chapter 1 I pointed out that the reasons generated by the
constitutive aim may conflict with the reasons generated by other aims, such as the aim
of teaching a child how to play chess. Consider such a case: I see that I could achieve
checkmate in one move, but I believe that doing so would not be edifying for the
child. I therefore consider not checkmating him, and instead moving in a way that will
prolong the game. It is natural to assume that in some such cases the balance of reasons
will be in favor of not checkmating my opponent. So here we have a pro tanto reason
to checkmate, a pro tanto reason to prolong the game, and (let’s stipulate) an all-things-
considered reason to prolong the game. As this example indicates, chess players do not
always have an overriding reason to fulfill the constitutive aim of checkmate.
Just as the constitutivist about chess will not be bothered by the fact that chess players
might have reason not to pursue checkmate in a given move, so too the constitutivist
about action will not be bothered by the fact that agents sometimes have reason not to
perform the action that best fulfills the constitutive aim. Consider an example: suppose
I can either loaf about and watch television (call this action A), or I can continue
working on a difficult philosophical problem (call this action B). Action B generates far
more resistance, and thus better satisfies will to power. However, having worked on
the problem for several hours, I am tired and strongly motivated to take a break. These
motives incline me toward A-ing. In some cases of this form, the balance of reasons
may favor A-ing.
Why aren’t these results troubling for the constitutivist? Although in certain circum-
stances there is reason not to perform the action that best fulfills the constitutive aim,
things typically change when we consider longer-term actions and projects. To see
why, let’s alter the above example. Rather than temporarily departing from my
philosophical projects in order to relax, I consider wholly abandoning these projects
and spending my life doing nothing more than loafing about and watching television.
So we have a series of loafing actions, A1 to An, making up a larger action, A. And, as an
alternative, we have a series of working actions, B1 to Bn, making up a larger action,
B. The Nietzschean claim is that for each Ai, I have a will-to-power-derived pro tanto
reason not to engage in it, whereas for each Bi, I have a will-to-power-derived pro
tanto reason to engage in it. I also have other motives, so for some—but usually not
39
I do not claim that Nietzsche himself would put the point in this way. Here, I am engaged in rational
reconstruction.
all—of the Ai’s, I may have a motive and thus a reason to engage in it. However, if each
reason is weighted equally, I will typically have more reason to B than to A.
As this example indicates, in the long run we typically have more reason to perform
the action that better fulfills the constitutive aim. It is not the weight of the constitutive-
aim-derived reasons that generates this normative conclusion; it is their ubiquity and
pervasiveness.40
However, consider the following potential objection: an agent could have a
motive—or a set of motives—that is directly opposed to, and as ubiquitous as, the
will to power. In such a case, the balance of reasons might consistently come out against
performing the action that best fulfills will to power. Wouldn’t this undermine the
Nietzschean conclusions?
The Nietzschean constitutivist could respond to such an objection as follows: while
this is conceivable, it is highly unlikely. (Notice that the Nietzschean theory, being
empirically grounded in facts about human psychology, needn’t commit itself to the
claim that such a scenario is impossible or incoherent.) The constitutive aim typically
does not compete with an agent’s first-order motives. Rather, it modifies the manner in
which these motives are pursued. Consider two examples. The desire for food could be
easily satisfied by most individuals with sufficient resources—we could simply eat the
same thing every day. However, very few people do this: almost no one wants to eat
just anything. So, too, the desire for sex would be fairly easy to satisfy, except that
almost no one desires to have sex with just anyone. Agents acting on the desires for
food or sex are pursuing these goals in a way that introduces resistances—in other
words, they are willing power—but they typically don’t experience a conflict between,
say, the (easily fulfilled) desire to have sex with just anyone and the (more difficult to
fulfill) desire to have sex with a certain individual (or with an individual who has
particular characteristics). For agents typically don’t have the former desire. The will to
power shapes our motives, in such a way that our motives come to express not just a
desire for their determinate goals, but a desire to achieve those goals in a way that
introduces resistances.
This makes long-term conflicts between will to power and other motives unlikely:
our other motives are likely to reinforce, rather than compete with, the will to power.
Accordingly, in many cases our motives will generate pro tanto reasons to engage in the
actions that better fulfill will to power, and it is unlikely that there would be a persistent
conflict between will to power and our other motives. Thus, weighting all aim-derived
reasons equally, the balance of reasons is likely to come out in favor of those actions
that better fulfill will to power. 41
40
Velleman endorses a similar claim. He writes that the fact that an aim is constitutive of action “does not
entail that it was of greater importance or influence in [agents’] lives [than their other aims]; on the contrary, it
was of minor importance and influence—minor but also persistent and pervasive” (2009, 157).
41
For a fascinating reflection on the way that easily fulfilled desires tend to transform into desires that
engender new resistances, see Gass (1971).
In sum, the constitutive aim functions not by generating overriding reasons in

particular cases of action, but rather by weighting the balance of reasons in favor of
the projects and long-term actions that best fulfill will to power. Accordingly, there is
no guarantee that in all cases of conflict we will have more reason to perform the action
that better fulfills will to power: Nietzsche writes that “everything unconditional
belongs in pathology,” and this would surely include an unconditional requirement
to maximize the fulfillment of will to power (BGE 154). However, in the long run, the
balance of reasons typically favors actions that better fulfill will to power.42 Thus,
although the constitutive aim generates only pro tanto reasons, the ubiquity of these
pro tanto reasons engenders substantive normative conclusions.
3. Gradations of power: the problem of

differential realizability
Above, I have been relying on two assumptions: there are different gradations of
power, and we have more reason to realize the higher gradations than the lower
gradations. In other words, I have been assuming that will to power is a differentially
realizable aim, and that we have reason to pursue greater realizations of this aim. In this
section, I will defend these assumptions.
The idea that there are different gradations of power might seem problematic. To
say that we will power is to say that we aim at encountering and overcoming resistances
in the course of pursuing other, more determinate ends. So, if we treat will to power as
a normative standard, it tells us to pursue those ends that generate resistances and
obstacles. But this standard seems too vague, too formal, to generate any substantive
conclusions.
This point applies to the evaluation of both values and actions. To see this, consider a
pair of opposed, discrete evaluative judgments such as “murder is wrong” and “murder
is not wrong.” Which of these judgments would maximize the encountering and
overcoming of resistance? There seems to be no satisfactory way of answering this
question. Certainly, attempting to murder another person would typically be quite
difficult, so perhaps the principle “murder is not wrong” promotes more resistance. On
the other hand, part of the reason why murder is so difficult in our society is that it is
strongly disvalued. Thus, one might argue that the valuation “murder is wrong”
promotes more resistance for those tempted to murder. A determinate verdict seems
unachievable here. The same reasoning applies to a person who is trying to decide
42
Incidentally, I think this explains why Nietzsche’s ethical writings have a peculiar feature. Unlike many
other writers interested in ethics, Nietzsche very rarely focuses on the rightness or wrongness of particular,
discrete actions. Rather, his writings tend to focus on long-term actions, projects, and patterns of behavior.
We can now see why: these longer-term actions are where the import of the constitutive aim becomes
apparent.
G R A DAT I O N S O F P OW E R 201
whether performing a particular action of murder would generate more or less resist-
ance than the action of not murdering.
Simon May draws attention to a version of this difficulty, writing,
the problem of defining and measuring ‘power’ would be very great even if it referred simply to
efficient force or political control or, in general, stateable ‘outcomes’. But this problem seems
insuperable if all human behaviour in its inexhaustible variety, including such activities as
knowing and self-discipline, is to be explained in terms of power. . . . Although Nietzsche speaks
of value as directly correlated to a ‘scale of force’ (WP, 710), it is hard to see what such a common
scale of force might be. (May 1999, 27)
According to May, the notion of power is simply too vague and indeterminate to play
the role that Nietzsche assigns it.
May’s objection might appear decisive. He is certainly correct that there is no
satisfactory way of placing different episodes of willing on a scale of power. As he
notes, this would be hard enough if we localized our inquiry to one type of activity:
who pursues and overcomes more resistance in writing, Goethe or Melville? That
question is baffling, but, as May goes on to note, the question grows even more
intractable when applied to activities belonging to different types. The pursuit of great
literature, the pursuit of knowledge, the pursuit of athletic prowess, and the pursuit of
political power are all difficult, but in quite different ways. They may be incomparable.
Who pursues and overcomes more resistance: Emily Dickinson or Stephen Hawking
or Jesse Owens or Ghandi? It is not clear how one could even begin to answer this
question. Consequently, one must concede May’s point: we cannot array all actions on
a scale, and simply pick the one that generates the most resistance.
In short, it seems that we lack a differentially realizable notion of will to power. And
if that is right, the will to power thesis cannot generate any substantive normative
conclusions.
Fortunately, I believe this objection to the will to power thesis can be answered.43
Two considerations are relevant. First, in order to generate substantive results, a theory
needn’t generate a determinate ranking of all actions or all values. As I pointed out in
Chapter 3, a theory can be substantive without being exhaustive. An adequate ethical
theory must be capable of judging that for some actions A and B, there is more reason to
A than to B. It needn’t, however, entail that for all actions A and B, there is a fact of the
matter about which action we should perform. Thus, it will be sufficient for our
purposes if Nietzsche’s will to power thesis can rank some subset of actions or values.
Put differently, a moral theory needn’t reach into every aspect of our lives; it is
sufficient if it structures some portions of our lives. Can Nietzsche’s theory do so?
43
May agrees with a version of this point, arguing that if the will to power thesis is supplemented in
certain ways (by Nietzsche’s notions of sublimation and form-creation), then it can generate determinate
results. See May (1999).
This brings us to the second point. Nietzsche’s texts display a sustained concern with
the broad patterns of behavior induced by particular moral systems; but they devote
relatively little time to the examination of discrete moral judgments such as “lying is
wrong” or “promises should be kept.” There is a principled reason for this. The will to
power thesis is easier to apply when we examine long-term actions and sets of values,
rather than discrete actions and particular values. There is often no way to tell in
advance whether a particular action constitutes a high realization of will to power.
Nietzsche draws attention to this point when he writes (with a touch of exaggeration),
“the only possible critique of morality” is “a rigorous and courageous [strengen und
muthigen] attempt to live in this or that morality” (D 195; cf. D 61). Here, he indicates
that problems with moral systems become apparent only when these systems are lived.
So I suggest that the Nietzschean theory plays a more retrospective than prospective
role. It is not addressed to an imaginary agent standing outside all valuations and
determining without any evaluative presuppositions which values to adopt. Rather,
it is intended to diagnose and correct courses of action and sets of values that we have
already embraced. It tells us how to go on, not how to begin.
This is one way in which Nietzsche’s theory differs from our familiar models, such as
Kantian ethics. Kant’s theory is prospective. As Korsgaard emphasizes, it concerns “the
first-person position of the agent,” for it is addressed to “the agent who is deliberating
about what to do” (Korsgaard 1996b, 16; 2009, 131). But, as we’ve seen above, there
are good reasons for doubting that the deliberating agent has any deep knowledge of
what she is doing. Deliberation is driven by motives that lie in the background; it is
pervasively influenced in ways that we cannot detect in the moment of choice. What
I am doing, when I deliberate, is often not transparent to me until much later. And for
this reason, an ethical theory that addresses itself solely to the deliberating agent is at
best incomplete. Atomic acts of choice are no doubt of immense importance; but they
are also a kind of surface, which disguises deeper tendencies in the person’s mental
economy.
Having recognized these points, Nietzsche develops a theory that is intended to
function in a different way; a way that, I think, is more psychologically realistic. Most
of our ordinary decisions are habitual or conventional. We act in familiar ways, on
familiar values. There is nothing wrong with this; there is no realistic alternative, and it
is not a failing. But we are capable, from time to time, of stepping back and doing
something more: we are capable, as Nietzsche likes to put it, of “vivisection.” He
enjoins us to apply “a vivisecting knife directly to the chest of the virtues of the age”
(BGE 212). The image of vivisection, which occurs several times in Nietzsche’s works,
suggests that real ethical inquiry does not consist merely in examining the surface
content of moral judgments; we don’t simply assess the discrete moral judgments
preached by the common man, judgments such as “murder is wrong” or “you should
help others.” Instead, we cut through these surface judgments, trying to find the deeper
motives, implicit principles, defunct ideals, conceptions of agency, and so forth, that
A B I PA RT I T E T H E O RY : A G E N T I A L A C T I V I T Y A N D P OW E R 203
underlie them. We try to understand how these values impact our affects and the ways
of classifying and distinguishing actions that seem natural to us.
So, part of Nietzsche’s point is that morality is not merely present in explicit moral
judgments. For the particular moral system that the agent embraces will influence not
just his explicit thoughts about what is right and wrong, good and bad, but his very
perceptions of the world. This is why Nietzsche writes that there “are no experiences
other than moral ones, not even in the realm of sense-perception” (GS 114). It is these
complex perceptions and evaluative orientations that must be dissected and examined.
It is not possible to examine these phenomena at the moment of choice; vivisection is
not something carried out in an instant. But it can stand alongside these moments of
choice, as a crucial corrective.
Once we keep this in mind, the objection—that the will to power thesis doesn’t
generate results—seems misguided. It’s true that the will to power thesis doesn’t tell us
whether to keep our promises or steal or jaywalk. It’s true, as well, that the will to
power thesis doesn’t enable us to rank Melville and Dickinson and Owens. But
Nietzsche is not imagining that it would. Rather, he envisions us from time to time
subjecting our values and long-term activities to this standard. He envisions us doing
what his own books do: engaging in extended, prolonged reflection on these values,
goals, and standards, gradually coming to shed some and retain others. It will be
enough for Nietzsche if the will to power standard leads us to rethink our attitudes
toward and evaluations of some of our central concerns, such as suffering, compassion,
and the like. In the next chapter, I will give an example of how this proceeds.
So, to answer the objection: will to power is differentially realizable, but this does
not mean that we can discern in advance whether particular actions are higher or lower
manifestations of power. Nor does it mean that, even after all the facts are in, we will be
able to rank (say) Melville and Goethe. What it does mean is that we can discern ways
in which certain values and courses of action reduce power. In the next chapter,
I provide a detailed examination of how this is so.
So we have answered one objection: will to power is differentially realizable. But we
are left with another objection: why are we committed to aiming at maximal power?
Why not just aim at a bit of power, instead? I address this point in the next section.
4. A bipartite theory: agential activity and power

Above, I have explored the way in which the will to power thesis generates normative
conclusions. Will to power functions in conjunction with the agent’s other values,
thereby generating conclusions about which actions the agent has reason to perform. In
addition, will to power serves as a principle of revaluation for these other values: some
of the agent’s values will generate conflicts with will to power, and hence must be
rejected.
However, I have so far ignored one final aspect of the constitutivist theory that
I wish to defend. In Chapter 5, I argued that agential activity is a constitutive aim of
action. According to my account, an agent is active iff the agent approves of her action,
and this approval is stable in light of further information about the action’s etiology.
Given that agential activity, so construed, is a constitutive aim, we can derive a standard
of success for action: an action is successful if the agent approves of it, and this approval
would be stable given further information about the action’s etiology. Chapter 6
pointed out that the results generated by this standard will depend on the agent’s
values. After all, what I approve of depends on what I value. As the standard of agential
activity did not by itself give us a way of assessing values, we turned to will to power in
an effort to locate such a standard. We have now seen that will to power does, indeed,
give us a way of assessing values.
But, in light of these results, agential activity might seem irrelevant as a constitutive
aim. All the work seems to be done by will to power. Couldn’t we therefore leave off
agential activity, and simply focus on power?
In this section, I will argue that agential activity does, indeed, play a crucial role in
Nietzschean constitutivism. To make this point, I begin, in Section 4.1, by introducing
a dispute between Nietzsche and Schopenhauer. I argue that Schopenhauer agrees
with a version of Nietzsche’s claim that all actions aim at power. However, Schopen-
hauer takes this as a reason to reject action. Section 4.2 asks why Nietzsche reaches a
different conclusion. I argue that the account of agential activity can be invoked to
explain why we ought to will power, rather than reject action. Section 4.3 shows that
agential activity also plays another role: it enables the Nietzschean ethical theory to rely
only on the sparest of claims about the connections between aims and reasons.
4.1 Schopenhauer’s argument for the negation of willing
Schopenhauer argues that all actions are manifestations of the “will to live.” The “will
to live,” Schopenhauer tells us, “constitutes the inner nature of everything, and lives in
all” (WWR, vol. I, 372). The character of this will to live is remarkably similar to
Nietzschean will to power. Schopenhauer describes it as follows:
The will dispenses entirely with an ultimate aim and object. It always strives, because striving is its
sole nature, to which no attained goal can put an end. Such striving is therefore incapable of final
satisfaction; it can be checked only by hindrance, but in itself it goes on forever. (WWR, vol. I,
308)
So the will to live is ceaseless striving, which cannot take satisfaction in any final
attainment, but compels us to will endlessly. In this respect, the will to live is analogous
to the will to power: both are characterizations of the fact that we aim at attaining rather
than attainment.44
44
I am not claiming that Schopenhauer’s will to live is exactly analogous to Nietzsche’s will to power, but
only that they share this common core. For a discussion of some important differences between will to live
and will to power, see Aydin (2007).
A B I PA RT I T E T H E O RY: A G E N T I A L A C T I V I T Y A N D P OW E R 205
Thus, while Schopenhauer and Nietzsche characterize willing in somewhat different

ways, they agree on the underlying idea that all actions manifest a ceaseless, indeter-
minate striving. This striving is not directed toward any particular end, but simply
toward activity. At this level of generality, Schopenhauer and Nietzsche are in agree-
ment.
However, Schopenhauer and Nietzsche react to the omnipresence of this will in
strikingly different ways. Put simply, Schopenhauer takes the presence of this will as an
objection to life, whereas Nietzsche takes it as something that is normative for us. Thus,
Schopenhauer writes, “this great intensity of willing is in and by itself and directly a
constant source of suffering . . . ” (WWR, vol. I, 363). He provides a vivid example:
The futility and fruitlessness of the struggle of the whole [of life] are more readily grasped in the
simple and easily observable life of animals . . . Instead of [imagining that there is some final aim or
purpose for animal life], we see only momentary gratification, fleeting pleasure conditioned by
wants, much and long suffering, constant struggle, war of all, everything a hunter and everything
hunted, pressure, want, need, and anxiety; shrieking and howling; and this goes on in secula
seculorum [forever], or until the crust of the planet breaks. Junghung relates that in Java he saw an
immense field entirely covered with skeletons, and he took it to be a battlefield. However, they
were nothing but skeletons of large turtles five feet long, three feet broad, and of equal height.
These turtles come this way from the sea, in order to lay their eggs, and are then seized by wild
dogs (Canis rutilans); with their united strength, these dogs lay them on their backs, tear open
their lower armor, the small scales of the belly, and devour them alive. But then a tiger often
pounces upon the dogs. Now all this misery is repeated thousands and thousands of times, year in
and year out. For all this, then, are these turtles born. For what offense must they suffer this
agony? What is the point of this whole scene of horror? (WWR, vol. II, 354)
Due to the omnipresence of the will to live, life overflows with pointless suffering.
Thus, Schopenhauer reasons,
Unless suffering is the direct and immediate object of life, our existence must entirely fail of its
aim. It is absurd to look upon the enormous amount of pain that abounds everywhere in the
world, and originates in needs and necessities inseparable from life itself, as serving no purpose at
all and the result of mere chance. Each separate misfortune, as it comes, seems, no doubt, to be
something exceptional; but misfortune in general is the rule. (Schopenhauer 2001, Vol. II, 291)
In the face of these facts, Schopenhauer reasons that it would be best to bring about a
“denial” or “self-suppression” of the will to live (WWR, vol. I, 404). Rather than
ceaselessly striving under the pressure of an unfulfillable aim, we should try to rid
ourselves of that aim. Schopenhauer claims that if we could bring about this “denial of
willing,” we would achieve an “entrance into freedom” (WWR, vol. I, 404).45
45
Although Schopenhauer bemoans the alleged futility of life, he does not think we can directly negate the
will to live. After all, if every action manifests will to live, then the action of negating the will to live would
itself manifest the will to live. Schopenhauer illustrates this point by examining the most dramatic way in
which one can negate the will to live: by committing suicide. Surprisingly, he argues that suicide turns out to
be an affirmation of the will to live, rather than its negation:
And yet Nietzsche draws the opposite conclusion: far from negating or denying “the
essence of life, its will to power” (GM II.12), we should treat the will to power as
standard-setting, and seek to fulfill it. So, whereas Schopenhauer takes ceaseless striving
as an objection to life, Nietzsche takes it as an inducement to life. Whereas Schopen-
hauer takes ceaseless striving as a mark of life’s futility, Nietzsche sees in it life’s value.
Are these reactions equally legitimate? If so, the Nietzschean theory would face a
potential problem: if an agent can coherently regret the presence of inescapable aims
and therefore seek the elimination of action, this might undercut the alleged normative
authority of action’s constitutive aim. In the next section, I argue that the notion of
agential activity enables us to answer this concern.
4.2 The solution: agential activity and power
It is here that the notion of agential activity returns to center stage. In Chapter 5,
I argued that agential activity is a constitutive aim of action. Whenever an agent
reflectively A’s, she aims both to approve of her A-ing, and to have this approval be
Far from being denial of the will, suicide is a phenomenon of the will’s strong affirmation. For denial has its
essential nature in the fact that the pleasures of life, not its sorrows, are shunned. The suicide wills life, and is
dissatisfied merely with the conditions on which it has come to him. Therefore he gives up by no means the
will to live, but merely life . . . (WWR, vol. I, 398)
In the above passage, Schopenhauer points out that suicide involves the shunning of life’s sorrows. The
suicide wants to live differently; he wants pleasures instead of pains. More generally, he wants life to have a
different character. So he doesn’t give up his will to live; on the contrary, by killing himself, he affirms his will
to live. In particular, he affirms his desire for pleasure so much that he would rather kill himself than accept life
with its current character.
This is a perfectly general point about the nature of action:
Just because the suicide cannot cease willing, he ceases to live; and the will affirms itself even here through the
cessation of its own phenomenon, because it can no longer affirm itself otherwise. (WWR, vol. I, 399, emphasis
added)
Every act of willing, including the act of willing the end of one’s willing, manifests the will to live.
Is there any way genuinely to negate the will to live, then? To do so, one would have to avoid performing
any actions whatsoever. Schopenhauer outlines two possible methods for bringing this about. First, he points
to a curious phenomenon, of which there are a few records: death by voluntary starvation. He claims that
certain forms of this phenomenon might indeed amount to rejections of the will to live. In particular, if the
individual becomes so indifferent to willing, so detached from his own needs, that he ceases to perform even
the rudimentary actions of feeding himself, then this would be a form of negating the will to live. Of course,
this would be very rare: the agent who deliberately tries to starve himself is different than the agent who is so
indifferent to his needs that he (non-actionally) ends up starving (cf. WWR, vol. I, 401). Second, negation of
the will to live can simply happen to us:
we have seen that self-suppression of the will comes from knowledge, but knowledge and insight as such are
independent of free choice, that denial of willing, that entrance into freedom, is not to be forcibly arrived at
by intention or design, but comes from the innermost relation of knowing and willing in man; hence it comes
suddenly, as if flying in from without. (WWR, vol. I, 404)
In other words, we can become so attentive to the omnipresence of suffering, or so attuned to the
senselessness of the will to live, that we non-voluntarily give up willing. The general desire for action simply
withers away in us, much as particular desires tend to wither away once we recognize their futility.
Thus, Schopenhauer claims that renunciation of the will to live is not a voluntary action, not something
one can set about to achieve under that description. It is, instead, something that must happen to us. For any
action at all—even the action of trying to negate the will to live—would express the will to live.
A B I PA RT I T E T H E O RY : A G E N T I A L A C T I V I T Y A N D P OW E R 207
stable in the face of further information about A-ing’s etiology. With the will to power
thesis at hand, we can see that a new wrinkle is added to this notion. If the Nietzschean
account of action is right, then the will to power is part of every action’s etiology. That
is, every action aims, in part, at encountering and overcoming resistance. Thus, we can
adduce the following criterion: when an agent A’s, she is committed to approving of
A given knowledge of the fact that A aims at power.
If the fact that A aims at power disrupts the agent’s approval of A-ing, she will fail to
meet the standards set by the constitutive aim of agential activity. More generally, if she
takes the fact that A aims at power to cancel her approval of A-ing, then all of her actions
will fail to be manifestations of agential activity, and all of her actions will therefore fail
to meet the constitutive standard of agential activity. In short: unless the agent approves
of the fact that her actions are manifestations of the will to power, she cannot be active.
Thus, we can give the following argument:
(1) If the agent performs an action A-ing, she is committed to agential activity. That
is, in A-ing she is committed to approving of her A-ing, and to having this
approval be stable given further facts about A-ing’s etiology.
(2) The etiology of every action includes will to power.
(3) Thus, in order for the agent to be active, the agent’s approval must be stable
given further facts about the way in which will to power motivates her.
(4) In this sense, the agent must approve of will to power as a motivating force.
In short, aiming at activity commits us to approving of will to power. Merely in virtue
of acting, we become committed to approving of will to power.46
Consider again what this means. We are committed to acting in ways that we can
endorse. The only way we can do that, though, is to endorse will to power. So the only
way we can act is to endorse will to power.
Accordingly, the difference between Schopenhauer and Nietzsche is more than a
difference in taste. The Nietzschean theory can employ the notion of agential activity
in order to support Nietzsche’s approval of will to power.
In fact, the notion of agential activity not only gives Nietzsche a response to
Schopenhauer, but also enables him to answer an additional, related objection. Recall
that in Chapter 2, I introduced the “Why Bother” objection to constitutivism. The
objection proceeded as follows. Suppose we show that action constitutively aims at G,
and suppose G can be fulfilled to different degrees. All that it takes for something to be
an action is for it to fulfill G to some extent. So, if an agent is committed to performing
actions, she is committed to fulfilling G to some extent. But why bother fulfilling G to
the highest degree? For example, if action constitutively aims at power, then all actions
46
This surprising result helps us to see why Nietzsche often speaks of affirming life or saying yes to life. After
all, he notes that “the essence of life” is “its will to power” (GM II.12). Thus, in “affirming” or “saying yes”
to life, we affirm or say yes to will to power.
manifest some degree of encountering and overcoming resistance. So why bother

striving to encounter and overcome maximal resistance?
To drive this point home, note that an agent will never fulfill all of her aims maximally.
Our aims are too disparate and conflicting to permit full realization. For example, an
agent might aim at living in Boston and living in New York, at spending all of her time
traveling and also at holding down a job, at being a doctor and at being a philosopher, and
so forth. Agents have too many aims, and experience too much conflict among them, to
fulfill all of them even to some degree, much less to fulfill all of them maximally. This
might be regrettable, but it is unavoidable. Accordingly, why not lump will to power in
with the host of other aims that I won’t even bother trying to fulfill maximally?
I gave one response to this objection in Section 2.2: due to the omnipresence of will
to power, systematically neglecting its fulfillment will involve systematically ignoring
reasons for action. I may have an aim of traveling as often as possible that I frequently
decline to fulfill maximally, but this kind of aim is not present in everything that I do.
Will to power, on the other hand, is. This is a crucial difference between will to power
and other aims.
However, the notion of agential activity gives us a stronger response. It might be
true that if you have an aim to which you are indifferent, or of which you disapprove,
then you will justifiably decline to fulfill it maximally. But if you have an aim of which
you approve, things seem different. I neglect my aim of constantly traveling in part
because I do not fully approve of it—though it holds considerable appeal, it also seems
comparatively trivial, impractical, and inconsistent with too many of my other goals.
Will to power, on the other hand, is an aim to which agents are committed in every
instance of action, an aim that agents affirm in every instance of action. In approving of
it, agents do accept it as a standard of success.
This gives us another response to the “why bother?” objection. If an agent had an
aim whose presence she regretted or did not approve of, it would make sense for her to
forgo fulfillment of that aim, or to fulfill it only partially, in order to fulfill other aims.
But we cannot coherently muster this attitude toward will to power. Merely in virtue
of acting, we are committed to approving of will to power. For, if we don’t approve of
will to power, we won’t be able to approve of any particular action at all. Insofar as
I acknowledge that I care about any particular end that I pursue, then, I commit myself
to approving of power. Consequently, the “why bother?” objection cannot arise:
I cannot simultaneously approve of fulfilling my aim in every instance of action and
ask why I should strive to fulfill it. In approving of it, I commit myself to its fulfillment.
4.3 An additional role for agential activity
Agential activity not only enables to answer Schopenhauer’s argument and the “why
bother?” objection; it also plays a third role. It enables us to weaken the normative principle
that the Nietzschean theory relies upon to generate ethical results. Let me explain.
In generating normative conclusions from the fact that action has a constitutive aim,
I have been relying on the following principle:
CONCLUSION 209
(Success) If X aims at G, then G is a standard of success for X-ing.

As I pointed out in Chapter 1, this principle is relatively uncontroversial and widely
accepted.
However, given that agential activity requires us to approve of our aiming at power,
we could employ an even weaker principle:
(Endorsement) If an agent aims at G, and the agent endorses this aim, then G is a
standard of success for the agent’s action.
This principle would also generate normative results from the fact that we aim at
power. After all, we are committed to endorsing will to power, so will to power would
be a standard of success.
The Endorsement principle is weaker than Success because it incorporates an
additional condition. We must ask not only whether the agent aims at G, but also
whether the agent approves of this aim. Presumably, agents have many aims of which
they disapprove. For example, suppose Bill has a desire to procrastinate, but disapproves
of this desire. If we embrace Success, Bill has a reason to procrastinate.47 If we embrace
Endorsement, he does not. In this respect, Endorsement might seem more appealing
than Success.48
Nietzsche himself does not distinguish between these two claims. However, it is
worth noting that either claim would enable the Nietzschean theory to succeed.
Readers who are skeptical of Success but accept Endorsement can be happy with the
Nietzschean argument, and the converse is also true.
5. Conclusion
In this chapter, I have explained how the Nietzschean version of constitutivism
functions. We assess our actions in light of will to power and our other values. These
additional values must themselves be assessed in terms of power. Some fail this test,
because they generate conflicts with our valuation of power.
47
Notice, though, that Success would not necessarily entail that Bill has an all-things-considered reason to
procrastinate. If his other aims conflict with procrastinating, he may have an all-things-considered reason not
to procrastinate.
48
In addition, those with Frankfurtian or Korsgaardian sympathies will prefer Endorsement to Success.
Frankfurt appeals to a principle analogous to Endorsement in his 2004, claiming that desires count as reasons
only when the agent identifies with the desire. Korsgaard claims that motives count as reasons only when the
agent endorses the motive (Korsgaard 1996b, 94). For a subtle discussion of these issues, see Cuypers (2000),
who argues that some “irresistible inclinations” and “insurmountable inhibitions” should count as “external”
to the agent because they are not “endorsed” by the agent (Cuypers 2000, 245). Cuypers continues “forces
and obstacles which are not under a person’s direct and immediate control may nevertheless be internal to
him, and hence autonomous, in virtue of his endorsement of them. This endorsement accounts for the
difference between the autonomy of volitional necessities and the heteronomy of compulsions and aversions
which are not so endorsed” (Cuypers 2000, 245). Thus, he concludes, “autonomy and necessity are perfectly
compatible” (Cuypers 2000, 245).
I have also discussed the way in which the constitutive aim of agential activity interacts
with the constitutive aim of power. By appealing to agential activity, we can establish that
merely in virtue of acting, we are committed to approving of will to power. Using this
result, the Nietzschean theory can answer Schopenhauer’s contention that we should
reject willing. In addition, the theory can be based on an exceedingly spare claim about
the relationship between aims and reasons (Endorsement).
I conclude that a Nietzschean theory of action provides the materials for a successful
argument for constitutivism. If we accept the Nietzschean account of action, as I have
argued that we should, then it follows that action has the constitutive aims of activity
and power. This theory succeeds in generating substantive normative conclusions.
But one part of the theory remains to be discussed: what are those normative
conclusions? What, specifically, does this Nietzschean ethical theory tell us to do and
to value? The next chapter will answer these questions.
8
The Normative Results Generated
by Nietzschean Constitutivism
The previous chapters argued that a Nietzschean theory of action yields two consti-
tutive aims: agential activity and power. The last chapter investigated the structural
features of this ethical theory. However, we have yet to examine the specific, substan-
tive normative conclusions entailed by the theory. That is the topic of the present
chapter.
There are two ways that we could pursue this goal. First, we could investigate what
Nietzsche himself believes the ethical consequences of his will to power thesis are.
After all, his texts are replete with claims about various values conflicting with will to
power, so an examination of these claims might be illuminating. Second, we might step
back from Nietzsche and simply determine, on our own, what results the theory
generates.
Each of these strategies has advantages and disadvantages. Examining the results of
Nietzschean constitutivism on independent grounds would be, in certain respects,
much easier; after all, we would thereby free ourselves from interpretive matters. On
the other hand, Nietzsche’s texts contain subtle, nuanced analyses of the relationship
between will to power and various values, so it seems foolhardy to set these aside.
Moreover, many readers associate Nietzsche’s will to power thesis with a host of
shocking and deeply counterintuitive evaluative claims; some readers will assume
that the Nietzschean theory that I am defending generates such paradoxical conse-
quences as to render it at best highly questionable. In light of these concerns, I think it
best to face these difficulties head on, by examining some of Nietzsche’s most difficult
claims about power.
Accordingly, I will pursue a mixed strategy. I will start with an exploration of the
normative results that Nietzsche claims for his will to power thesis; however, I will then
argue that in certain respects, Nietzsche’s conclusions go wrong. In short, I will
combine analyses of Nietzsche’s texts with independent critique. By the end of this
chapter, I hope to show what sorts of ethical results we should expect from the
Nietzschean version of constitutivism.
As I mentioned, I want to start by facing a difficulty head on. Many readers have
thought that a Nietzschean ethical theory generates horrific consequences, enjoining us,
for example, to subjugate other individuals whenever doing so is to our own advantage.
212 T H E N O R M AT I V E R E S U LT S
This interpretation is certainly understandable: Nietzsche’s texts seem designed to

induce, in his readers, the idea that willing power involves absolutely no constraints on
interpersonal behavior. Indeed, he often seems to suggest that will to power beckons us to
engage in the most revolting acts of oppression, cruelty, and so on. A few passages:
The weak and the failures shall perish: first proposition of our love for humanity. And one shall
even help them to it. (A 2)
A human being who strives for something great considers everyone he meets on his way either as
a means or as a delay and obstacle—or as a temporary resting place. (BGE 273)
These passages suggest that Nietzsche’s view yields an obligation to express one’s
power on others, without regard for the deleterious effects this may have on them.
In addition, Nietzsche frequently condemns values such as equality and justice:
We simply do not consider it desirable that a realm of justice and harmony should be established
on earth. (GS 377)
‘Equality’, a certain factual growing-similar which merely brings itself to expression in the theory
of ‘equal rights’, belongs essentially to decline. (TI IX.37)
The overall degeneration of humanity down to what today appears to the socialist dolts and
flatheads as their ‘human being of the future’—as their ideal!—this degeneration and diminution
of humanity into perfect herd animals (or, as they say, to human beings of the ‘free society’), this
animalization of humanity into dwarf animals of equal rights and claims, is possible, there is no
doubt of it! Anyone who has once thought through this possibility to the end, knows a greater
disgust than other human beings—but perhaps also a new task! (BGE 203)
Here, open praise of hierarchy is combined with the claim that equality is harmful.
It would be a mistake to ignore these passages. After all, they are ubiquitous in
Nietzsche’s works, and may represent just the kind of disagreeable truths that he beseeches
us to endure. However, it is important to recognize that Nietzsche also writes things like:
It goes without saying that I do not deny—unless I am a fool—that many actions called immoral
ought to be avoided and resisted, or that many called moral ought to be done and encouraged—
but I think one should be encouraged and the other avoided for other reasons than hitherto. (D 103)
The good four.—Honest towards ourselves and whoever else is a friend to us; brave towards the
enemy; magnanimous towards the defeated; polite—always: this is what the four cardinal virtues
want us to be. (D 556)
When the exceptional human being treats the mediocre more tenderly than himself and his
peers, this is not mere politeness of the heart—it is simply his duty. (A 57)
In these passages, Nietzsche suggests that there are constraints on how we should
interact with others: so far from an injunction to manifest open aggression and
dominance, Nietzsche suggests that there is an obligation to be honest, magnanimous,
polite, and so forth.
T H E N O R M AT I V E R E S U LT S 213
In addition, it is revealing that Nietzsche’s ethical exemplars are not the sadists, the
murderers, and their ilk, but are instead individuals such as Goethe, Beethoven, and
Nietzsche himself (see, for example, TI IX.49 and EH IV). These are individuals who
pursued great, immensely difficult tasks, and in that way fulfilled the demands set by
will to power. However, the tasks that these individuals engaged in were primarily
intellectual, and hardly involved the subjugation of others. As Nietzsche himself puts it,
“the men of great creativity” are “the really great men, according to my understand-
ing” (KSA 11:37[8]).1
Indeed, none of these intellectual accomplishments in literature, music, or philoso-
phy would have been possible without a fairly sophisticated society underpinning and
supporting them. Goethe wasn’t engaged in acts of domination and oppression of a
servant class; he was a figure who devoted the majority of his life to writing literature
and practicing law. Beethoven wasn’t an oppressive, savage member of the aristocracy;
on the contrary, he was dependent on the financial support of his aristocratic patrons.
Nietzsche wasn’t an aggressive, domineering figure; he was a sickly man relying for his
survival on a lifelong university pension. These institutions and practices—the ability to
have a profession, to enjoy support for one’s artistic endeavors, to receive a lifelong
income for past services rendered—hardly seem to be impeded by egalitarian social
institutions. On the contrary, the societies that one thinks of as most egalitarian tend to
be most heavily invested in these institutions and practices.
As these remarks indicate, there is a real ambivalence in Nietzsche’s texts concerning
the interpersonal normative consequences of his will to power thesis. On the one hand,
he sometimes suggests that the will to power thesis either recommends a complete
absence of constraint on interpersonal relations or a positive injunction to engage in
vicious behavior. On the other hand, he sometimes claims that we must conform even
to the minutiae of politeness and etiquette, and he heaps praise on individuals who
would manifestly fail to flourish without the support of their societies. In the following
sections, I will try to make sense of these claims about will to power. Exactly what sorts
of interpersonal normative claims does the Nietzschean theory generate?
I will suggest that while the Nietzschean constitutivist theory does have some
counterintuitive results, and while it does require reassessment of some cherished
values, it hardly licenses oppression, domination, and the like. On the contrary, it is
compatible with certain forms of egalitarianism, and enjoins us to engage primarily in
creative endeavors. I will show, then, that the normative conclusions generated by
Nietzschean constitutivism are not, as some might fear, the conclusions that might be
hoped for by a sociopath or a tyrant. They are, instead, conclusions that urge us to strive
for a challenging vision of flourishing. Indeed, one of the first scholars to take
Nietzsche seriously put it well: “Nietzsche attacks morality in every contemporary
1
Nietzsche’s frequent praise of a less docile individual—Napoleon—raises more questions. I will address
this point below.
form in which he finds it, not in order to remove men’s chains, but rather to force men,
under a heavier burden, to attain a higher rank” (Jaspers 1997, 140).
I begin, in Section 1, by considering the way in which Nietzschean constitutivism
undermines a central tenet of modern morality: its insistence that the same set of values
is appropriate for all rational agents. Section 2 asks how egalitarian values fare in the
Nietzschean ethical theory. Finally, Section 3 examines the individuals who Nietzsche
presents as ethical exemplars. At the conclusion of this chapter, then, we will have an
overview of some of the central normative implications of Nietzschean constitutivism.
1. Nietzsche’s rejection of universalism

One of Nietzsche’s most frequently voiced complaints against modern morality is its
assumption that one set of values is appropriate for all agents. Nietzsche regularly asserts
that, on the contrary, different sets of normative claims—different moralities—are
appropriate for different kinds of agents. For example, he writes,
My philosophy aims at an ordering of rank: not at an individualistic morality. The ideas of the
herd should rule in the herd—but not beyond it: the leaders of the herd require a fundamentally
different valuation for their own actions, as do the independent, or the ‘beasts of prey,’ etc. (KSA
12:7[6])
Call universalism the claim that the selfsame values are appropriate for all agents. In the
above passage, Nietzsche rejects universalism.2
How, exactly, does the will to power thesis entail a rejection of universalism? To
start, recall that conflicts between will to power and other values come in two forms:
internal and external. An internal conflict arises when willing power and adopting
some value V is self-contradictory. An external conflict arises when, although adopting
V and willing power is not self-contradictory, adopting V does undermine one’s
capacity to will power, either in particular instances or in general.
Let’s begin with internal conflicts. Notice that internal conflicts between power and
a given value will hold for all individuals. Thus, if a value preached by contemporary
morality leads to an internal conflict, it will be bad for everyone.3 To pick one of
Nietzsche’s examples, contemporary morality teaches that suffering is always disvalu-
able. Yet Nietzsche argues that this generates what I have called an internal conflict with
will to power. Accordingly, everyone is harmed, at least to some extent, by the
2
There are many similar passages. Nietzsche writes that a “morality that takes itself for unconditional and
addresses itself to all does not only sin against taste . . . Moralities must be forced to bow first of all before the
order of rank; their presumption must be brought home to their conscience—until they finally reach
agreement that it is immoral to say: ‘what is right for one is fair for the other’ ” (BGE 221). Analogously, he
contends that “what is fair for one cannot by any means for that reason alone also be fair for others . . . the
demand of one morality for all is detrimental to the higher men; in short . . . there is an order of rank between
man and man, hence also between morality and morality” (BGE 228). See also D 174; GS 55; BGE 198, 225,
259, 260; GM I.13; A 11; EH IV.4.
3
Moreover, it will be bad at all times.
NIETZSCHE’S REJECTION OF UNIVERSALISM 215
disvaluation of suffering. In teaching that suffering is bad for everyone, contemporary

morality generates conflicts for everyone.4,5
Internal conflicts, then, are universally applicable. There is, however, a caveat:
notice that contemporary morality may be a more severe problem for singular individ-
uals. For the extent of the harm and the particular kinds of harm that contemporary
morality generates will vary depending on the characteristics of the individual in
question.
Take the disvaluation of suffering. This valuation generates an internal conflict
precisely because it inclines us not to pursue resistances and challenges. However, it
seems natural to say that some individuals suffer greater harms by not pursuing
challenges than do others. Consider an individual whose capacities are well below
average: of poor intelligence and ability, he holds down a job, and is sometimes
tempted to try to become (say) a writer, but decides that the difficulty of doing so
outweighs the benefits. This individual is never inclined to push himself to great yet
difficult accomplishments; yet, had he done so, he would not have accomplished much
anyway. If, on the other hand, a nascent Goethe or Nietzsche declines to pursue
resistances, he will be deprived of great accomplishments. As these simplistic examples
illustrate, internal conflicts between will to power and other values can generate more
and less severe problems for individuals, depending on their capacities, abilities, and
projects.
So far, nothing we have said is incompatible with universalism; the proponent of
universalism could accept these points. However, things change when we consider
external conflicts between will to power and other values. These conflicts can be far
more localized. Given facts about my psychology, my pursuing certain values might
conflict with my pursuit of power; whereas, given facts about your psychology, your
pursuing these same values might not conflict with your pursuit of power.
4
BGE 225 offers a nice summary of these points: “You want, if possible—and there is no more insane ‘if
possible’—to abolish suffering. And we? It really seems that we would rather have it higher and worse than
ever . . . The discipline of suffering, of great suffering—do you not know that only this discipline has created all
enhancements of man so far?” Analogously, HH I.235 claims that a life free from suffering “would destroy
the soil out of which great intellect and the powerful individual in general grow,” and BGE 229 contends that
“almost everything we call ‘higher culture’ rests on the spiritualization of, and giving depth to, cruelty.” See
also D 18; GS 338, 357; Z Prologue 5; BGE 202, 230, 260; GM III.11; A 2.
5
Bernard Reginster endorses a similar reading, claiming that Nietzsche offers a conception of human
flourishing that is valuable for all individuals. However, Reginster notes that Nietzsche does believe that most
individuals won’t be able to achieve this ideal: “the distinction between ‘higher’ and ‘lower’ men . . . is
interpreted more plausibly as underwriting a contrast between capacities to have a good life rather than
between different types of good life. Specifically, in contrast with those who are strong, those who are weak
are not able to overcome resistance and so to enjoy the distinctive happiness found in that activity. Nietzsche,
therefore, would not be a relativist but an ethical elitist: there is only one good life for human beings, and
some human beings are more capable of achieving it than others. This elitism, moreover, allows us to make
relatively easy sense of Nietzsche’s claim that ‘morality’ is ‘hostile to life,’ and not just to the life of the ‘higher
men.’ In being detrimental to the ‘higher men,’ ‘morality’ would simply be inimical to a good human life,
since only the higher men are capable of human excellence” (2006, 264).
A good example is self-understanding. Nietzsche has an ambivalent attitude toward

the pursuit of self-understanding. He claims that self-understanding is bad, in that it
typically leads to psychic disharmony and thereby undermines the capacity to pursue
power. But he also claims that self-understanding is good, in that it is a necessary
condition for the highest expression of will to power. The key to making sense of this
apparent inconsistency is noting that Nietzsche is claiming that self-understanding is
good for one type of person, and bad for another type of person.
Let’s start with the ways in which self-understanding can be bad. Nietzsche consist-
ently emphasizes that the pursuit of self-understanding tends to compromise psychic
integrity.6 In learning more about one’s drives, affects, and values, one may become
increasingly dissatisfied with oneself. A person who discovers truths about himself that
he cannot bear may be forced to resort to repression and self-deception, increasing his
inner conflict, magnifying his psychic disharmony. In many respects, such an individual
would be better off without self-knowledge. For the psychic disharmony will make
him less capable of encountering and overcoming resistances: it will cripple him. This is
why Nietzsche writes, “It could be useful and important for one’s activity to interpret
oneself falsely” (KSA 11:40[21]).7
But Nietzsche does not conclude that self-understanding should be avoided. He
claims that “truth has had to be fought for every step of the way,” and “almost
everything else dear to our hearts has had to be sacrificed to it”; for these reasons
“the service of truth is the hardest service” (A 50). The pursuit of truth is an exemplary
expression of will to power. Elsewhere, he writes, “How much truth does a spirit
endure, how much does it dare? More and more that became for me the real measure of
value” (EH P 3). So Nietzsche clearly values the pursuit of truth. Indeed, he makes
it clear that self-understanding is needed for revaluation of values: “to be allowed to
have a say about value and disvalue, one must see five hundred convictions beneath
oneself,—behind oneself ” (A 54). He notes that the “Revaluation of all values” is his
“formula for an act of the highest self-reflection by humanity” (EH IV.1).8
So the Nietzschean theory offers different ideals to different types of people. Put at
its baldest: self-understanding is good for those who can bear it. If an agent’s pursuit of
self-understanding will lead to internal conflict that cannot be overcome, then the
agent’s ability to encounter and overcome resistances will be undermined. Thus, an
individual who will not be able to bear truths about himself and his values has reason to
avoid self-understanding. On the other hand, an individual who will be able to bear
self-understanding has reason to pursue it. In an important passage, Nietzsche makes
this point completely explicit:
6
See especially D 18, A 50, and EH II.9 (parts of which are quoted below).
7
Compare GM Preface 1, where Nietzsche writes that “we have to misunderstand ourselves.” See also
H I.491 and the passages cited in the prior note.
8
See also D 479; GS 88, 123, 283, 301, 319, 324, 335; BGE 30, 227; GM I.17n; EH P 3, IV.3.
NIETZSCHE’S REJECTION OF UNIVERSALISM 217
Let us assume that the task [of freedom] . . . transcends the average man very significantly: in that
case, nothing could be more dangerous than catching sight of oneself with this task . . . To
become what one is, one must not have the faintest notion what one is . . . this can express a
great prudence, even the supreme prudence: where nosce te ipsum [know thyself] would be the
recipe for ruin, forgetting oneself, misunderstanding oneself . . . become reason itself. (EH II.9)
Nietzsche here considers the effect of the imperative “know thyself ” on the “average
man.” He asserts that the imperative “know thyself ” can be a recipe for ruin, whereas
forgetting oneself, misunderstanding oneself, can be eminently reasonable. In other
words, Nietzsche thinks that the pursuit of self-knowledge is dangerous for many—
perhaps most—people. Most people will not be able to bear self-knowledge. Those
who can bear self-knowledge, however, will be able to achieve the task that “tran-
scends the average man”—the task of freedom. Setting aside the question of why
freedom requires self-understanding, we can see that the will to power doctrine
generates different evaluative conclusions for different types of people.9
This brings us back to the claim that contemporary morality generates different kinds
and different extents of conflict for different individuals. A powerful individual would
be well-served by a claim such as “self-understanding is good.” After all, such a claim
would encourage her to pursue self-understanding, and this would lead her to a new
type of freedom. However, an ordinary individual would not be well-served by the
claim that self-understanding is good. If an ordinary individual attempted to pursue
self-understanding, she would be harmed. Thus, a universalistic morality preaching
that self-understanding is good for everyone would harm some individuals (the average
ones) and help others (the exceptional ones). This is, in effect, a mirror image of the
structure that Nietzsche attributes to certain Judeo-Christian values, which he takes to
harm the exceptional and help the average (cf. GM I and A).
One solution to the above problem is rejecting the idea that certain norms are
universally binding. Rather than accepting a claim such as “self-understanding ought to
be pursued,” we endorse a claim such as “self-understanding ought to be pursued by
individuals of type T, and ought not to be pursued by individuals who are not of type
T.” (Type T might be “shows evidence of having the potential to become an
exceptional individual,” or some such.)
In sum, Nietzsche’s rejection of universalism turns on the idea that external conflicts
between will to power and other values will vary across individuals, depending on their
traits and abilities. For this reason, the will to power thesis rules out universalism, and
instead entails that different values are appropriate for different types of people.10
9
Compare Z IV.13, where Nietzsche tells us not to be virtuous “beyond our strength.”
10
One could attempt to preserve universalism by contextualizing these evaluative judgments. For
example, rather than claiming that self-understanding is valuable for everyone, a universalist could claim
that self-understanding is valuable for everyone so long as its pursuit does not conflict with other values (such
as the value of psychic integrity). In a sense, this preserves the universal character of the moral judgment: the
judgment applies to everyone, but directs different individuals to act in different ways. It seems to me that
Nietzsche would have no objection to this: it merely recasts his point in different terminology.
Notice, though, that Nietzsche’s rejection of universalism does not commit him to
relativism. Relativism is typically defined as follows: the truth or falsity of normative
judgments is relative to the traditions, practices, or beliefs of a group of individuals. For
example, “cannibalism is wrong” might be true for one group of individuals and false
for another, depending on their traditions, practices, and beliefs. Nietzsche is not a
relativist, for he holds that claims such as “pursuing resistance is good” are true for all
agents. In other words, he treats will to power as a universal, non-relative standard of
success.11 Rather than a relativist, Nietzsche is what is sometimes called a parametric
universalist: he holds that there is a universally valid normative standard, but argues
that the particular results generated by this standard vary across different types of
individuals.12
With that in mind, we can summarize the above results. Nietzsche objects to a
structural feature of modern morality: its insistence that all values are universal. The
Nietzschean constitutivist theory entails that rather than treating all values as universal,
we should accept parametric universalism: while there is a universally valid normative
standard, other, more particular values can vary. That is, certain values can be justified
for one individual and not justified for another, depending on the capacities of the
individuals in question.
2. Egalitarianism
The previous section examined Nietzsche’s objection to a structural feature of modern
morality: its insistence that all values are universally applicable. Let’s now turn to some
of the particular values endorsed by modern morality. We can ask how thoroughly our
values must be modified in order to accommodate Nietzsche’s will to power thesis. Is
Nietzsche correct in suggesting that some of our most central values conflict with will
to power? I will argue that there is less reason for concern than Nietzsche seems to
think, and that the consequences of his view, while dramatic, are not as disagreeable as
they initially appear.
2.1 Nietzsche’s critique of egalitarian social institutions
I will focus upon Nietzsche’s critique of one of our most cherished values: the positive
valuation that we place on egalitarian social institutions. Most contemporary moral and
political philosophers accept egalitarianism as a starting point. Here is a typical state-
ment of the view:
11
The textual evidence for this claim was discussed in Chapters 6 and 7. See especially Chapter 6,
Section 1.1.
12
Here I am in agreement with Shaw (2007, 110). For a definition of parametric universalism, see Scanlon
(2000, 238–9). To reiterate the point from note 10: there is a sense in which Nietzsche can be described as a
universalist. He is not rejecting all forms of universalism; he is not rejecting parametric universalism. Rather, he
is rejecting the form of universalism that claims that values do not need to be contextualized to different
individuals.
E G A L I TA R I A N I S M 219
An egalitarian . . . maintains that people ought to be treated as equals—as possessing equal

fundamental worth and dignity and as equally morally considerable. In this sense, a sample
nonegalitarian would be one who believes that people born into a higher social caste, or a
favored race or ethnicity, or with an above-average stock of traits deemed desirable, ought
somehow to count for more than others in calculations that determine what morally ought to be
done. (Arneson 2009)
As Arneson notes, an egalitarian views individuals as deserving of equal treatment,

having equal worth, and meriting equal moral concern.
Nietzsche rejects each of these claims. He writes,
‘Equality’, a certain factual growing-similar which merely brings itself to expression in the
theory of ‘equal rights’, belongs essentially to decline: the cleft between human being and
human being, class and class, the multiplicity of types, the will to be oneself, to stand out—what
I call the pathos of distance, that is proper to every strong age. The tension, the range between the
extremes is today growing less and less – the extreme themselves are finally obliterated to the
point of similarity . . . All our political theories and state constitutions . . . are consequences,
necessary effects of decline . . . Declining life, the diminution of all organizing power, that is to
say the power of separating, of opening up chasms, of ranking above and below, formulates itself
in the sociology of today as the ideal . . . (TI IX.37)
Every enhancement of the type ‘man’ has so far been the work of an aristocratic society . . . that
believes in the long ladder of an order of rank and differences in value between man and man.
(BGE 257)
Aristocracy represents the belief in an elite-humanity and a higher caste. Democracy represents
the disbelief in great humans and an elite-society: ‘Everyone is equal to everyone.’ ‘At bottom
we are all together self-serving cattle and mob.’ (KSA 11:26[282])
In these passages, Nietzsche praises forms of social organization predicated upon the
idea that individuals neither have equal worth nor merit equal concern. Whereas the
idea of equality is said to foster decline, anti-egalitarian arrangements “enhance”
human beings.
Developing these claims, Nietzsche elsewhere enjoins us to promote in society a
sense of “rank order,” a sense that the great individual in society should “experience
itself not as a function (whether of the monarchy or the commonwealth) but as their
meaning and highest justification” (BGE 257). Indeed, he even suggests that in order to
promote this sort of individual, we should cultivate the attitude of the ancient nobility,
who accepted “with a good conscience the sacrifice of untold human beings who, for its
sake, must be reduced and lowered to incomplete human beings, to slaves, to instru-
ments” (BGE 258). In sum, Nietzsche claims that rather than promoting the idea that
individuals are equal, society should cultivate a sense of inequality, according to which
certain individuals are taken to have far more worth than others.
The general form of Nietzsche’s complaint against egalitarian social institutions is
clear enough: by treating individuals as equals, these institutions undermine will to
power. However, in order to assess his critique, we will need a more precise account of
the particular ways in which this is so. In the passages quoted above, Nietzsche argues
that:
(1) The idea of equality disposes us to reject any value that is not universally applicable.
(2) The idea of equality disposes us to be content with and valorize the common or
the average, rather than striving for and valorizing the exceptional.
(3) In virtue of (1) and (2), the idea of equality disposes us to become more equal.
(4) In virtue of (1)–(3), the idea of equality disposes us to think that individuals
should be equal, and that the exceptional is a defect.
Start with the first point. Nietzsche believes that the widespread acceptance of egali-
tarianism inclines us to be suspicious of values that are not universally achievable. After
all, if everyone is equal, it may seem rather odd to say that certain values are appropriate
only for a select few. Consider an example: if all agents are equal, how could the
valuation of freedom or happiness be justified for some agents but unjustified for others?
This claim becomes more plausible when we examine it in connection with (2).
Claim (2) asserts that the valuation of equality has an effect on our other values: it leads
us to value the average, rather than the exceptional. The easiest way to illustrate this
point is by considering the way in which the content of one of our central values—
happiness—has shifted over time.
To begin, consider Aristotle’s account of happiness. Aristotle spends most of the
Nicomachean Ethics arguing that the forms of happiness sought and acquired by most
agents are illusions. Genuine happiness is “the activity of philosophical wisdom,”
which is “admittedly the pleasantest of virtuous activities; at all events the pursuit of
it is thought to offer pleasures marvelous for their purity and their enduringness”
(Nicomachean Ethics 1177a). Accordingly, happiness is available only to the few who
are capable of philosophical activity. Moreover, this happiness is available only for
limited stretches of time. Aristotle tells us that a completely contemplative life “would
be too high for man” (1177b). It would be a life fit for the divine. Nevertheless,
Aristotle urges us to strive to achieve this life: “But we must not follow those who
advise us, being men, to think of human things, and, being mortal, of mortal things, but
must, so far as we can, make ourselves immortal, and strain every nerve to live in
accordance with the best thing in us; for even if it be small in bulk, much more does it
in power and worth surpass everything” (1177b–1178a).
On Aristotle’s account, then, genuine happiness is achieved in philosophical con-
templation: it is available only to a very few, and the most that these few can hope for is
to achieve it in fleeting moments. Contrast this with Bentham’s conception of happi-
ness. Bentham tells us that happiness is the having of pleasurable sensations (Bentham
2007, Chapters I and IV). Anyone—indeed, anything sentient—can have that. Thus, in
moving from Aristotle to Bentham, happiness is reconceived such that everyone is
capable of achieving it: no one is in principle barred from having pleasurable sensations.
Indeed, not only is everyone capable of achieving pleasure, but everyone does, to some
extent, achieve it.
So happiness is transmuted from a rare state achievable only in fleeting moments by

the very few, to something universally available and readily achieved. This illustrates
the kind of diminishment of our values that Nietzsche believes is fostered by egalitar-
ianism. As he puts it, “the instinct of the herd esteems the middle and the mean as the
highest and most valuable” (KSA 12:10[39]).13
This brings us to (3): in part by treating all values as universally achievable and
valorizing the universally available middle, the valuation of egalitarianism causes
individuals to become more equal:
One calls that in which the distinction of the European is sought ‘civilization’ or ‘humanization’
or ‘progress’; one calls it simply . . . Europe’s democratic movement . . . Behind all the moral and
political foregrounds, to which some formulas point, a tremendous physiological process is taking
place, which comes ever more on,—the process of the becoming-similar of Europeans. (BGE
242; cf. Z II.7)
In other words, Nietzsche complains that the positive valuation of egalitarianism has
the effect of incrementally rendering people equal in ability. It pulls individuals down
to the lowest common denominator. After all, if our ideal is some universally achiev-
able quality such as pleasure, we will not strive to encounter and overcome resistance.
This harms the potentially exceptional individuals, tempting them to rest content with
the average. However, it also harms the average individual, by disinclining him from
engaging in any struggles whatsoever. Thus, Nietzsche worries that we are headed
toward the “most contemptible thing”: the emergence of the “last man,” the individ-
ual whose ideal is nothing more than the “wretched contentment” of having pleasur-
able sensations (Z Preface).
The apotheosis of this process, Nietzsche believes, is the idea that individuals should
be equal in capacities. This is claim (4): Nietzsche contends that the valuation of
13
John Richardson offers a helpful gloss: “The ideal lies at a merely average level, and if one has trouble
even so, society helps one up to it. Thus there’s very little gap between the life conceived to be best and the
one attained by mere default and inertia” (Richardson 1996, 168). So, the ideal becomes something readily
achievable (such as having pleasurable sensations), and society is organized such that all achieve it. For an
insightful discussion of the general point, see Clark (1999). In a passage worth quoting at length, she writes
that in egalitarian societies, “the desire to achieve, to be a winner, is very strong; it is just that the measure of
winning, of achieving excellence, is common, is reduced to increments of the lowest common denominator,
of what everyone can relate to without having to change any fundamental aspect of their character. What is
missing is any room for the idea that there are higher states of soul, virtues or excellences of character, ones
that are not mere increments of that to which everyone can already relate. It was not always this way. Under
the influence of religion, human beings have traditionally believed in ‘higher human beings,’ in human
beings who have achieved a level of spirituality or virtue of which most people are not capable. In the
discourse of our now more democratic culture, on the other hand, virtue has only a minor presence. And
there is certainly nothing in this discourse to challenge and stir the soul, to induce a craving for higher virtues
or degrees of virtue. This is, I think Nietzsche would say, because the only virtues about which we talk are the
old, tired virtues, virtues interpreted under the auspices of the old ideal, hence ones of which we assume
everyone is already capable. But ‘what can be common,’ says Nietzsche, ‘always has little value’ (BGE 43)—
which is to say that what we perceive as ordinary cannot inspire us to extraordinary passion or effort. And
without such passion and effort, Nietzsche very plausibly claims, no enhancement of the human type, no new
level of spiritual achievement, will be possible . . . ” (Clark 1999, 132).
egalitarianism will lead us to treat exceptional individuals as defective. In an extended

reflection on this point, Nietzsche writes,
The more the feeling of unity with one’s fellow humans gains the upper hand, the more human
beings are made uniform, the more they will perceive all difference as immoral. In this way,
the sand of humanity necessarily comes into being: all very similar, very small, very round,
very accommodating, very boring. Christianity and democracy have done the most to drive
humanity along the path toward sand. A small, weak, glowing feeling of contentment equally
distributed among all . . . would that be the last image that humanity could offer? Inevitably, if
we remain on the path of moral sensibilities until now. A great reflection is needed, perhaps
humanity must draw a line under its past, perhaps it must address a new canon to all singular
individuals: be different from all others, and take pleasure in each being different from the
other; the crudest monsters have certainly been eradicated under the prevailing regime of
morality thus far—that was its task; but we do not wish to live on thoughtlessly under a regime
of fear in the face of wild beasts. For so long, far too long, the word has been: one like all, one
for all. (KSA 9:3[98])
In this passage, Nietzsche claims that egalitarian social institutions lead to the thought
that individuals should be uniform and that difference as such is immoral. The
exceptional individual, who sets herself difficult tasks, is viewed as defective or
objectionable. (It may be helpful to illustrate this contention with a simple example:
consider the 2008 US presidential election, in which traits such as education at elite
universities and high degrees of intelligence were sometimes presented as objection-
able, whereas middling accomplishments were sometimes presented as exemplary.14)
In sum, Nietzsche’s concern is that a culture that fosters a sense of equality will not
promote the achievement of human excellence. Instead, it will incline individuals to
view mediocrity as perfectly acceptable, and distinction as at least vaguely objectionable.
2.2 Nietzsche’s focus on moral culture rather than moral norms
There is an obvious and straightforward response to Nietzsche’s concerns: we can
simply modify our conception of egalitarianism to avoid these problematic conse-
quences. After all, (1)–(4) are not logical consequences of the claim that individuals
deserve equal rights or equal moral standing. As a result, it seems that we could simply
pry apart the claim that individuals deserve equal rights from the claim that all
individuals ought to be equal in capacities and achievements (claim 4), that the average
is the ideal (claim 2), and so on.
This is, after all, the strategy that many contemporary egalitarians pursue. Consider
first a strong version of egalitarianism:
Egalitarians have the deep and (for them) compelling view that it is a bad thing—unjust and
unfair—for some to be worse off than others through no fault of their own. (Temkin 1986, 100)
14
See, for example, Bigg (2008) and Dean (2008). For a classic study, see Hofstadter’s Anti-Intellectualism in
American Life (1966).
Temkin contends that the mere fact that individuals are worse off due to factors beyond
their control is immoral. In so doing, he moves, perhaps quite naturally, beyond the
basic egalitarian idea that individuals deserve equal treatment or equal negative rights to
the much stronger idea that individuals deserve equal outcomes.
However, it hardly seems that all egalitarians are committed to Temkin’s strong
form of egalitarianism. A number of philosophers explicitly reject the idea that
egalitarianism requires equal outcomes. For example, Dworkin (1977) and Nagel
(1991) distinguish equal rights or equal treatment from equal outcomes, and argue
that egalitarianism requires only the former.
This brings us to our objection: Nietzsche seems to elide the differences between
extreme and more restricted versions of egalitarianism. His attacks focus on aspects
of an extreme form of egalitarianism, which few contemporary philosophers en-
dorse. Once we distinguish a restricted form of egalitarianism—say, a form that
requires only equal treatment rather than equal outcomes—won’t it be immune to
Nietzsche’s critiques? Put differently, wouldn’t this restricted form of egalitarianism
avoid problems (1)–(4)?
The ease with which we can draw these distinctions between extreme and restricted
egalitarianism should make us suspect that Nietzsche has a response to this strategy. And
indeed, Brian Leiter has shown that he does. Leiter writes that although egalitarianism
doesn’t explicitly require equal outcomes, valorization of the mean, and so forth, a
society in which egalitarianism is valued and promoted does tend to lead to these
outcomes. In particular, Leiter writes that Nietzsche “thinks a culture in which such
norms prevail as morality will be a culture which eliminates the conditions for the realization
of human excellence—the latter requiring, on Nietzsche’s view, concern with the self,
suffering, a certain stoic indifference, a sense of hierarchy and difference, and the like”
(Leiter 2010, Section 1.3). He explains,
The normative component of an MPS [Leiter uses MPS to abbreviate “morality in a pejorative
sense,” or modern morality] is harmful not because its specific prescriptions and proscriptions
explicitly require potentially excellent persons to forego that which allows them to flourish—
that is, Nietzsche’s claim is not that a conscientious application of the “theory” of MPS would be
incompatible with the flourishing of higher men. Rather, Nietzsche’s claim is that an MPS in
practice simply does not make such fine distinctions: under a regime of MPS values—and
importantly because of MPS’s embrace of the idea that one morality is appropriate for all—
potentially higher men will come to adopt such values as applicable to themselves as well. Thus,
the normative component of MPS is harmful because, in reality, it will have the effect of leading
potentially excellent persons to value what is in fact not conducive to their flourishing and
devalue what is in fact essential to it. (Leiter 1997, 274)
As Leiter notes, Nietzsche’s claim is not that a society that values equality (for example)
will undermine excellence and threaten achievement because it explicitly disavows
such things. Rather, the valuation of equality will undermine excellence for two
reasons. First, the valuation fosters a culture in which there is pressure not to value
excellence and achievement. Second, the valuation of equality inclines us to value

things that diminish struggle, such as uniformity and conformity.15 Leiter provides an
illustration: “rather than learn how to look down on himself, to desire to overcome his
present self and become something better, he will embrace the prevailing rhetoric of
equality—captured nicely in the pop-psychology slogan ‘I’m OK, you’re OK’—and
thus never learn to feel the contempt for self that might lead one to strive for something
more” (Leiter 1997, 281).
Herman Siemens makes a similar point, writing
The political is usually taken as symptomatic of something else, something much larger, much
broader, that Nietzsche comes to call the ‘whole democratic movement’ [demokratische Ge-
samtbewegung]’ (KSA 11:26[352], p. 242). Perhaps we can speak of a pervasive cultural
tendency or a general disposition. This is also suggested by the frequent occurrence of the
term ‘taste’ (Geschmack) in connection with democracy in expressions such as ‘der demokra-
tische Geschmack’ or ‘der demokratische Grundgeschmack aller Wertschätzung’ (KSA 11:35
[22], p. 518). ‘Democracy,’ then, refers primarily to a set of values or ideals . . . but also to
a disposition, attitude, or type that flourishes and dominates under those values. (Siemens
2009, 21)
It is not necessarily the explicit evaluative claims that undermine will to power and lead
to (1)–(4), then. It may, instead, be the “disposition, attitude, or type” that is promoted
and fostered by these evaluative claims. Leiter and Siemens say little about the actual
mechanisms by which this phenomenon occurs. However, a helpful analogy may be
certain feminist critiques of pornography: while there may be nothing about the
specific normative claim “pornography is acceptable” that undermines the flourishing
of women, some writers argue that the portrayal of sexuality leads to a system of largely
tacit assumptions, attitudes, and dispositions that do undermine the flourishing of
women (see, for example, Easton 1994 and Dyzenhaus 1992).
In sum, Nietzsche’s claim is that egalitarian societies—even if they explicitly endorse
only a restricted form of egalitarianism—foster dispositions, attitudes, and cultural
norms that generate conflicts with will to power. Thus, Nietzsche’s complaint against
the valuation of egalitarianism focuses on the moral culture induced by this valuation,
rather than the moral norm itself.
2.3 Does Nietzsche endorse non-egalitarian institutions?
I have presented Nietzsche’s critique of egalitarianism. Let’s now examine the other
side of that critique: the political arrangements that Nietzsche endorses.
Readers sometimes assume that if Nietzsche critiques egalitarianism, he must embrace
non-egalitarian views. This is facile: one can—and Nietzsche often does—critique a
15
As David Owen points out, “Nietzsche’s criticism accords with that general nineteenth-century
concern articulated by perfectionist thinkers—most famously, John Stuart Mill—that the democratic move-
ment expresses a will to equality that acts to level down and, more particularly, to foster conformity” (Owen
2002, 119).
value or practice without endorsing its opposite. For example, Nietzsche critiques certain
aspects of the scientific enterprise, but hardly takes this to entail that we should reject it;
he critiques the university, without thinking it should be abolished; and he critiques the
will to truth, while himself seeking truth.16 Just so, one can critique certain features of
egalitarianism while nonetheless thinking that egalitarianism is, all things considered,
something we ought to adopt. Thus, David Owen is exactly right in noting that “we
should not leap from the fact that Nietzsche offers pointed criticisms of the democratic
movement (e.g., BGE 202–3) to the conclusion that Nietzsche is an anti-democratic
thinker” (Owen 2002, 120).
With that in mind, let’s take a closer look at why Nietzsche thinks non-egalitarian
systems avoid problems (1)–(4), above. When Nietzsche praises non-egalitarian
systems, he typically focuses on the idea that these systems generate a “pathos of
distance.”
Every elevation of the type “man,” has hitherto been the work of an aristocratic society—and so
it will always be—a society believing in a long scale of gradations of rank and differences of worth
among human beings, and requiring slavery in some form or other. Without the pathos of distance
such as grows out of the incarnated difference of classes, out of the constant looking out and
looking down of the ruling caste on subordinates and instruments, and out of their equally
constant practice of obeying and commanding, of keeping down and keeping at a distance—that
other more mysterious pathos could never have arisen, the longing for an ever new widening of
distance within the soul itself, the formation of ever higher, rarer, further, more extended, more
comprehensive states, in short, just the elevation of the type “man,” the continued “self-
surmounting of man,” to use a moral formula in a supermoral sense. (BGE 257; compare TI
IX.37, quoted above)
A “pathos of distance” is a sense that human beings are fundamentally unequal: that
some human beings are more valuable than others.17 In the above passage, Nietzsche
claims that this sense of evaluative distance between human beings in aristocratic
societies led to heightened manifestations of will to power: it disposed certain individ-
uals to strive for “higher, rarer” states and to “surmount” themselves.
Non-egalitarian systems foster this pathos of distance by “incarnating” it in differing
legal statuses, rights, expectations, and so forth. By contrast, egalitarian systems elimin-
ate this pathos. A line from Whitman’s poem “The Prairie-Grass Dividing” illustrates
this point: Whitman speaks of Americans as “those that look carelessly in the faces of
Presidents and governors, as to say Who are you?” This unconcern with rank, this
disposition to view all individuals as equal, this perception of even those in the highest
positions of authority as fundamentally the same as oneself—this is the opposite of the
pathos of distance.
16
For critiques of the scientific enterprise, see BGE 211, BGE 253, BGE 270, and GM III.23–7; for the
university, see BT 20, UM III, and BGE 58; for the will to truth, see GM III.23–7, A 50, and EH P 3.
17
For additional remarks on the pathos of distance, see GM I.2, GM III.14, A 43, A 57.
Nietzsche’s remarks on the pathos of distance suggest an argument of the following

form:
(i) The pathos of distance is promoted by non-egalitarian values, and undermined
or eliminated by egalitarian values.
(ii) The pathos of distance is a necessary condition for maximal fulfillment of the
will to power.
(iii) If a value conflicts with a necessary condition for maximal fulfillment of will to
power, then this value is to be rejected.
(iv) Therefore, we ought to reject egalitarian values, and adopt non-egalitarian values.
Would this argument be convincing? In fact, there are several problems.
First, claim (ii) seems excessively strong. Even if the pathos of distance were a spur to
achievement, it is hard to believe that it is a necessary condition for maximal fulfillment
of the will to power. After all, there are many individuals who pursue and overcome
immense challenges within egalitarian societies. Whitman himself is perhaps an
example of a great individual who flourished in a (relatively) egalitarian society.
Indeed, Nietzsche himself seems to acknowledge this point, writing that although
“Europe’s democratic movement . . . will on average lead to the leveling and mediocri-
tization of man,” it is also “likely in the highest degree to give birth to exceptional
human beings of the most dangerous and attractive quality” (BGE 242). He goes on to
claim that the emergence of democracy in Europe has the result that “in single,
exceptional cases the strong human being will have to turn out stronger and richer
than perhaps ever before.”
For these reasons, claim (ii) seems implausible. A weaker version of (ii) might be true:
for example, the pathos of distance might promote or encourage maximal fulfillment of
will to power. However, we would then need to modify claim (iii) as follows: if a value
fails to promote maximal fulfillment of will to power, then it should be rejected. Might
this modified argument succeed?
In fact, the argument would then become far more dubious. Suppose it is true that
non-egalitarian arrangements promote a pathos of distance, and thereby dispose
individuals to fulfill the will to power maximally. Still, we would have to weigh this
against the ways in which non-egalitarian systems undermine will to power. There are
glaringly obvious problems, here. I will mention just two.
First, non-egalitarian arrangements systematically exclude many individuals from the
very possibility of achievement. Consider the actual forms of non-egalitarian societies
that have existed. Whether we consider the aristocratic societies of the middle ages, the
restricted forms of democracy present more recently in American and European
history, or many of the contemporary societies in the Middle East, they are non-
egalitarian in virtue of excluding certain groups from positions of authority: such
groups might include those not descended from royal blood, the poor, those who do
not own land, women, racial or ethnic minorities, or certain religious groups. Perhaps
this generates a “pathos of distance”: perhaps early nineteenth-century white male
Americans felt themselves superior to their black slaves and to women. Perhaps, as
Nietzsche’s claims imply, this induced certain white males to strive for maximal
fulfillment of will to power. However, this effect would have to be balanced against
the fact that the social arrangement systematically prevented the emergence of talented,
exceptional individuals from whole swathes of the non-white, non-male population.
Put bluntly: one doubts that the American slave trade was especially beneficial in
promoting the emergence of singular black individuals, or that the repression of
women increased their chances for achieving greatness.18 In short, even if we grant
that egalitarianism in certain ways hinders the emergence of singular creative individ-
uals, we have to ask whether a non-egalitarian society would do any better.
A second problem is that in non-egalitarian societies, this pathos is directed at those
who occupy positions of authority solely by way of birthright, race, gender, or other
factors beyond their control. One wonders why this is supposed to be better than a
society in which authority and respect have to be earned. After all, in egalitarian
societies there are certainly individuals who are regarded as (e.g.) artistic, literary,
intellectual, or athletic exemplars. David Owen draws attention to a version of this
point, arguing that we can generate a pathos of distance without there being “a fixed
social hierarchy,” because
all that is required for an order of rank [and hence a pathos of distance] to exist is that there is a
common acknowledgement of a range of evaluative distinctions between the virtuous man and
the vicious man or, say, the strong and the weak, and that this acknowledgement is given
practical expression in the distribution of respect and contempt within the ethical culture. (Owen
2008, 163)
In other words, while the hierarchical social arrangements instituted by non-egalitarian

societies are one way of fostering a pathos of distance, a non-hierarchical society could
generate the same pathos by drawing and enforcing strong evaluative distinctions that
are not tethered to social rank. For example, a society could distinguish and valorize one
sort of person—the artist, the athlete, the soldier, the doctor, the virtuous individual—
and thereby induce in individuals a sense of rank.
I mention these problems to illustrate a significant difficulty with Nietzsche’s
critique of egalitarianism: although Nietzsche’s ethical theory does entail that certain
aspects of egalitarian social institutions are problematic, it also entails that aspects of non-
egalitarian institutions are problematic. With this in mind, let’s step back and evaluate
Nietzsche’s critique of egalitarianism, asking what consequences follow from it.
18
Of course, one does think of certain great individuals who emerged from these oppressed groups:
Martin Luther King, Frederick Douglass, Emily Dickinson, and so on. However, the relevant question is
not whether any great individuals emerged from among the oppressed groups; it is, instead, whether more
great individuals would have emerged absent the oppression. (Nor is the question whether, for example,
Martin Luther King would have been a great individual if he had emerged in an egalitarian society. There is
no reason to assume that the same individuals who flourished in oppression would flourish absent the
oppression.)
2.4 Evaluating Nietzsche’s critique of egalitarianism

I have now presented Nietzsche’s analysis of the way in which egalitarianism conflicts
with will to power. In this section, I will evaluate the critique, asking how convincing
it is.
Nietzsche sometimes does suggest that the will to power thesis issues a damning
indictment of egalitarianism, and that it instead favors non-egalitarian societies. But it is
important to note that, put bluntly, he may be mistaken. After all, philosophers are
often confused about the consequences of their own theories. Take Kant: he claims
that his moral theory entails that masturbation is worse than suicide, that suicide is
always impermissible, that homosexuality is a perversion, and so on. Few Kantians
believe that Kant’s theory actually entails these results; the prejudices of Kant’s time
simply led him astray. Just so, Nietzsche may be mistaken about what his will to power
thesis entails.19
With this in mind, let’s ask how persuasive Nietzsche’s arguments are. Do his
arguments show that our positive valuation of egalitarianism should be abandoned?
In order to pose this question in its strongest form, let’s simply grant that all of
Nietzsche’s arguments, above, succeed: egalitarianism has all of the disagreeable,
disvaluable features that Nietzsche attributes to it. Nevertheless, we still have to ask
whether egalitarianism might have positive features that outweigh these negative
attributes. After all, Nietzsche would be the first to admit that nothing in the world
is likely to be wholly positive or wholly negative.
Notice that in order to value something, we need not (and typically will not) regard
it as an unalloyed good. I might value a relationship, a job, or a political party while
knowing that it has a host of disagreeable, disadvantageous, or downright bad aspects.
For example, imagine an individual who considers whether he can continue valuing his
relationship with an absent, dysfunctional, troubled parent. The question that this
individual poses isn’t whether he finds the relationship wholly positive, for he obvi-
ously doesn’t. The question is whether, while recognizing all that is bad about the
relationship, he can continue to value it. Sometimes, one can.20
So the question concerning egalitarianism is whether, while holding its pernicious
aspects clearly in view, we can continue to affirm it. Above, I pointed out that nothing
in the Nietzschean theory directly conflicts with restricted egalitarianism (the claim that
individuals ought to enjoy equal standing). Nietzsche’s arguments instead focus on the
alleged consequences of instituting restricted egalitarianism:
19
It also bears keeping in mind that Nietzsche was writing in an age of ideological politics, when
philosophers and ideologues were openly debating the respective merits of fascism, communism, capitalism,
and the like. His suggestions were, in a way, less radical in the past than they are today.
20
One thinks, here, of Nietzsche’s eternal recurrence test. As Janaway points out, the relevant question is
not whether each event in one’s life was wholly good; the question is, instead, whether one can see all that is
wretched about one’s life and the world and nonetheless continue to value it (Janaway 2007, 257–60).
A. Egalitarian societies foster attitudes, dispositions, and cultural norms that generate
conflicts with will to power (see (1)–(4), above).
B. Egalitarian societies eliminate the pathos of distance.
To assess Nietzsche’s critique, we should ask how convincing these claims are.
I have already pointed to some problems with (B). First, Nietzsche’s claim that the
pathos of distance is a spur to achievement is somewhat dubious, and the strong claim
that it is a necessary condition for achievement is implausible. Second, there seem to be
ways of fostering a pathos of distance in egalitarian societies. Third, even if the first and
second responses were untenable, we would need to determine which is worse: a non-
egalitarian society with a pathos of distance, or an egalitarian society without a pathos
of distance. I have suggested, above, that even on Nietzsche’s theory the latter society
might well come out ahead.
In sum, claim (B) seems both dubious and incapable of generating, on its own, an
indictment of egalitarianism. But this does leave us with (A). What can we say in
response? I will suggest that certain forms of restricted egalitarianism can be maintained,
without fostering the attitudes that concern Nietzsche.
David Owen has recently argued that we should respond to Nietzsche’s critique not
by abandoning our commitment to egalitarianism, but by distinguishing different forms
that egalitarian (or democratic) institutions might take. Some of these forms will
mitigate the negative consequences adumbrated by Nietzsche, whereas other forms
will heighten them. In support of this point, Owen quotes the following passage:
Two kinds of equality.—The thirst for equality can express itself either as a desire to draw everyone
down to oneself (though diminishing them, spying on them, tripping them up) or to raise oneself
and everyone else up (through recognizing their virtues, helping them, rejoicing in their success).
(HH 300)
Nietzsche seems to believe that modern calls for egalitarianism promote the first kind of
equality. However, Owen argues that agonistic elements can be incorporated into
egalitarian social institutions, so that they promote the second kind of equality:
Modern democracy can avoid the pitfalls that Nietzsche identifies in ‘the democratic movement of
our times’ to the extent that it cultivates an agonal political culture in which citizens strive to develop
their capacities for self-rule in competition with one another, a culture that honors exemplary
democratic citizens as setting standards that we should seek to match and surpass. (Owen 2002, 126)
That is, Owen argues that egalitarian social institutions needn’t create a culture of
leveling-down, but could instead foster increased competition amongst citizens. An
example would be ancient Athens, which, Nietzsche claims, had an agonistic democ-
racy (see HC).21
21
Owen’s reading is complicated by the fact that Nietzsche himself seems to view the ancient Athenian
agon as restricted in its domain: the aristocratic elite were engaged in various forms of competition, but it is far
from clear that this attitude was widespread among the slave class. This is a point made by Appel (1998, 140).
For another reading along Owen’s lines, see Hatab (1995).
With this in mind, it seems that we can sketch the general shape of a society that
would fulfill the Nietzschean demands without departing too radically from our own
current ideals. Consider a hypothetical society with the following three features:
(1) Rights are evenly distributed. For example, no one should enjoy different legal
rights on the basis of gender, race, sexual orientation, and so forth. Following
Dworkin (1977, 370), we can say that this is a demand not that people be treated
equally, but that they be treated as equals. Treating people as equals would
include, for example, granting everyone the freedom of speech, the right to
vote, and equality of opportunity.
(2) A “pathos of distance” or an idea of “spiritual inequality” is fostered. That is, the
society promotes the idea that some lives are better or more valuable than others.
Most societies already do this in restricted domains: some athletes are regarded as
far superior, qua athletes, to other athletes; some literary figures are regarded as
superior, qua writers, to others writers. However, the hypothetical society
would promote an unrestricted version of this idea, by maintaining that certain
individuals should be regarded as better human beings than others. In other words,
individuals are not just better at what they do, but are better at what they are.22
(3) The society is structured so that great accomplishments are fostered. That is, the
society provides support and encouragement for these accomplishments. Great
achievements are valorized, and institutions are designed to support these
achievements. For example, if one wanted to promote the emergence of great
literary figures, the government could provide subsidies for writers, and could
attempt to promote widespread respect and admiration for these writers. If one
wanted to promote great scientific endeavors, one could devote more of
society’s resources to basic research and education.
Condition (1) makes the society egalitarian; conditions (2) and (3) are attempts to
mitigate Nietzsche’s concerns about the damaging attitudes fostered by typical egali-
tarian systems, as discussed in Section 2.2.
Contemporary societies do not, in general, meet these conditions, but neither are
these conditions so outlandish as to defy belief. One could imagine a society of this
sort.23,24 It would seem to meet the Nietzschean conditions reasonably well, while
22
Compare Richardson: “Nietzsche wants to redesign society, so that it functions not for average or
overall well-being, but for high individual achievements . . . One of the ways society should favor such
achievements is by dispersing widely the recognition or acceptance that there are higher and lower human
levels—that individual types lie not on a plain of equal validity or worth, but vertically on a ‘ladder’ of steps or
levels” (Richardson 2004, 204). Notice that this would not be easy. It would require that many individuals—
those who lack the capacities for great accomplishments—view themselves as bad or wretched, and perhaps
irredeemably so. It would thus run the risk of sparking ressentiment, in exactly the way that Nietzsche outlines
in the first essay of the Genealogy.
23
Notice, though, that it might not be possible democratically to institute such a society. Tamsin Shaw
discusses a version of this problem in Chapter 6 of her 2007.
24
Indeed, one might even go further. Many philosophers believe that society should be designed to
promote everyone’s well-being. While Nietzsche would reject certain versions of this idea, it is harder to see
what his objection would be to other forms. For example, Harry Frankfurt argues that inequality as such is
NIETZSCHE’S ETHICAL EXEMPLARS 231
avoiding some of the negative consequences of both extreme egalitarianism and non-
egalitarianism.
2.5 Summary
Above, I have examined Nietzsche’s critique of one of our most cherished values:
egalitarianism. I have argued that the will to power thesis does not entail that egalitar-
ianism should be rejected. The Nietzschean view is far more nuanced. Nietzsche thinks
egalitarianism runs the risk of promoting certain disadvantageous tendencies, and he
suggests that, in response, we need to reinforce in human beings the tendency to strive
for distinction and greatness. He thinks this is best promoted by fostering, in human
beings, the idea that some lives are worth far more than others: that a Goethe or a
Nietzsche is worth more than an average life. Thus, Nietzsche enjoins us to promote in
society a sense of “rank order,” a sense that the great individual in society “experience
itself not as a function (whether of the monarchy or the commonwealth) but as their
meaning and highest justification” (BGE 257). Indeed, he even suggests that in order to
promote these sorts of individuals we should cultivate the attitude of the ancient
nobility, who accepted “with a good conscience the sacrifice of untold human beings
who, for its sake, must be reduced and lowered to incomplete human beings, to slaves,
to instruments” (BGE 258). Whatever we may think of that, we can agree that
Nietzsche has identified some potential problems with egalitarianism. Egalitarian social
institutions are not unalloyed goods; they have disadvantageous features. But the fact
that egalitarian institutions are not wholly good does not in the least imply that non-
egalitarian institutions would be better; nor does it imply that we should abandon the
promotion of egalitarianism.
Thus, examining Nietzsche’s critique of egalitarianism reveals that although a
Nietzschean ethical theory doesn’t leave everything as it is, it also doesn’t generate
appalling or abhorrent consequences. The Nietzschean theory identifies some prob-
lematic aspects of egalitarianism, and urges us to modify our valuation in order to
resolve, as best as possible, these problems.
3. Nietzsche’s ethical exemplars

In the previous section, I discussed some aspects of egalitarianism that do not fare well
on the Nietzschean theory. I have argued that while the Nietzschean theory rules out
extreme versions of egalitarianism, it is compatible with more restricted forms, and it
does not require that we adopt non-egalitarian policies.
unimportant: it doesn’t matter whether some are worse off than others (Frankfurt 1987). Certainly Nietzsche
would agree with this. However, Frankfurt goes on to argue that it does matter whether the worst off have a
decent life, passing some minimal threshold. Once this threshold is reached, massive inequality is fine. While
Nietzsche himself would seem to disavow such an idea, it is harder to see what, in his ethical theory, really
opposes it: it is not as if a nascent Nietzsche or Goethe, emerging from the lowest strata of society, would have
been spurred to greatness by starving to death, or suffering from lack of basic medical care.
Let’s turn now to the positive side of the Nietzschean ethical theory, examining the
way in which it commits us to certain values. Again, it will be helpful to begin by
examining Nietzsche’s texts, considering the individuals whom Nietzsche presents as
ethical exemplars. The individual whom Nietzsche presents most frequently as the
embodiment of his ethical ideal is Goethe.25 Other individuals who are presented in
a uniformly positive fashion are Nietzsche himself, Beethoven, and Napoleon.26
Although this is a diverse group, Nietzsche focuses on the several traits that they
have in common: their commitment to immensely difficult goals; their continuous
seeking of new attainments; their commitment to truthfulness; their self-affirmation;
and their “untimeliness” or “independence” from their age. I will discuss each of these
characteristics in turn.27
First, these individuals strove to accomplish immensely difficult tasks. They encoun-
tered and overcame extreme amounts of resistance. Although countless individuals
have tried to produce works on a similar scale, very few have succeeded; there have
been many novelists, many musicians, and many political and military leaders, but few
are the equal of Goethe, Beethoven, and Napoleon. But it is not just the objective
magnitude of their accomplishments that Nietzsche praises: in addition, he consistently
emphasizes that these individuals manifest a severity with themselves, including the
self-discipline necessary to pursue great goals. He writes,
The noble human being honors himself as one who is powerful, also as one who has power over
himself, who knows how to speak and be silent, who delights in being severe and hard with
himself and respects all severity and hardness. (BGE 260)
Goethe’s sustained devotion to Faust, Beethoven’s struggles after his loss of hearing,
and Napoleon’s series of victories come to mind.
Second, Nietzsche focuses upon the way in which these individuals continuously
sought new achievements. For example, Goethe achieved immense fame with his early
publication of The Sorrows of Young Werther. But he did not rest content with that: he
25
In Nietzsche’s texts, Goethe is mentioned more than any other individual. There are 135 discussions of
Goethe in Nietzsche’s texts, and—quite unusually for Nietzsche—each passage is strikingly positive. See
Brobjer (1995, Appendix 2).
26
Beethoven is mentioned 27 times, whereas Napoleon is mentioned 26 times (Brobjer 1995, Appendix 2).
For Nietzsche as the embodiment of his own ethical ideal, see Ecce Homo, especially part IV.
27
See also Leiter (2002, 115–25) for a helpful discussion of these individuals. Although Leiter offers a
somewhat different list of the characteristics definitive of Nietzsche’s exemplars, I am largely in agreement
with his reading. In particular, Leiter takes the following features as definitive of Nietzsche’s “higher man”:
(1) he “seeks burdens and responsibilities, as he is driven towards the completion of a unifying project” (2002,
117); (2) he has a form of self-reverence (2002, 120); (3) he is solitary (2002, 116); (4) he “is essentially healthy
and resilient” (2002, 118); and (5) he affirms life (2002, 119). My discussions of difficult goals and continuous
seeking of new attainments correspond to Leiter’s (1); my discussion of self-affirmation corresponds to
Leiter’s (2); and my discussion of independence corresponds to Leiter’s (3). I do not discuss Leiter’s (4), as
I take this to be equivalent to the claim that the individual manifests a high degree of power (see Chapter 6);
and I omit (5), as I have discussed the relevant material above in Chapter 7, Section 4. Leiter omits discussion
of the way in which Nietzsche’s exemplars manifest a commitment to truthfulness; as I explain below, I take
this to be an important and pervasive concern.
went on to labor over Faust and other works for the remainder of his life. Nietzsche
describes this as a form of self-overcoming: whenever one completes some difficult
task, one does not pause and experience contentment with one’s current level of
attainment. Rather, one immediately sets oneself a new, more difficult task. Or, as
Nietzsche puts it, “Whatever I create and however much I love it—soon I must oppose
it and my love—thus my will wills it” (Z II.12).
A third trait that Nietzsche focuses upon is self-affirmation. He writes that “the
noble soul has reverence for itself ” (BGE 287). Thus, he valorizes “the ideal of the
most high-spirited, alive, and world-affirming human being who has not only come to
terms and learned to get along with whatever was and is, but who wants to have what
was and is repeated into all eternity” (BGE 56). These individuals, Nietzsche tells us,
are entirely content with their lives—they affirm their lives despite the many difficulties
and struggles that they encounter. They are free from ressentiment and associated
emotions such as guilt (EH II.2; TI IX.49).28
Fourth, these individuals manifest a certain form of honesty or truthfulness. Nietzsche
writes,
What is good-heartedness, refinement, and genius to me, when the human being who has these
virtues tolerates slack feelings in his faith and judgments, and when the demand for certainty is
not to him the inmost craving and the deepest need—that which distinguishes the higher from
the lower men. Among certain pious ones, I found a hatred of reason and appreciated it: at least
they thus betrayed their bad intellectual conscience. (GS 2)
Similar remarks occur in the Antichrist, where Nietzsche claims that he “wages war”
against contemporary morality and religion, which, he tells us, are so distorted that they
lack “even a single point of contact with reality” (A 9, A 15). He valorizes those who
seek truth: “At every step one has to wrestle for truth . . . That requires greatness of soul:
the service of truth is the hardest service . . . ” (A 50). Nietzsche tells us that his character
Zarathustra’s “doctrine, and his alone, posits truthfulness as the highest virtue” (EH
IV.3). He summarizes his view as follows: “How much truth does a spirit endure, how
much truth does it dare? More and more that became for me the genuine measure of
value” (EH Preface 3).
This brings us to the final point: Nietzsche presents his exemplars as “untimely,” in
the sense that they achieved a form of independence from their age (cf. BGE 211–12).
They did not merely adopt the dominant values and established culture, but attempted
to transform them. (This is, of course, closely related to their honesty: they can see
through defunct ideals.) Nietzsche praises this independence: “the concept of greatness
entails being noble, wanting to be by oneself, being able to be different, standing alone
and having to live independently” (BGE 212). As he elsewhere puts it, “Independence
28
Risse (2003) provides a helpful discussion of the way in which these traits comport with Nietzsche’s
broader concerns about morality and freedom. As Risse points out, “guilt and ressentiment are pivotal both in
generating and maintaining the kind of ethical life that Nietzsche wants to overcome” (2003, 155).
of the soul—that is at stake here! No sacrifice can then be too great: even one’s dearest
friend one must be willing to sacrifice for it, though he be the most glorious human
being, embellishment of the world, genius without peer” (GS 98).
Nietzsche presents Goethe as the embodiment of many of these qualities. He writes,
Goethe conceived a human being who would be strong, highly educated, skillful in all bodily
matters, self-controlled, reverent toward himself, and who might dare to afford the whole range
and wealth of being natural, being strong enough for such freedom; the man of tolerance, not
from weakness but from strength because he knows how to use to his advantage even that from
which the average nature would perish; the man for whom there is no longer anything that is
forbidden, unless it be weakness, whether called vice or virtue. Such a spirit who has become free
stands amid the cosmos with a joyous and trusting fatalism, in the faith that only the particular is
loathsome, and that all is redeemed and affirmed in the whole—he does not negate any more.
Such a faith, however, is the highest of all possible faiths . . . (TI IX.49)29
Nietzsche also presents himself as having these qualities; see especially EH I.2 and
EH IV.
In sum, Nietzsche’s ethical exemplars devote themselves to difficult accomplish-
ments, continuously set aside these prior accomplishments in order to seek new ones,
affirm their own lives, are honest, and display a form of independence from their age.
How might these features relate to will to power? Notice that these features have
something in common: no determinate goal is posited. All of these features relate to the
manner in which we act, rather than the particular goals that we pursue.30 Thus, what
unifies Goethe, Nietzsche, Beethoven, and Napoleon is not some shared goal, but the
fact that they undertook immensely difficult tasks—tasks that required a great deal of
creativity and perseverance—and nevertheless managed to achieve at progressively
greater heights.
But this raises a question. Some of Nietzsche’s exemplars, such as Goethe, Beetho-
ven, and Shakespeare appear rather tame and unobjectionable by light of conventional
moral values. Others, including Napoleon, raise more qualms: after all, these are
individuals who engaged in warfare and political domination. And this focus on
Napoleon (not to mention Nietzsche’s occasional praise of Cesare Borgia and his ilk)
perennially leads readers of Nietzsche to wonder: how would the will to power thesis
assess a more modern conqueror, Hitler? Insofar as Nietzsche’s ethic praises the pursuit
of power, won’t individuals like Hitler turn out to be at least as exemplary as Goethe?
After all, no one can deny that Hitler pursued immensely difficult tasks, overcame great
resistances, and so forth.
29
For an illuminating discussion of this passage, see Risse (2003).
30
Reginster offers an insightful discussion of the way in which Nietzsche’s values do not have determinate
contents. As he puts it, it is a “crucial requirement” of Nietzsche’s view that “the standard of perfection
should remain devoid of determinate content: it represents the indeterminate and ever-receding objective of
an essentially insatiable creative impulse that is valued for its own sake” (Reginster 2007, 51).
Commentators sometimes attempt to sidestep this issue by pointing out that Hitler
himself had characteristics that Nietzsche deplored.31 That is no doubt true; Nietzsche
would have condemned the actual Hitler. But consider a slight variant of Hitler: a
conqueror just as brutal who is not motivated by resentment. Would the Nietzschean
ethic praise this individual? Or, to pose the question in a more general form: manifest-
ing the five traits discussed above is necessary for greatness, but is it also sufficient?32 If it
is, we face a problem. For one can manifest these traits when pursuing art or literature
or philosophy; but one can also manifest them when pursuing military conquest or
murder or physical dominance. Are there any grounds for ruling out some of the latter
pursuits?
There are. As I explained in Chapter 7, will to power is not the sole criterion by
means of which we evaluate actions. We are also to assess our actions in light of
other values—values that, though they may be merely conventional and may lack
any deep foundations, have been vetted in terms of will to power. If these values
rule out oppression, murder, and so forth, then the Nietzschean theory will indeed
condemn Hitler and his ilk. Or, put differently: if our disvaluation of murder and
oppression is compatible with willing power—and everything above suggests that it
is—then the Nietzschean ethical theory can, indeed, condemn individuals who
engage in these acts.33
I suggest that this is why an overriding theme in Nietzsche’s work is the claim that
his ethic promotes creative genius. Nietzsche writes that “men of great creativity” are
“the really great men according to my understanding” (KSA 11:37[8]). The five traits
that Nietzsche takes as characteristic of greatness—striving for difficult tasks, continu-
ously overcoming oneself, self-affirmation, honesty, and independence—are traits
exemplified both in creative and in aggressive actions, but the creative actions—unlike
the aggressive ones—remain consistent with our other values. So we can treat these
five characteristics as necessary but not sufficient conditions for fully realizing the
demands of Nietzschean constitutivism.
In the end, that is what the Nietzschean theory recommends: not so much a specific
set of determinate goals for everyone to pursue, but a particular way of pursuing goals
that are chosen for their own sake. Of course, the Nietzschean theory does generate
31
Most obviously, Hitler was unquestionably motivated by ressentiment, which Nietzsche takes to be
characteristic of the slaves. As Ansell-Pearson puts it, “Hitler was a man whose whole being was pervaded by
feelings of deep-seated resentment and poisonous revenge, and he can hardly be held up as an example of
Nietzsche’s model of the noble individual” (Ansell-Pearson 1994, 31).
32
Nehamas raises this question in a perceptive essay (1999, 7). Reginster claims that although Nietzsche
himself “did not offer a developed response to this question,” we can hold that “the difficulty of an
achievement is a necessary condition of its being great. It is not sufficient for greatness, however, because
the determinate content of the achievement must itself be valuable” (2007, 43).
33
For an insightful defense of a related point, see Hurka (2007, 29–30). (Hurka’s interpretation of
Nietzsche’s ethical theory differs from my own, but Hurka’s claims concerning the compatibility of
Nietzschean ethics with other values are more broadly applicable.)
some constraints on what is to be willed: it rules out goals such as the complete
avoidance of suffering, the utter absence of struggle, and so forth. But the theory
does not attempt to single out particular goals that are to be pursued by all agents.34
There will be some specific values regulating interpersonal relationships—such as the
positive valuation of restricted egalitarianism—that are selected so as to promote this
way of pursuing ends. But these values leave a great deal of latitude, enabling agents to
choose a range of ends for their own sake.
Thus, Nietzschean constitutivism enjoins us to select ends that can be willed in a
particular manner: a struggle-ridden manner. Nietzsche presents this as a path to a
“new happiness” (GS Preface). It is not a happiness of equanimity, not the feeling of
abiding in some pleasant state—Nietzsche derides those forms of happiness as
“wretched contentment” (Z Preface), telling us that they consist merely in “the
happiness of resting, of not being disturbed, of satiety” (BGE 200). The happiness
that the Nietzschean theory holds out to us is, instead, a happiness born of continuous
achievement: we are to experience joy in constant struggle and overcoming.
In struggling for this happiness, we often fail:
You higher men, for you it longs, joy . . . for your woe, your failures. All eternal joy longs for
failures. For all joy wants itself, hence it also wants agony. (Z IV.19)
But this is a necessary precondition for Nietzsche’s “new happiness”: in wanting the
joy of achievement, we also want struggle and “agony.” In this sense, we long for
failures. Thus, although this struggle involves pain, the
Pain is sanctified: the ‘pains of childbirth’ sanctify pain in general—all becoming and growing, all
that guarantees the future, postulates pain . . . For the eternal joy in creating to exist, the will to life
eternally to affirm itself, the ‘torment of childbirth’ must also exist eternally. (TI IX.4)
Nietzsche’s “new happiness,” then, involves pain and difficulty in production or

creation.
This, Nietzsche thinks, is how we flourish, how we best satisfy our deepest and most
ineluctable aim. We are not content to abide in any given state, no matter how good or
pleasant. Rather, we take “joy in the destruction of the noblest and at its progressive
ruin: in reality joy in what is coming and lies in the future, which triumphs over
existing things, however good” (KSA 10:8[14]/WP 417).
This is a hopeful vision, calling us toward intellectual and artistic achievement.
But it is also an immensely challenging vision, for it denies the possibility of a final
resting point. What awaits us is not some state of contentment, but only continuous
34
In this respect, the Nietzschean theory is closer to Kant’s than to Bentham’s and Mill’s: whereas
utilitarian ethical theories entail that there is typically only one correct action to perform, the deontological
theories such as Kant’s give us more latitude, prohibiting certain courses of action but leaving us free to select
from among many others.
CONCLUDING REMARKS 237
overcoming.35 No attained ideal can satisfy us, for “whoever reaches his ideal tran-
scends it eo ipso” (BGE 73).
4. Concluding remarks
I have devoted the majority of this book to exploring how the structure of agency
might disclose substantive normative commitments. I have argued that action has two
constitutive aims: activity and power. In that sense, my primary goal in this book has
been to show how a constitutivist theory can even begin, not how it in fact ends.
However, in this chapter I have scratched the surface of that ending, by examining a
few of the particular normative conclusions entailed by the constitutivist theory. I have
highlighted three such conclusions. First, the theory requires a rejection of the idea that
all values are universally applicable. Second, the theory reveals some potential problems
with extreme versions of egalitarianism, though it is perfectly compatible (Nietzsche’s
protests to the contrary notwithstanding) with more modest forms of egalitarianism.
Third, the theory recommends, as its highest ideal, a life of continuous attainment.
35
Kafka gives us a wonderful image of this: “Human nature, which is fundamentally changeable and
unstable as billowing dust, endures no restraint. If it restricts itself, it will soon begin to shake the restraints
madly until it rends everything asunder, the wall, the bonds, and its very self ” (Kafka, “The Great Wall of
China”).
9
Activity, Power, and the
Foundations of Ethics
I began this book by posing the foundational question in ethics: how are normative
claims justified? In particular, how could universal normative claims—claims that apply
to all agents, independently of their idiosyncratic psychological states—be justified?
I have argued that constitutivism is the best strategy for answering this question.
Constitutivism is based on the idea that every action shares a common, higher-order
aim, whose presence generates a standard of success for action. As this aim is present in
every action, it generates universal reasons: every agent has reason to perform those
actions that meet the standard of success.
Constitutivism avoids the metaphysical, practical, and epistemological problems
plaguing traditional methods of justifying ethical claims. It does not require us to posit
irreducible normative truths or to rely on unsupported intuitions about what there
is reason to do. Instead, it simply anchors universal normative claims in agents’ aims.
However, constitutivism does face the difficult task of articulating a conception of
action that entails the presence of a constitutive aim. After all, our actions appear to have
wildly diverse, conflicting aims. It is hard to believe that every action—from brushing
one’s teeth to studying philosophy to going for a jog—shares a common aim. The
constitutivist must show how, in the face of this fact, actions have a constitutive aim.
I have shown that a roughly Nietzschean account of agency enables us to establish
the presence of constitutive aims in action. I argued that this account of agency is
independently plausible: regardless of whether we are attracted to constitutivism, we
can see that the conception of motivation and agency is supported both by contem-
porary empirical work on human psychology and by philosophical arguments. We thus
have independent reason for accepting the account. However, it turns out that if we do
accept the roughly Nietzschean account of agency, then a straightforward argument
commits us to the claim that action has two constitutive aims: activity and power.
Let’s start with activity. I maintained that there is a distinction between actions that
spring from our own activity and actions that are brought about in a more passive
fashion. Actions that spring from our own activity can be said to be manifestations of
“agential activity.” I argued that merely in virtue of acting reflectively, agents aim to
manifest agential activity. In other words, we aim at being the cause of our own
actions.
A C T I V I T Y, P OW E R , A N D T H E F O U N DAT I O N S O F E T H I C S 239
But what is it to manifest agential activity? Traditionally, philosophers have argued

that reflective or deliberative actions are paradigmatic cases of agential activity. This
contention is supported by the claim that reflection or deliberation enables one to
suspend the effects of one’s motivational states: in reflecting on a motive, I achieve a
distance from it and suspend its influence on my deliberation. Actions that spring from
my own reflection therefore seem significantly different than actions that are merely
caused by some motivational state.
However, I have argued that this traditional model of agential activity is mistaken.
This model, which is inspired by Locke and Kant, treats motivation as something that is
independent from reflective thought. By contrast, I have argued that motivation is
properly seen as operating through reflective thought. This renders untenable the
Kantian/Lockean claim that we are capable of suspending our motives: once we see
how motives impact reflective thought, we can see that an agent might reflect on a
motive, take herself to have suspended its effects, and all the while be under its thrall.
Once we recognize the ways in which motives operate through reflective thought, we
will be unable simply to identify agential activity with reflective or deliberative agency.
Instead, I have argued that we should adopt a new account of agential activity.
According to this account, an agent is active in the production of her action if (a) she
approves of her action, and (b) this approval would be stable in the face of further
information about the etiology of her action. This account of agential activity is
compatible with the more accurate account of motivation, according to which motives
impact reflective thought itself. In addition, it correctly accounts for certain paradig-
matic cases of activity and its absence.
This account of agential activity provides a constitutive standard for action: an action
is successful to the extent that (a) the agent approves of it, and (b) this approval is stable
in the face of further information about the action’s etiology. However, I pointed out
that this approval is based, at least in part, on the values that the agent currently accepts.
For example, if I accept the claim “jealousy is bad,” I am unlikely to approve of any
actions that spring from jealous motives. On the other hand, if I accept the opposite
valuation—“jealousy is good”—I am likely to approve of these actions.
The account of agential activity does not by itself provide us with a way of assessing
these values. However, this is where the second constitutive aim—will to power—
plays its part.
Nietzsche’s account of agency is based on the idea that our actions are motivated by
a distinctive kind of psychological state: the drive. Drives are dispositions that engender
affective orientations. Drives admit an aim/object distinction, where the aim is a
characteristic form of activity and the object is a chance occasion for expression. For
example, the sex drive aims at sexual activity, and might take as its temporary object a
particular person. Drives are not satisfied by the attainment of their objects, for these
objects are simply adventitious. Accordingly, there is no object the attainment of which
completes or satisfies the drive. Rather than disposing us to acquire their objects, then,
drives simply lead us to engage in their characteristic form of activity.
240 A C T I V I T Y, P OW E R , A N D T H E F O U N DAT I O N S O F E T H I C S
This fact about drives entails that all drive-motivated actions share a common,
higher-order aim of encountering and overcoming resistance. For part of what it is
to be motivated by a drive—to manifest a drive’s activity—is continuously to seek new
objects on which to direct the drive’s activity. In seeking new objects, we seek new
resistances. In virtue of being motivated by drives, then, we seek continuous encoun-
tering and overcoming of resistance, rather than the attainment of definite states.
Nietzsche describes this fact about our aims by saying that we manifest “the will to
power.”
I have argued that this aim of power generates its own standard of success: we have
reason to perform those actions that provide opportunities for encountering and
overcoming resistance. Moreover, we can use this standard as a “principle of revalu-
ation.” That is, we can use it to assess our current values. The standard tells us to reject
those values that generate internal conflicts with will to power, and to modify our
other values so that they minimize external conflicts with will to power.
The constitutive aim of agential activity interacts with the constitutive aim of power
in two ways. First, the principle of revaluation generated by the will to power gives us a
way of assessing the values upon which activity depends: we reject those values that
generate conflicts with will to power. Second, recall that agential activity requires the
agent’s approval of her action, given further knowledge of the action’s etiology.
Because will to power is present in the etiology of every action, in order to manifest
agential activity we must approve of will to power. Put differently, insofar as we act at
all, we aim to approve of will to power.
With that, we can summarize the full ethical theory: an action is successful if it is a
manifestation of agential activity and it encounters and overcomes resistance. Both of
these standards arise merely in virtue of facts about the nature of action. They are
therefore universal, applying to all agents.
However, some of the particular normative claims generated by this theory do vary
across individuals. For, on this account, the majority of values are not derived from the
constitutive standard. Rather, the constitutive standard is used to assess values that arise
from disparate sources. Where we end up depends, in part, on where we begin.
Though my primary task has been to articulate the structure of a successful con-
stitutivist theory, I have concluded by giving some examples of the determinate
normative conclusions generated by this Nietzschean version of constitutivism. In
order to show how the theory rules out certain valuations, I have used egalitarianism
as a case study. The Nietzschean version of constitutivism requires us to rethink some
central aspects of egalitarianism. In particular, it commits us to rejecting radical
egalitarianism, though it is compatible with a more restricted form of egalitarianism
according to which agents have equal rights. In addition to ruling out certain
values, the Nietzschean theory endorses other values. As an example, I pointed out
that it enjoins us to value creativity, the seeking of challenges, achievement, and
independence of mind. Perhaps most dramatically, it requires us to rethink our
negative evaluation of suffering and our positive evaluation of pleasure.
A C T I V I T Y, P OW E R , A N D T H E F O U N DAT I O N S O F E T H I C S 241
The normative results generated by the theory, then, are in conformity with some
elements of our current evaluative framework, and deeply incompatible with others.
This should not be surprising: as I pointed out in Chapter 1, it would be facile to
assume that our current evaluative framework is wholly correct.
So much for the summary. What else can be said in favor of the theory? To begin,
notice that the Nietzschean theory differs from the versions of constitutivism defended
by Christine Korsgaard and David Velleman in three ways. First, the Nietzschean
theory is non-foundationalist. Second, it is empirical or a posteriori. Third, it is
bipartite. I will explain these features in turn.
Let’s begin by examining the non-foundationalist structure of the theory. The
Nietzschean version of constitutivism does not treat all values as derived from some
foundational principle, such as the Categorical Imperative. Rather, it uses the consti-
tutive standards to assess values that arise from disparate sources. So there is no
aspiration, here, to reduce all legitimate norms to one source. That quest, I think, is
quixotic. In this respect, the Nietzschean theory is more modest in its aspirations than
its competitors.
Second, the Nietzschean theory is empirical or a posteriori. Korsgaard’s theory
operates with a series of conceptual claims about the nature of autonomous action.
Velleman’s theory begins with the idea that action involves immediate knowledge.
The Nietzschean theory, by contrast, investigates the psychological structure of human
motivation. We have a constitutive aim of power because the etiology of every human
action includes one or more drives. There is no conceptual necessity here; we might
have evolved differently, such that we were not motivated by drives. If that had
happened, we would be subject to different normative claims. In this respect, the
Nietzschean theory is again more modest than its competitors.
Finally, the Nietzschean theory is bipartite. Whereas Velleman’s theory posits only
one constitutive aim—self-understanding—the Nietzschean theory claims that we aim
jointly at agential activity and power. In this respect, it bears some similarity to
Korsgaard’s theory. Recall that Korsgaard’s account posited not one, but two consti-
tutive standards: the Categorical and Hypothetical Imperatives. Korsgaard sometimes
describes this feature of her theory by claiming that action’s constitutive standards are
both autonomy and efficacy. Roughly, the norms of autonomy give us a way of
determining which ends we should have, whereas the norms of efficacy simply tell us
what to do given that we have an end.
The Nietzsschean theory can be seen as drawing on, and inverting, this idea from
Kant. The Nietzschean theory, too, posits two constitutive aims. First, the account
holds that our actions aim at something we could call “autonomy,” namely agential
activity. That is, autonomy is interpreted in terms of agential activity rather than as
being subject to the Categorical Imperative. Agential activity is the Nietzschean ana-
logue of Kantian autonomy: it is an aim that a reflective agent must have merely in
virtue of being active. Yet the Nietzschean theory also posits a second constitutive aim:
the will to power. Power is an aim that arises not from our nature as reflective agents, but
242 A C T I V I T Y, P OW E R , A N D T H E F O U N DAT I O N S O F E T H I C S
merely from our nature as agents. It is shared, Nietzsche claims, by all creatures that act:
it is the one aim that is intrinsic to action as such. Moreover, the will to power can
be understood as the Nietzschean analogue of Kantian efficacy: whereas Kantian
efficacy requires us to take the necessary and available means to our ends, Nietzschean
power requires us to pursue our ends by seeking obstacles and resistances to those
ends—in other words, by selecting resistance-inducing means.
Crucially, though, the Nietzschean theory inverts the Kantian schema. Agential
activity (the Nietzschean analogue of Kant’s autonomy) generates minimal normative
content, whereas will to power (the Nietzschean analogue of Kant’s efficacy) generates
a great deal of normative content. Or, to put the point differently, very little normative
content arises from our nature as reflective agents (via agential activity), but a great deal
of normative content arises from our nature as agents (via will to power). This is
precisely the opposite of Kant’s theory, which derives the bulk of its normative
conclusions from the Categorical Imperative, a principle applying solely to reflective
agents.
In sum, a roughly Nietzschean theory of agency enables us to defend a version of
constitutivism that is non-foundationalist, a posteriori, and bipartite. This model
promises to resolve the foundational question in ethics. For it shows that we can
ground certain universal normative claims in a model of agency that, I have argued, is
well supported by both empirical results and philosophical arguments.
The picture of agency and value articulated by Nietzsche is striking. On this view, it
is constitutive of human agency that it is never satisfied, that it always seeks more. As
Nietzsche puts it, “Whatever I create and however much I love it—soon I must oppose
it and my love—thus my will wills it” (Z II.12). This image of human life as a process of
continuous attaining, never satisfied, was perhaps best captured by Dostoevsky:
Man is a frivolous and incongruous creature, and perhaps, like a chess player, loves the process of
the game, not the end of it. And who knows (there is no saying with certainty), perhaps the only
goal on earth to which mankind is striving lies in this incessant process of attaining, in other
words, in life itself, and not in the thing to be attained. (Dostoevsky, Notes From Underground )
By modeling its theory of agency on this conception of human nature, the Nietzschean
version of constitutivism resolves the foundational question in ethics. It entails a host of
substantive and often surprising universal normative claims, and holds up for us as the
highest ideal the life devoted to creative activity. Constantly seeking, impelled toward
ends that do not satisfy, human action’s constitutive standards generate an ethic of
attaining, rather than attainment.
Appendix: Is Nietzsche Really
a Constitutivist?
I have argued on roughly Nietzschean grounds that will to power is a constitutive aim of action.
As noted earlier, my primary goal in this volume has been the defense of a constitutivist theory
rather than full-fledged textual exegesis. I do think that the arguments given in Chapters 6
through 8 represent Nietzsche’s actual view, but given the complexity of Nietzsche’s texts,
establishing his commitment to this view would necessitate a book of its own.1 Accordingly,
I have presented the view articulated in these chapters as a Nietzschean view rather than
Nietzsche’s view. For present purposes, the question of whether the theory is true is more
important than the question of whether it is Nietzsche’s.
Nevertheless, I am aware that many readers of Nietzsche will be astonished by my suggestion
that his theory is best interpreted along constitutivist lines. How could Nietzsche, of all philoso-
phers, think that ethics is grounded in action? Could he really believe that it’s possible to derive
universal reasons from facts about agency? Isn’t Nietzsche a great skeptic about ethics, attempting
not to construct a theory but to knock down all theories?
In short, my answer is no. It is certainly easier to see what Nietzsche is against than what he is
for; his texts devote far more space to critique than to the construction of a positive view.
However, as I pointed out in Chapter 6, Nietzsche’s ethical critiques are inextricably bound up
with the articulation of a positive view: his critiques presuppose that power has a privileged
normative status. “Only as creators can we destroy,” he writes (GS 58).
Given that many readers may find my use of Nietzsche surprising, I do want to spend a
moment addressing some central interpretive concerns. My goal in this appendix is not to
establish, decisively, that my reading of Nietzsche is correct; I lack the space for such a task.
My goal is more modest: I will show that my interpretation of Nietzsche cannot be casually
dismissed. In particular, I will argue that many initial misgivings about attributing a constitutivist
view to Nietzsche are based on confusions.
With this in mind, let me mention six reasons for resisting my reading of Nietzsche. This list
may not be exhaustive, but it includes the most common reactions to my reading.
1) The constitutivist theory is a version of Kantianism, so the constitutivist reading of

Nietzsche implies that Nietzsche is a Kantian. But this is absurd; Kant is one of Nietzsche’s
central targets.
2) The constitutivist theory implies that there are universal reasons, but Nietzsche denies that
there are any universal reasons.
3) The various distinctions that I draw in the previous chapters—such as the distinction
between the force and the ubiquity of reasons—aren’t present in Nietzsche’s texts.
1
I do offer interpretive defenses of some aspects of this view in a manuscript in progress entitled The
Nietzschean Self: Agency and the Unconscious, as well as in Katsafanas (2011a, 2011b forthcoming a, and
forthcoming b).
244 A P P E N D I X : I S N I E T Z S C H E R E A L LY A C O N S T I T U T I V I S T ?
4) The constitutivist theory is incompatible with Nietzsche’s model of agency.

5) The constitutivist theory is a theory, an articulated ethical view; but Nietzsche does not
offer any positive theories at all. He is engaged in a purely critical project.
6) The constitutivist theory is based on two ideas: (a) Nietzsche gives power a privileged
normative status, and (b) Nietzsche thinks we can justify the claim that power has a
privileged normative status. But we can reject one or both of these claims.
Objections (1)–(4) can be dealt with rather quickly; I will address them in Section 1. Section 2
examines (5), and Section 3 considers (6).
1. The first four objections

Let’s start with objection (1). I take it as a datum that Nietzsche is not endorsing a Kantian ethical
theory; his texts are unambiguously critical of Kant. Thus, claim (1) would be a decisive objection
to my reading. However, claim (1) is based on a misunderstanding. As I explained in Chapter 1,
constitutivism is more properly Humean than Kantian. It is simply a historical accident that the
currently prominent versions of constitutivism, defended by Korsgaard and Velleman, are
Kantian. However, the version of constitutivism that I attribute to Nietzsche is quite close to
what I called, in Chapter 1, a Humean constitutivism. Put simply: I am not claiming that
Nietzsche is a Kantian; I am claiming that he is a more sophisticated version of Hume.
This brings us to a related concern, expressed in objection (2). At the very least, constitutivism
resembles Kantianism in holding that there are certain universal reasons. But doesn’t Nietzsche
deny that there are any universal reasons?
I addressed this concern in Chapter 8, arguing that Nietzsche is a parametric universalist.
Nietzsche certainly does reject universalism in one sense—the universalism present in figures
such as Kant, who holds that there is one normative standard that generates the same normative
results for all rational agents. But I’ve argued that Nietzsche’s texts make it clear that he is
committed to a quite different type of universalism, which I’ve called parametric universalism.
On this reading, Nietzsche does think that there is one universally valid normative standard:
power. However, other, more particular values can vary.
Let’s turn to objection (3). I concede this point: in order to make the Nietzschean version of
constitutivism succeed, we must introduce various distinctions and complications that are not
explicitly present in Nietzsche’s texts. For example, as Chapter 7 explained, we must be more
precise about the way in which values can conflict with power; we must recognize a distinction
between pro tanto and all things considered reasons; and, if we adopt certain views of reasons, we
may need to appeal to a notion of agential activity in order to explain why we should treat the
will to power as normative rather than merely compulsive. This does not show that Nietzsche is
not a constitutivist; it simply shows that Nietzsche failed to consider certain sophisticated
philosophical objections to his view. This is hardly surprising.
Finally, consider objection (4). Is constitutivism really compatible with Nietzsche’s model of
agency? In particular, doesn’t the model of agency required by the constitutivist reading seem
altogether too reflective for Nietzsche?
Nietzsche certainly doesn’t imagine agents reflectively considering their reasons for action at
every moment, or engaging in bouts of deliberation prior to each action. Indeed, as I pointed out
in Chapter 7, Section 3, Nietzsche claims that most of our actions are habitual, relatively
unreflective, and automatic. But, as I noted in the same section, this is compatible with my
DOES NIETZSCHE OFFER THEORIES? 245
reading of Nietzsche’s will to power thesis. As I interpret him, Nietzsche does not imagine that
agents will reflectively consider the relation between power and particular actions and values at
each moment; rather, he envisions agents periodically assessing their actions and values, primarily
in a retrospective fashion. So the will to power thesis, as I interpret it, does not rely on an
excessively intellectualistic or reflective model of agency; on the contrary, it requires only the
minimal claim that agents periodically reflect on their actions and values and sometimes adjust
these actions and values in light of this reflection.
Of course, some interpreters will object even to this modest thesis. For a few interpreters read
Nietzsche as denying the very possibility of reflective thought’s having a causal impact on action.
Brian Leiter, for example, has argued that Nietzsche is an epiphenomenalist about conscious
willing (Leiter 2001 and 2007). If this interpretation is correct—if Nietzsche treats conscious
thought or conscious willing as causally inert—then my reading of his will to power thesis
would, indeed, be problematic. However, in other works I have argued that Leiter’s epipheno-
menalist reading of Nietzsche is indefensible.2 If my arguments on that point are correct, then my
interpretation of the will to power thesis is perfectly compatible with Nietzsche’s model of
agency.
2. Does Nietzsche offer theories?

Above, I’ve addressed various concerns about the details of the theory that I attribute to
Nietzsche. However, some readers will object not to the details of my interpretation, but to
the very idea that Nietzsche offers philosophical theories. For some readers interpret Nietzsche as
an opponent of all philosophical theorizing. To cite just two prominent examples: Bernard
Williams (1994) claims that Nietzsche’s texts are “booby trapped” against the extraction of
philosophical theories—in other words, that Nietzsche’s concerns are purely critical rather than
constructive. Robert Pippin has recently defended Williams’ claim at some length (2010).3 If this
is correct—if Nietzsche’s texts are designed so that any attempt to extract a theory will fail—then,
of course, my reading will be unsuccessful.
However, my response to this objection is straightforward: what looks like a “trap” or a
contradiction in Nietzsche’s texts is often revealed, upon closer examination, to arise from an
impoverished view of the philosophical options. Once we rethink some of our own presuppos-
itions, these alleged contradictions often dissolve. Consider a simple example: in Twilight,
Nietzsche both denies and affirms the causal efficacy of the will. He first writes, “The ‘inner
world’ is full of phantoms and illusions: the will is one of them. The will no longer moves
anything, hence does not explain anything—it merely accompanies events; it can even be
absent” (TI VI.3). This passage seems to deny the causal efficacy of the will. Yet, a few pages
2
See, in particular, Katsafanas (2005) and a work in progress entitled “Nietzsche and Kant on the Will:
Two Models of Reflective Agency.”
3
This kind of reading is sometimes supported by quoting Nietzsche’s famous remark: “I mistrust all
systematizers. The will to a system is a lack of integrity” (TI I.26). However, as Reginster has argued (2006,
3), Nietzsche most likely has in mind a rather specific sense of system: the post-Kantian demand, articulated
(in quite different ways) by thinkers ranging from Reinhold to Fichte to Hegel to Schelling, to account for all
philosophical problems in an interconnected way. We can avoid attributing that kind of systematic aspiration
to Nietzsche while still maintaining that Nietzsche’s thought is, as Reginster puts it, “organized and logically
ordered, and not a haphazard assemblage of brilliant but disconnected ideas” (2006, 3).
later, Nietzsche says that certain individuals have the capacity to will, where willing is defined as
the power “not to react at once to a stimulus, but to gain control of all the inhibiting, excluding
instincts . . . the essential feature is precisely not to ‘will’, to be able to suspend decision. All
unspirituality, all vulgar commonness, depend on an inability to resist a stimulus: one must react,
one follows every impulse” (TI VIII.6). So, within a few pages, we have the denial and the
affirmation of the causal efficacy of the will. On the Williams/Pippin reading, this is a booby trap,
designed to thwart philosophical theorizing. But, in fact, there is no deep mystery or ineluctable
tension here: Nietzsche is merely denying one conception of the will and replacing it with
another.4 As the texts make clear, he denies the existence of the libertarian will, the will conceived
as a faculty exempt from determination by prior events. But he accepts the existence of the will
conceived in more naturalistically respectable terms: as a capacity, not exempt from causal
determination, which enables agents to counter the effects of certain motives, and whose
strength varies from individual to individual.5
In short: what looks like a series of “booby traps” or contradictions turns out to be perfectly
consistent and sensible, provided that we don’t assume that the only legitimate conception of the
will is the libertarian account. This is just one example, but the point generalizes: when Nietzsche
seems to be asserting both P and not-P, a more careful reading typically reveals that he is drawing
subtle distinctions or introducing novel ways of dividing the conceptual space.
In light of this, let’s take stock of the state of the debate. I claim that Nietzsche is offering an
ethical theory. My opponents claim that Nietzsche offers no theories of any kind, because his
texts are designed to thwart theorizing. If the claim that Nietzsche is not a theorist is to be more
than a bare assertion, then we need to show what, in particular, is wrong with my reading of him
as a constitutivist. The objection cannot be: Nietzsche offers no theories, therefore your reading
is wrong. The objection must be more specific than this. It must show where, in particular, my
reading goes wrong. It will then collapse into one of the other, more specific objections.
3. Objections to the reliance on the will to power thesis

Chapter 6 relied on two interpretive claims: Nietzsche gives power a privileged normative status,
and Nietzsche claims that each instance of action aims at power. These interpretive claims are
relatively uncontroversial: they constitute an exceedingly common interpretation of Nietzsche.6
Nonetheless, some skeptical commentators argue that there are grounds for rejecting one or both
of the claims. In this and the following sections, I will argue that these skeptics are mistaken;
Nietzsche’s commitment to the two claims is clear.
4
Other passages from the late works make it explicit that Nietzsche is rejecting only some conceptions of
the will. For example, Nietzsche writes, “today we have taken [man’s] will away altogether, in the sense that
we no longer admit the will as a faculty [Vermögen]” (A 14). Notice that Nietzsche says he is rejecting the will
as a faculty. While Nietzsche does not explain what he means by “faculty,” it seems natural to assume that
conceiving of the will as a faculty involves conceiving of it as causally isolated from the agent’s drives and
motives.
5
I explore Nietzsche’s view of the will in more depth in “Nietzsche and Kant on the Will: Two Models
of Reflective Agency.”
6
To give a brief sampling, all of the following works defend some version of these claims: Kaufmann
(1974), Wilcox (1974), Heidegger (1979), Schacht (1983), Hunt (1991), Richardson (1996, 2004), Jaspers
(1997), May (1999), Reginster (2006), and Hussain (2011).
O B J E C T I O N S T O T H E R E L I A N C E O N T H E W I L L T O P OW E R T H E S I S 247
There are three potential objections, which I will address in turn. First, some commentators
argue that Nietzsche eventually abandoned his will to power thesis. Second, Brian Leiter has
recently argued that Nietzsche views will to power as unimportant. Third, Leiter has also claimed
that Nietzsche rejects the idea that each action aims at power. If these claims are correct, they
might suggest a different resolution to the interpretive problem that I canvassed in Chapter 6:
perhaps Nietzsche resolves the interpretive problem simply by giving up his claim that power has
a privileged normative status. Below, I argue that this is not the case.
3.1 Does Nietzsche abandon his will to power thesis?

Let’s start with the first point. Commentators sometimes claim that Nietzsche ultimately rejected
his will to power project. For example, Leiter writes that “recent scholarship has cast doubt on
whether Nietzsche ultimately accepted” a will to power thesis according to which “all life
(actions, events) reflects the will to power” (Leiter 2002, 139). Hollingdale claims that Nietzsche
“abandoned” his project of revaluing values in terms of will to power (Hollingdale 1999, 220).
Young goes even further, writing that Nietzsche rejected the very idea that all actions manifest
will to power (Young 2010, 542–6). If this is right, then my interpretation of Nietzsche would
not be consistent with his mature view.
However, these claims are highly misleading; they are speculations based on dubious inter-
pretations of textual evidence. To see why, we need to disentangle a few threads.7
Part of the controversy surrounding Nietzsche’s attitudes toward will to power arises from the
nature of the book published as The Will to Power. This book is a haphazard, disorganized
collection of extracts from Nietzsche’s notebooks, pasted together and organized under often
arbitrary headings by Nietzsche’s sister and her editors; it not only falsifies the order of
Nietzsche’s notes, but includes notes that Nietzsche expressly discarded. I take it as obvious
that this pseudo-book has no value; I have not relied on it here.
Another controversy arises over whether Nietzsche himself intended to publish a work titled
The Will to Power. At times, he did: from 1884 to 1888, Nietzsche’s notebooks are full of plans to
write either a book or a series of books that would focus on the revaluation of values and the will
to power. His notebooks reveal that Nietzsche tried out several different titles, often settling on
some variant of The Will to Power or The Revaluation of All Values. By September 1888, Nietzsche
has settled on a four-book series called The Revaluation of All Values. Then we come to two stories
that are thought by Young, Leiter, Hollingdale, and others to have a decisive bearing on
Nietzsche’s thought.
Let’s start with the first story. From June to September 1888, Nietzsche stayed at a guesthouse
in Sils-Maria and composed many notes on will to power and other topics. Hollingdale (1999,
250) claims that when Nietzsche departed, he left behind and instructed his landlord to throw
out many of these 1888 notebook writings; however, Hollingdale tells us that the landlord saved
the manuscripts, and some of them were ultimately published in the Will to Power. Others
7
I will be concerned only with the claim that will to power is present in all actions. BGE 36 and
Nietzsche’s notebooks contain speculations about whether will to power might extend even more broadly—
for example, he wonders whether the concept of force can be analyzed in terms of will to power. This
metaphysical version of the will to power thesis seems less important to Nietzsche; he appeals to it only once,
in a highly hedged manner, in the published works (BGE 36). In any case, it is irrelevant for our purposes.
Accordingly, I will not address it here. (For a helpful discussion of this metaphysical version of the will to
power thesis, see Clark [1990, 209ff.].)
have repeated Hollingdale’s story, claiming that Nietzsche wanted his notebooks destroyed; see,
for example, Leiter (2002, xvii) and Young (2010, 628 note 9). This story is often cited in support
of the idea that Nietzsche abandoned his will to power project: after all, if Nietzsche consigned so
many of his writings on will to power to the wastebasket, he can hardly have regarded those
notes as important! However, the story is apocryphal. Hollingdale’s only cited source for this
story is a magazine article from 1893, but as Magnus (1986) points out, this article is flatly
inconsistent with Hollingdale’s claims: the article says that Nietzsche left behind and instructed
his landlord to throw out not notebooks, but page proofs of Twilight of the Idols.8 So this story
appears to be a mere myth that has somehow managed to live on in certain areas of Nietzsche
scholarship.
This brings us to the second story. As mentioned above, by September 1888, Nietzsche has
settled on a four-book series on the revaluation of values. His notes from this period treat The
Antichrist as one of these four books. Then we come to the point which concerns Young,
Hollingdale, Leiter, and others: in certain notes from mid to late November 1888, Nietzsche
presents The Antichrist not as one part of, but as the totality of, his Revaluation of All Values.9
Nietzsche’s productive life ends shortly thereafter: by mid December 1888, Nietzsche is slipping
into insanity.10
What should we make of this? I am inclined to say: very little. After all, we do not need to
speculate about Nietzsche’s plans for unwritten works in order to support his claims about will to
power. Nietzsche’s actual published works from Zarathustra to the Antichrist provide ample
evidence for the claim that all actions manifest will to power and that will to power is the
principle of revaluation, as I have indicated above.11
But some commentators think we can infer a great deal from the history outlined above. In
particular, some commentators conclude that because several notes from mid to late November
1888 treat the Antichrist as the totality of the Revaluation of All Values, Nietzsche must have
abandoned his will to power thesis. For example, Hollingdale writes that Nietzsche “abandoned”
his revaluation project because he recognized that “the Revaluation was no advance upon the
philosophy completed in Zarathustra” (Hollingdale 1999, 220). Leiter makes a similar claim
(2002, xvii). Analogously, in his recent biography of Nietzsche, Julian Young reviews the
notebook evidence mentioned above and concludes that “the claim that all human motivation
can be reduced to will to power” is “abandoned in the works of 1888” (Young 2010, 546).
Young speculates that Nietzsche’s “intellectual integrity required him to reject his original,
8
Magnus speculates that Hollingdale’s actual source for the story must be a “similar but by no means
identical tale” in a 1908 work entitled Franz Overbeck und Friedrich Nietzsche: Eine Freundschaft by Carl
Bernoulli. For a discussion, see Magnus (1986, 88).
9
In letters from November 13 and 14, Nietzsche is still treating the Antichrist as “the first book of the
transvaluation” (Middleton 1996, 324). But in a letter from November 20 to Brandes, Nietzsche refers to the
Antichrist and says “the ‘transvaluation of all values’ . . . lies finished before me” (Middleton 1996, 326); a
letter of November 26 to Deussen makes a similar claim. Based on this type of evidence, Mazzoni Montinari
argues that Nietzsche abandoned his plan for a four-volume series in mid-November 1888 (Montinari 1982,
92-119).
10
For an overview of Nietzsche’s notebooks during the 1884–8 period, see for example Magnus (1986),
Hollingdale (1999, 217–27), and Young (2010, 534–50).
11
The published works do not provide sufficient evidence for the metaphysical interpretations of will to
power, as discussed by Heidegger and others. But I have not relied on that material; I have focused on the
psychological reading of will to power.
all-embracing system”—that is, prevented him from the allegedly reductive attempt to find a
basic principle for all action (Young 2010, 542). For, according to Young, in late 1888 Nietzsche
“becomes open to the variety of human motivations and no longer tries to force them all onto
the procrustean bed of the will to power” (Young 2010, 546).12
This conclusion strikes me as both extremely dubious and of only biographical interest. To
start with the latter point: suppose we assume that in his notebook writings of November 1888,
Nietzsche came to reject his own will to power thesis. (Suppose, for example, a note turned up in
which Nietzsche wrote “my will to power thesis was an error: values should not be evaluated in
terms of will to power!”) Then we would need to ask whether unpublished notebook material
from the last weeks of Nietzsche’s sane life should be taken as more important than Nietzsche’s
great 1883–8 works—from Zarathustra to Beyond Good and Evil to the Genealogy to Twilight and
the Antichrist—which clearly endorse the will to power thesis. It seems to me obvious that the
answer would be no.13
But this brings us to the first point: the claim that Nietzsche rejects the will to power thesis is
nothing more than speculation based on the thinnest of textual evidence. Two points bear
emphasizing. First, one can certainly question the inference from a changing title to a rejection of
a central doctrine. That a philosopher no longer intends to write a series of books focusing on a
particular topic does not imply that the philosopher has abandoned or modified his beliefs about
the topic. Perhaps Nietzsche came to feel that one book—The Antichrist—offered a sufficient
analysis of will to power. Or perhaps other problems had come to seem more interesting or more
gripping. It is an immense interpretive leap to infer that because Nietzsche might have abandoned
plans to write a four-book series focusing on will to power, he came to reject his will to power
doctrine.14
12
A word on Young’s rather bizarre claim that Nietzsche abandoned the will to power thesis because he
became aware of its inconsistency with the diversity of human motivation (Young 2010, 546). First, one
suspects that Nietzsche was aware of the diversity of human motivation before 1888. Second, as Chapter 6,
Section 2 explained, the will to power thesis does not conflict with the claim that human beings have diverse
motivations. Magnus offers a more nuanced view, arguing that Nietzsche “abandoned the will to power and
eternal recurrence as cosmological principles,” but not as psychological principles (Magnus 1986, 84; emphasis
added).
13
Consider an analogy: suppose we discover that, in the last weeks of his life, Kant penned a note reading
“I no longer believe that my Categorical Imperative is the supreme principle of morality.” Would this show
that Kantians should stop talking about the CI, that they should abandon their reliance on the Groundwork
and the Second Critique? Of course not. It would be an interesting biographical detail, but it would not affect
the evaluation of the Kantian arguments. Just so with Nietzsche.
14
Moreover, Nietzsche’s notebooks reveal that he is constantly shifting and modifying his plans,
rearranging and shuffling material under different headings and titles. I see no reason to attach any great
importance to the titles of future books or the projected lengths of future series. Had Nietzsche’s sanity lasted
a few more weeks, we might very well have found yet another projected table of contents for a multi-book
Revaluation of Values. Bernard Reginster makes a related point, noting that Nietzsche wrote at least 25 plans
for this work. As Reginster puts it, although there are “significant differences among these plans . . . it is their
broad structural similarities that I find most striking. Most of the plans require (1) an examination of the
nature and history of European nihilism; (2) a critique of the dominant values, particularly what are referred
to as Christian and moral values; (3) a revaluations of these values, which takes the will to power as its
principle; and, finally, (4) the doctrine of the eternal recurrence . . . The order and the manner in which those
themes are treated vary from one plan to the next, but these four issues retain their place and their basic
significance throughout Nietzsche’s revisions” (Reginster 2006, 17).
Second, the evidence for the claim that Nietzsche regarded the Antichrist as the totality of the
Revaluation is highly questionable. To be sure, there are a few letters from mid and late November
1888 in which Nietzsche seems to refer to the Antichrist as the totality of the Revaluation. But, as
Thomas Brobjer points out in a recent study of Nietzsche’s notes,
when Nietzsche very carefully revised his EH manuscript—“weighing each word on a gold scale” (KSB 8:
letter to Köselitz, 9 December 1888)—in early December, he did not change the reference to A as “the first
book” . . . Still more important, when he revised his NCW in the latter part of December, which he had
begun writing and compiling on December 12, he then again refers to A as the “first book of the revaluation
of values.” (Brobjer 2010, 21–2)
So a few letters from mid to late November seem to treat the Antichrist as the totality of
the revaluation project, whereas notes and published material prior to mid November and
throughout December consistently treat the Antichrist as one part of a four-book series. The
idea that Nietzsche rejects his will to power thesis, then, is based on giving an overridingly
authoritative status to a few brief remarks from two weeks of Nietzsche’s life.
In sum: the claim that Nietzsche abandoned his will to power thesis turns out to be based
either on an apocryphal tale originating in Hollingdale’s biography or on a highly dubious
inference from mixed evidence about the number of volumes in a series that Nietzsche planned
to write. This is hardly the stuff on which to base a rejection of a central theme in a philosopher’s
published works.
3.2 Does Nietzsche view will to power as unimportant?

This brings us to the second question: is will to power a central theme in the published works?
Given the evidence adduced in Chapter 6, Section 1.1, this question might seem surprising.
Nevertheless, Brian Leiter has repeatedly claimed that Nietzsche views will to power as “unim-
portant” (Leiter forthcoming a, Section 3). Leiter argues that if Nietzsche really viewed will to
power as an important component of his philosophical thought, then
it is hard to understand why he says almost nothing about will to power . . . in the two major self-reflective
moments in the Nietzschean corpus: his last major work, Ecce Homo, where he reviews and assesses his life and
writings, including specifically all his prior books (EH III); and the series of new prefaces he wrote for the
Birth of Tragedy, Human, All-too-Human, Daybreak, and The Gay Science in 1886, in which he revisits his major
themes. That this putative “fundamental principle” [will to power] merits no mention on either occasion
suggests that its role in Nietzsche’s thought has been overstated. (Leiter 2002, 142; cf. Leiter forthcoming a,
Section 3)
In this passage, Leiter first claims that Nietzsche says “almost nothing” about will to power in his
autobiography Ecce Homo. He then notes that Nietzsche doesn’t mention will to power in the
series of 1886 prefaces for his pre-Zarathustra books. This, allegedly, casts doubt on the import-
ance of the will to power thesis.
However, both of these arguments are deeply problematic. Start with Leiter’s claim about the
1886 prefaces. In fact, it is completely unsurprising that Nietzsche’s prefaces for his pre-Zarathus-
tra books don’t mention will to power—for will to power is first introduced in Zarathustra itself!
The fact that the prefaces do not mention a topic which none of these works discuss is no evidence
against will to power’s importance.15
So let’s consider Leiter’s other argument: the claim that Ecce Homo says “almost nothing”
about will to power. Two points are relevant. First, Nietzsche wrote Ecce Homo as a retrospective
looking-back and summing-up of his developments. As he penned Ecce Homo, his notebooks are
full of plans for a volume entitled Revaluation of All Values or Will to Power. Given that he
intended to write a new work on will to power, it would make sense if he devoted no space to it
in his retrospective work Ecce Homo. (Indeed, in a letter of November 20, 1888, Nietzsche
describes Ecce Homo as a “prelude to the Revaluation of All Values”; see Middleton [1996, 326] for
a translation of the letter.)
But this brings us to a second point: the claim that Ecce Homo says “almost nothing” about will
to power is at best misleading. The phrase “will to power” [Wille zur Macht] occurs four times in
Ecce Homo, “power” occurs an additional five times, and “will of life” or “will to life” occurs
three times.16 In total, then, there are no less than twelve separate references to power and its
cognate terms in Ecce Homo.
Perhaps Leiter’s claim is that because will to power is mentioned only twelve times—or
because the precise phrase “Wille zur Macht” occurs only four times—it cannot be an important
topic for Nietzsche. But even this charitable reading won’t help Leiter. Suppose we take a given
topic that Leiter thinks is among Nietzsche’s central concerns, and ask how many times it is
mentioned in Ecce Homo. For example, Leiter notes that one of Nietzsche’s concerns is the
rejection of a certain conception of free will (Leiter 2002, 87–101). Discussions of free will occur
only four times in EH. Surely we cannot conclude, from this, that Nietzsche abandons or
considers unimportant his critique of free will.
In light of these considerations, I conclude that there are no good reasons for denying that will
to power is one of Nietzsche’s central concerns. His commitment to the will to power thesis is
clear throughout his late works.
3.3 Does Nietzsche reject the idea that each action aims at power?
This brings us to a final interpretive objection. I have relied on the claim that will to power is
present in every action. However, Brian Leiter has argued that Nietzsche in fact rejects this claim.
Leiter writes,
Nietzsche repeatedly makes claims inconsistent with the thesis that “the essence of life” is will to power. So,
for example, he writes: “Life itself is to my mind the instinct for growth, for durability, for accumulation
of forces, for power: where the will to power is lacking there is decline. It is my contention that all the
supreme values of mankind lack this will” (A 6). But if all actions manifested this will, then this will could
never be found lacking. Yet Nietzsche thinks it can be lacking, which means he must countenance
the possibility that not all organic phenomena are will to power. (Leiter 2002, 141)
15
To be sure, some of Nietzsche’s 1886 prefaces discuss later developments in his thought (the preface for
Birth of Tragedy is most notable in this regard). But they do not attempt to provide a synoptic overview of his
later thought.
16
“Wille zur Macht” occurs in EH Preface 4; EH III “Birth” 4; EH III “Case of Wagner” 1; EH IV.4.
“Macht” occurs in EH III “Zarathustra” 2, 3, and 7; EH III “Genealogy”; EH IV.1 and IV.7. The phrase “will
to life” occurs in EH II.1, EH III “Birth” 3, and EH III “Case of Wagner” 2.
In short, Leiter takes these two claims to be incompatible:

1) Each action manifests will to power, and
2) Certain actions lack will to power.
Nietzsche certainly accepts (2). Does this give us a reason to reject claim (1)?
In fact, it does not. There are two reasons to doubt Leiter’s argument. First, if the passage cited
above really were incompatible with the claim that all actions manifest will to power, then the
principle of charity would dictate interpreting it as a sloppy phrasing rather than Nietzsche’s
deliberate rejection of claim (1). After all, this is just one passage, whereas versions of claim (1)
occur in many places throughout Nietzsche’s texts (as Chapter 6 demonstrated).17
However, this brings us to the second problem: there is no need to dismiss the above passage.
Given our reconstructions above, we can see that claims (1) and (2) are in fact perfectly compatible.
Claim (2) should be interpreted not as the claim that certain actions express no will to power whatsoever,
but rather that certain actions manifest only minimal amounts of will to power. After all, when some
property X can be present to different degrees, we often use the phrase “X is lacking” to mean that
X is present only to a minimal extent. (The German word translated as “lacking” is “fehlen”, which
operates analogously.) For example, the claim “Bill’s athletic prowess is lacking” does not mean
that Bill has absolutely no ability to engage in athletic endeavors, but that his ability to do so is
below average. The claim “Tom’s cooking abilities are lacking” does not mean that Tom has
absolutely no capacity to cook, but that his ability to do so is limited. Just so, claim (2) indicates that
the will to power is present to a low degree, rather than completely absent.18
For these reasons, there is no conflict between Nietzsche’s claim that will to power is lacking
and his claim that will to power is omnipresent. Nietzsche does, indeed, assert that every action
manifests will to power.
4. Conclusion
Above, I have argued that central objections to and misgivings about my reading of Nietzsche as
a constitutivist are groundless. I hope to have shown that there are no straightforward or obvious
17
Leiter also cites two additional passages, but neither passage claims that the will to power is absent. First,
he notes that Nietzsche writes, “wherever the will to power declines in any form” (A 17). However, to
decline is not to be absent: the claim that certain actions manifest diminished or declining amounts of will to
power is certainly compatible with the claim that all actions manifest will to power. Second, Leiter cites
Nietzsche’s claim that liberal institutions “undermine the will to power” (TI IX.38). Again, the claim that
something is undermined does not suggest that it is absent. An athlete’s aim of completing a race may be
undermined by a sudden knee injury, but this hardly means that he did not aim to complete the race.
18
Maudemarie Clark offers a different argument against will to power. She interprets will to power as “a
second-order desire for the ability to satisfy one’s other, or first-order, desires” (1990, 211). She argues that
this must be an empirical generalization: we should read Nietzsche as showing that will to power is “an
important human motive” (1990, 212). However, like Leiter, Clark denies that this motive is omnipresent:
while she concedes that “in calling attention to this motive, Nietzsche does illuminate large areas of human
life and behavior,” Clark says that she “resist[s], however, the idea that Nietzsche believed that all behavior is
motivated by the desire for power because I do not see any way in which this could be a plausible or
interesting hypothesis about human behavior” (1990, 212). I agree with Clark that an omnipresent second-
order desire to fulfill first-order desires would be neither plausible nor interesting; but I take this to be a
problem with Clark’s interpretation of will to power, rather than with Nietzsche’s theory. For I aim to show
that if we interpret will to power as the aim of encountering and overcoming resistance, the will to power
thesis is both plausible and interesting.
CONCLUSION 253
reasons for rejecting my constitutivist reading of Nietzsche. Of course, producing a full-fledged

defense of this reading of Nietzsche would require a level of textual work necessitating a volume
of its own. I haven’t produced that volume. I’ve simply argued that (1) if we interpret Nietzsche
as a constitutivist, then we can resolve a generations-long puzzle about the status of his ethical
critiques; (2) the constitutivist interpretation is consonant not just with a stray remark here and
there in Nietzsche’s texts, but with his most abiding philosophical concerns and some central
features of his works; and (3) the obvious objections to interpreting Nietzsche as a constitutivist
can be answered. I think this strongly suggests that Nietzsche’s ethical theory is, indeed, a version
of constitutivism.
References
Allison, Henry (1990), Kant’s Theory of Freedom. New York: Cambridge University Press.
Anderson, R. Lanier (2005), “Nietzsche on Truth, Illusion, and Redemption,” European Journal
of Philosophy 13(2): 185–225.
Annas, Julia (1995), The Morality of Happiness. Oxford: Oxford University Press.
Anscombe, G. E. M. (1958), “Modern Moral Philosophy,” Philosophy 33: 1–18.
—— (2000), Intention. Cambridge, MA: Harvard University Press.
Ansell-Pearson, Keith (1994), An Introduction to Nietzsche as Political Thinker. Cambridge: Cam-
bridge University Press.
Appel, Fredrick (1998), Nietzsche Contra Democracy. Ithaca: Cornell University Press.
Arneson, Richard (2009), “Egalitarianism,” The Stanford Encyclopedia of Philosophy (Spring 2009
Edition), ed. Edward N. Zalta, <http://plato.stanford.edu/archives/spr2009/entries/egalitar-
ianism/>.
Arpaly, Nomy (2004), Unprincipled Virtue. Oxford: Oxford University Press.
Augustine (1991), Confessions, trans. R. S. Pine-Coffin. New York: Penguin Classics.
Aydin, Ciano (2007), “Nietzsche on Reality as Will to Power: Toward an ‘Organization-
Struggle’ Model,” Journal of Nietzsche Studies 33: 25–48.
Bargh, J. A., and T. L. Chartrand (1999), “The Unbearable Automaticity of Being,” American
Psychologist 54: 462–79.
Barnes, Jonathan (1984), The Complete Works of Aristotle. Princeton: Princeton University Press.
Baron, Marcia (1995), Kantian Ethics Almost Without Apology. Ithaca: Cornell University Press.
Baumeister, Roy, Alfred R. Mele, and Kathleen D. Vohs (eds.) (2010), Free Will and Conscious-
ness: How Might They Work? New York: Oxford University Press.
Bennett, R. L., A. G. Motulsky, A. Bittles, L. Hudgins, S. Uhrich, D. L. Doyle, K. Silvey,
C. R. Scott, E. Cheng, B. Mcgillivray, R. D. Steiner, and D. Olsen (2002), “Genetic
Counseling and Screening of Consanguineous Couples and their Offspring: Recommenda-
tions of the National Society of Genetic Counselors,” Journal of Genetic Counseling 11: 97–119.
Bentham, Jeremy (2007), An Introduction to the Principles of Morals and Legislation. New York:
Dover.
Bergmann, Frithjof (1994), “Nietzsche and Analytic Ethics,” in Richard Schacht (ed.), Nietzsche,
Genealogy, Morality. Berkeley: University of California Press, 76–94.
Bigg, Matthew (2008), “Elitism Charge Hurts in U.S. Politics,” Reuters (April 16, 2008 edition),
<http://www.reuters.com/article/idUSN1547078420080416>.
Blackburn, Simon (2001), Ruling Passions. Oxford: Oxford University Press.
Bloomfield, Paul (2001), Moral Reality. New York: Oxford University Press.
Bower, G. H. (1981), “Mood and Memory,” American Psychologist 36: 129–48.
Bradburn, Norman (1969), The Structure of Psychological Well-Being. New York: Walter de
Gruyter.
Bratman, Michael (1999), Faces of Intention. New York: Cambridge University Press.
—— (2007), Structures of Agency. Oxford: Oxford University Press.
Brewer, Talbot (2002), “Maxims and Virtues,” Philosophical Review 111: 539–72.
REFERENCES 255
Brink, David O. (1986), “Externalist Moral Realism,” Southern Journal of Philosophy, 24: 23–40.
Brobjer, Thomas (1995), Nietzsche’s Ethics of Character: A Study of Nietzsche’s Ethics and its Place in
the History of Moral Thinking. Uppsala: Uppsala University Press.
—— (2010), “The Origin and Early Context of the Revaluation Theme in Nietzsche’s
Thinking,” Journal of Nietzsche Studies 39: 12–29.
Broome, John (1999), “Normative Requirements,” Ratio 12(4): 398–419.
Buss, Sarah, and Lee Overton (2002), Contours of Agency. Cambridge, MA: MIT Press.
Clark, Maudemarie (1990), Nietzsche on Truth and Philosophy. Cambridge: Cambridge University
Press.
—— (1994), “Nietzsche’s Immoralism and the Concept of Morality,” in Richard Schacht (ed.),
Nietzsche, Genealogy, Morality. Berkeley: University of California Press, 15–34.
—— (1998), “On Knowledge, Truth, and Value: Nietzsche’s Debt to Schopenhauer and the
Development of his Empiricism,” in Christopher Janaway (ed.), Willing and Nothingness:
Schopenhauer as Nietzsche’s Educator. Oxford: Oxford University Press.
—— (1999), “Nietzsche’s Antidemocratic Rhetoric,” Southern Journal of Philosophy 37 (Supple-
mentary Issue 1): 119–41.
Clark, Maudemarie, and David Dudrick (2009), “Nietzsche on the Will: An Analysis of BGE
19,” in Ken Gemes and Simon May (eds.), Nietzsche on Freedom and Autonomy. Oxford:
Oxford University Press, 247–68.
Clark, M. S., and B. A. Waddell (1983), “Effects of Moods on Thoughts about Helping,
Attraction and Information Acquisition,” Social Psychology Quarterly 46: 31–5.
Clark, Philip (2001), “Velleman’s Autonomism,” Ethics 111: 580–93.
Clarke, Samuel (1705), A Discourse Concerning the Unchangeable Obligations of Natural Religion, and
the Truth and Certainty of the Christian Revelation: The Boyle Lectures, 1705, in J. B. Schneewind
(ed.), Moral Philosophy from Montaigne to Kant. Cambridge: Cambridge University Press, 2002,
295–310.
Cohen, G. A. (1996),“Reason, Humanity, and the Moral Law,” in Christine M. Korsgaard (ed.),
The Sources of Normativity. Cambridge: Cambridge University Press, 167–88.
—— (2001), Karl Marx’s Theory of History: A Defence, 2nd edition. Oxford: Oxford University
Press.
Conway, Daniel (1997), Nietzsche’s Dangerous Game. Cambridge: Cambridge University Press.
Copp, David, and David Sobel (2004), “Morality and Virtue: An Assessment of Some Recent
Work in Virtue Ethics,” Ethics 114(3): 514–54.
Cordner, Christopher (1994), “Aristotelian Virtue and Its Limitations,” Philosophy 69: 291–316.
Cox, Christoph (1999), Nietzsche: Naturalism and Interpretation. Berkeley and Los Angeles:
University of California Press.
Crisp, Roger (2006), “Aristotle on Greatness of Soul,” in Richard Kraut (ed.), The Blackwell
Guide to Aristotle’s Nicomachean Ethics. Oxford: Blackwell, 158–78.
Csikszentmihalyi, Mihaly (1988), Optimal Experience. New York: Cambridge University Press.
—— (1990), Flow: The Psychology of Optimal Experience. New York: Harper & Row.
Cullity, Garrett, and Berys Gaut (eds.) (1997), Ethics and Practical Reason. Oxford: Oxford
University Press.
Cuypers, Stefaan E. (2000), “Autonomy Beyond Voluntarism: In Defense of Hierarchy,”
Canadian Journal of Philosophy 30: 225–56.
—— (2001), Self-Identity and Personal Autonomy. Burlington, VT: Ashgate.
256 REFERENCES
Dancy, Jonathan (2002), Practical Reality. Oxford: Oxford University Press.

Davidson, Donald (1980), Essays on Actions and Events. Oxford: Clarendon Press.
Dean, John W. (2008), “Is Senator Barack Obama Truly too Elite to be Elected President?,”
Findlaw (May 16, 2008, edition), <http://writ.news.findlaw.com/dean/20080516.html>.
Deci, Edward, and Richard Ryan (1985), Intrinsic Motivation and Self-Determination in Human
Behavior. New York: Springer.
Doris, John M. (2010), The Moral Psychology Handbook. Oxford: Oxford University Press, 2010.
Dostoevsky, Fyodor (2006), Notes from Underground. New York: Bantam Classics.
Dreier, James (1997), “Humean Doubts about the Practical Justification of Morality,” in Garrett
Cullity and Berys Gaut (eds.), Ethics and Practical Reason. Oxford: Oxford University Press,
81–100.
Dworkin, Ronald (1977), Taking Rights Seriously. Cambridge, MA: Harvard University Press.
Dyzenhaus, D. (1992), “John Stuart Mill and the Harm of Pornography,” Ethics 102: 534–51.
Easton, Susan M. (1994), The Problem of Pornography. London: Routledge.
Enoch, David (2006), “Agency, Schmagency: Why Normativity Won’t Come from What is
Constitutive of Action,” Philosophical Review 115: 169–98.
—— (2009), “How is Moral Disagreement a Problem for Realism?,” Journal of Ethics 13: 15–50.
—— (2011), “Schmagency Revisited,” in Michael Brady (ed.), New Waves in Metaethics
(New York: Palgrave Macmillan), 208–33.
Falvey, Kevin (2000), “Knowledge in Intention,” Philosophical Studies 99: 21–44.
Ferrero, Luca (2009), “Constitutivism and the Inescapability of Agency,” in Russ Shafer-Landau
(ed.), Oxford Studies in Metaethics, Volume 4. Oxford: Oxford University Press, 303–33.
Feshbach, S., and R. D. Singer (1957), “The Effects of Fear Arousal and Suppression of Fear
upon Social Perception,” Journal of Abnormal and Social Psychology 55: 283–8.
Fitzpatrick, William (2005), “The Practical Turn in Korsgaard’s Ethical Theory,” Ethics 115:
651–91.
Foley, Richard (1977), “Deliberate Action,” Philosophical Review 86: 58–69.
Foot, Philippa (2003), Natural Goodness. Oxford: Oxford University Press.
Forgas, J. P. (1994), “The Role of Emotion in Social Judgments: An Introductory Review and
an Affect Infusion Model (AIM),” European Journal of Social Psychology 24: 1–24.
Forgas, J. P., and P. Vargas (2004), “The Effects of Mood on Social Judgment and Reasoning,”
in Michael Lewis and Jeannette M. Haviland-Jones (eds.), The Handbook of Emotions, 2nd
edition. New York: Guilford Press, 350–67.
Frankfurt, Harry (1987), “Equality as a Moral Ideal,” Ethics 98: 21–42.
—— (1988), The Importance of What We Care About. New York: Cambridge University Press.
—— (1999), Necessity, Volition, and Love. New York: Cambridge University Press.
—— (2004), The Reasons of Love. Princeton: Princeton University Press.
—— (2006), Taking Ourselves Seriously and Getting it Right. Stanford: Stanford University Press.
Freud, Sigmund (1957), The Standard Edition of the Complete Psychological Works of Sigmund Freud,
ed. J. Strachey. London: Hogarth Press.
Frierson, Patrick (2005), “Kant’s Empirical Account of Human Action,” Philosophers’ Imprint
5(7): 1–34.
Garner, Richard (1990), “On the Genuine Queerness of Moral Properties and Facts,” Australa-
sian Journal of Philosophy 68: 137–46.
REFERENCES 257
Gass, William (1971), “The Stylization of Desire,” The New York Review of Books, 16.3 (February
25).
Gemes, Ken, and Simon May (eds.) (2009), Nietzsche on Freedom and Autonomy. Oxford: Oxford
University Press.
Gerhardt, Volker (1996), Vom Willen zur Macht. Anthropologie und Metaphysik der Macht am
exemplarishcen Fall Friedrich Nietzsches. Berlin: Walter de Gruyter.
Geuss, Raymond (1997), “Nietzsche and Morality” European Journal of Philosophy 5(1): 1–20.
Gibbard, Allan (2003), Thinking How to Live. Cambridge, MA: Harvard University Press.
Ginet, Carl (1990), On Action. New York: Cambridge University Press.
Gollwitzer, Peter, and Paschal Sheeran (2006), “Implementation Intentions and Goal Achive-
ment: A Meta-Analysis of Effects and Processes,” Advances in Experimental Social Psychology 38:
69–119.
Griswold, Charles (1999), Adam Smith and the Virtues of Enlightenment. Cambridge: Cambridge
University Press.
Haidt, J. (2001), “The Emotional Dog and its Rational Tail: A Social Intuitionist Approach to
Moral Judgment,” Psychological Review 108: 814–34.
Hamermesh, Daniel, and Amy Parker (2005), “Beauty in the Classroom: Instructors’ Pulchritude
and Putative Pedagogical Productivity,” Economics of Education Review 24: 369–76.
Hampshire, Stuart (1949), “Fallacies in Moral Philosophy,” Mind 58: 466–82.
Hampton, Jean (1998), The Authority of Reason. Cambridge: Cambridge University Press.
Harman, Gilbert (1977), The Nature of Morality: An Introduction to Ethics. Oxford: Oxford
University Press.
Hassin, R., J. Uleman, and J. Bargh (eds.) (2005), The New Unconscious. Oxford: Oxford
University Press.
Hatab, Lawrence (1995), A Nietzschean Defense of Democracy. Chicago: Open Court.
Hegel, G. W. F (1991), Elements of the Philosophy of Right, trans. H. B. Nisbet. Cambridge:
Cambridge University Press.
Heidegger, Martin (1979), Nietzsche, trans. David F. Krell, 4 vols. New York: Harper & Row.
Hobbes, Thomas (1994), Leviathan, ed. Edwin Curley. Indianapolis: Hackett Publishing.
Hofstadter, Richard (1966), Anti-Intellectualism in American Life. New York: Vintage.
Hollingdale, R. J. (1999), Nietzsche: The Man and His Philosophy. New York: Cambridge
University Press.
Holton, Richard (2009), Willing, Wanting, Waiting. Oxford: Oxford University Press.
Hornsby, Jennifer (2004), “Agency and Actions,” in H. Steward and J. Hyman (eds.), Agency and
Action. Cambridge: Cambridge University Press, 1–23.
Hubin, Donald (2001), “The Groundless Normativity of Instrumental Reason,” Journal of
Philosophy 98: 445–68.
Hume, David (1987), Essays Moral, Political, and Literary, ed. E. F. Miller. Indianapolis: Liberty
Press.
Hunt, Lester (1991), Nietzsche and the Origin of Virtue. New York: Routledge.
Hurka, Thomas (2007), “Nietzsche: Perfectionist,” in Brian Leiter and Neil Sinhababu (eds.),
Nietzsche and Morality. Oxford: Oxford University Press, 9–31.
Hursthouse, Rosalind (1999), On Virtue Ethics. Oxford: Oxford University Press.
258 REFERENCES
Hussain, Nadeem (2011), “The Role of Life in Nietzsche’s Genealogy,” in Simon May (ed.),
Nietzsche’s On the Genealogy of Morality: A Critical Guide. Cambridge: Cambridge University
Press, 142–69.
Huxley, T. H., and William J. Youmans (1868), The Elements of Physiology and Hygiene: A Text-
book for Educational Institutions. New York: D. Appleton and Company.
Hyman, John, and Helen Steward (eds.) (2004), Agency and Action. New York: Cambridge
University Press.
Isen, Alice (1984), “Towards Understanding the Role of Affect in Cognition,” in R. S. Wyer and
T. K. Srull (eds.), Handbook of Social Cognition Volume 3. Hillsdale, NJ: Erlbaum, 179–236.
—— (1987), “Positive Affect, Cognitive Processes and Social Behaviour,” in L. Berkowitz (ed.),
Advances in Experimental Social Psychology, Volume 20. San Diego, CA: Academic Press, 203–53.
James, William (1890), The Principles of Psychology. New York: Dover.
Janaway, Christopher (2007), Beyond Selflessness: Reading Nietzsche’s Genealogy. Oxford: Oxford
University Press.
Jaspers, Karl (1997), Nietzsche: An Introduction to the Understanding of his Philosophical Activity, trans.
C. F. Walraff and F. J. Schmitz. Baltimore: Johns Hopkins University Press.
Joyce, Richard (2006), The Evolution of Morality. Cambridge, MA: MIT Press.
Kahneman, Daniel, and Amos Tversky (1981), “The Framing of Decisions and the Psychology
of Choice,” Science 211: 453–8.
Kane, Robert (1996), The Significance of Free Will. Oxford: Oxford University Press.
Kant, Immanuel (1996), The Metaphysics of Morals, ed. Mary Gregor. New York: Cambridge
University Press.
—— (1998), Groundwork of the Metaphysics of Morals, ed. Mary Gregor. New York: Cambridge
University Press.
—— (1999a), Critique of Practical Reason, ed. Paul Guyer and Allen Wood. New York:
—— (1999b), Religion within the Boundaries of Mere Reason, ed. Allen Wood and George di
Giovanni. New York: Cambridge University Press.
Katsafanas, Paul (2005), “Nietzsche’s Theory of Mind: Consciousness and Conceptualization,”
European Journal of Philosophy 13: 1–31.
—— (2011a), “The Concept of Unified Agency in Nietzsche, Plato, and Schiller,” Journal of the
History of Philosophy 49(1): 87–113.
—— (2011b), “Deriving Ethics from Action: A Nietzschean Version of Constitutivism,”
Philosophy and Phenomenological Research 83(3): 620–60.
—— (2011c), “Activity and Passivity in Reflective Agency,” in Russ Shafer-Landau (ed.),
Oxford Studies in Metaethics, Volume 6. Oxford: Oxford University Press, 219–54.
—— (2011d), “The Relevance of History for Moral Philosophy,” in Simon May (ed.),
Nietzsche’s On the Genealogy of Morality: A Critical Guide. Cambridge: Cambridge University
Press, 170–92.
—— (forthcoming a), “Nietzsche’s Philosophical Psychology,” in John Richardson and Ken
Gemes (eds.), The Oxford Handbook of Nietzsche. Oxford: Oxford University Press.
—— (forthcoming b), “Value, Affect, Drive,” in Manuel Dries and Peter Kail (eds.), Nietzsche on
Mind and Nature. Oxford: Oxford University Press.
Kaufmann, Walter (1974), Nietzsche: Philosopher, Psychologist, Antichrist. Princeton: Princeton
University Press.
REFERENCES 259
Kierkegaard, Sren (1983), The Sickness unto Death, trans. H. V. Hong and E. H. Hong.
Princeton: Princeton University Press.
Korsgaard, Christine M. (1996a), Creating the Kingdom of Ends. Cambridge: Cambridge Univer-
sity Press.
—— (1996b), The Sources of Normativity. Cambridge: Cambridge University Press.
—— (1997), “The Normativity of Instrumental Reason,” in Garrett Cullity and Berys Gaut
(eds.), Ethics and Practical Reason. Oxford: Oxford University Press, 215–54.
—— (1999), “Self-Constitution in the Ethics of Plato and Kant,” Journal of Ethics 3: 1–29.
—— (2008), The Constitution of Agency: Essays on Practical Reason and Moral Psychology. Oxford:
Oxford University Press.
—— (2009), Self-Constitution: Agency, Identity, and Integrity. Oxford: Oxford University Press.
—— (2010), “Reflections on the Evolution of Morality,” The Amherst Lecture in Philosophy 5:
1–29.
Kuhn, T. S. (1970), The Structure of Scientific Revolutions. Chicago: University of Chicago Press.
Kunda, Ziva (1990), “The Case for Motivated Reasoning,” Psychological Bulletin 108: 480–98.
Langton, Rae (2004), “Intention as Faith,” in J. Hyman and H. Steward (eds.), Agency and Action.
New York: Cambridge University Press, 243–58.
Laudan, Larry (1977), Progress and Its Problems: Toward a Theory of Scientific Growth. London:
Routledge & Kegan Paul.
Lavin, Douglas (2004), “Practical Reason and the Possibility of Error,” Ethics 114: 424–57.
Lear, Jonathan (2006), Radical Hope: Ethics in the Face of Cultural Devastation. Cambridge, MA:
Harvard University Press.
Leibniz, G. W., and S. Clark (2000), G. W. Leibniz and Samuel Clark: Correspondence, ed. Roger
Ariew. Indianapolis: Hackett Publishing.
Leiter, Brian (1997), “Nietzsche and the Morality Critics,” Ethics 107: 250–85.
—— (2000), “Nietzsche’s Metaethics: Against the Privilege Readings,” European Journal of
Philosophy 8: 277–97.
—— (2001), “The Paradox of Fatalism and Self-Creation in Nietzsche,” in J. Richardson and
B. Leiter (eds.), Nietzsche. Oxford: Oxford University Press, 281–321.
—— (2002), Nietzsche on Morality. London: Routledge.
—— (2006), “The Hermeneutics of Suspicion: Recovering Marx, Nietzsche, and Freud,” in
Brian Leiter (ed.), The Future for Philosophy. Oxford: Oxford University Press, 74–105.
—— (2007), “Nietzsche’s Theory of the Will,” Philosophers’ Imprint 7(7): 1–15.
—— (2010), “Nietzsche’s Moral and Political Philosophy,” The Stanford Encyclopedia of Philosophy
(Summer 2010 Edition), ed. Edward N. Zalta, <http://plato.stanford.edu/archives/sum2010/
entries/nietzsche-moral-political/>.
—— (forthcoming a), “Nietzsche’s Naturalism Reconsidered,” in John Richardson and Ken
Gemes (eds.), The Oxford Handbook of Nietzsche. Oxford: Oxford University Press.
—— (forthcoming b), “Moral Skepticism and Moral Disagreement in Nietzsche.”
Leiter, Brian, and Neil Sinhababu (eds.) (2007), Nietzsche and Morality. Oxford: Oxford Univer-
sity Press.
Libet, B., C. A. Gleason, E. W. Wright, and D. K. Pearl (1983), “Time of Conscious Intention
to Act in Relation to Onset of Cerebral Activity (Readiness-Potential): The Unconscious
Initiation of a Freely Voluntary Act,” Brain 106: 623–42.
Locke, John (1975), An Essay Concerning Human Understanding. Oxford: Oxford University Press.
260 REFERENCES
Loeb, Don (1998), “Moral Realism and the Argument from Disagreement,” Philosophical Studies
90(3): 282–303.
Loeb, Paul (1995), “Is There a Genetic Fallacy in Nietzsche’s Genealogy of Morals?” International
Studies in Philosophy 27(3): 125–41.
McCarty, Richard (2009), Kant’s Theory of Action. Oxford: Oxford University Press.
McDowell, John (2001), Mind, Value, and Reality. Cambridge, MA: Harvard University Press.
Mackie, John (1977), Ethics: Inventing Right and Wrong. New York: Pelican Books.
Magnus, Bernd (1986), “Nietzsche’s Philosophy in 1888: The Will to Power and the Über-
mensch,” Journal of the History of Philosophy 24(1): 79–98.
Marlowe, C., S. Schneider, and C. Nelson (1996), “Gender and Attractiveness Biases in Hiring
Decisions,” Journal of Applied Psychology 81: 11–21.
May, Simon (1999), Nietzsche’s Ethics and his War on ‘Morality’. Oxford: Oxford University Press.
—— (ed.) (2011), Nietzsche’s On the Genealogy of Morality: A Critical Guide. Cambridge:
Melden, A. I. (1961), Free Action. London: Routledge & Kegan Paul.
Mele, Alfred (2003), Motivation and Agency. Oxford: Oxford University Press.
—— (2009), Effective Intentions: The Power of Conscious Will. Oxford: Oxford University Press.
Middleton, Christopher (1996), Selected Letters of Friedrich Nietzsche. Indianapolis: Hackett
Publishing.
Mill, John Stuart (2002), Utilitarianism. Indianapolis: Hackett Publishing.
Millgram, Elijah (1995), “Was Hume a Humean?” Hume Studies 21(1): 75–93.
Molden, D. C., and E. T. Higgins (2005), “Motivated Thinking,” in K. Holyoak and
R. Morrison (eds.), The Cambridge Handbook of Thinking and Reasoning. New York: Cambridge
University Press, 295–318.
Montinari, Mazzoni (1982), Nietzsche Lesen. Berlin: Walter de Gruyter.
Moore, G. E. (1971), Principia Ethica. Cambridge: Cambridge University Press.
Moore, Gregory (2002), Nietzsche, Biology, and Metaphor. New York: Cambridge University
Press.
Müller-Lauter, Wolfgang (1971), Nietzsche: Seine Philosophie der Gegensatze und die Gegensatze
seiner Philosophie. Berlin: Walter de Gruyter.
Muraven, Mark, and Roy F. Baumeister (2000), “Self-Regulation and Depletion of Limited
Resources: Does Self-Control Resemble a Muscle?” Psychological Bulletin 126: 247–59.
Nagel, Thomas (1991), Equality and Partiality. Oxford: Oxford University Press.
Nehamas, Alexander (1985), Nietzsche: Life as Literature. Cambridge, MA: Harvard University
Press.
—— (1999), “Nietzsche and Hitler,” Southern Journal of Philosophy 37: 1–17.
Nozick, Robert (1981), Philosophical Explanations. Cambridge, MA: Belknap Press.
O’Shaughnessy, Brian (1973), “Trying (As the Mental ‘Pineal Gland’),” Journal of Philosophy 70
(13): 365–86.
—— (1980), The Will: A Dual-Aspect Theory, 2 vols. Cambridge: Cambridge University Press.
Owen, David (2002), “Equality, Democracy, and Self-Respect: Reflections on Nietzsche’s
Agonal Perfectionism,” Journal of Nietzsche Studies 24: 113–31.
—— (2007), Nietzsche’s Genealogy of Morality. Montreal: McGill-Queen’s University Press.
—— (2008), “Nietzsche, Ethical Agency and the Problem of Democracy,” in Herman Siemens
and Vasti Roodt (eds.), Nietzsche, Power and Politics. Berlin: Walter de Gruyter, 143–68.
REFERENCES 261
Owens, David (2003), “Does Belief Have an Aim?” Philosophical Studies 115: 283–305.
Parfit, Derek (1997), “Reason and Motivation,” Proceedings of the Aristotelian Society Supplementary
Volume 71: 99–146.
—— (2006), “Normativity,” in Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, Volume 1.
Oxford: Oxford University Press, 325–80.
—— (2011), On What Matters. Oxford: Oxford University Press.
Phelps, E. A. (2005), “The Interaction of Emotion and Cognition: The Relation Between the
Human Amygdala and Cognitive Awareness,” in R. Hassin, J. Uleman, and J. Bargh (eds.),
The New Unconscious. Oxford: Oxford University Press, 61–76.
Pippin, Robert (2009), “Natural and Normative,” Daedelus 138: 35–43.
—— (2010), Nietzsche, Psychology, and First Philosophy. Chicago: University of Chicago Press.
Pockett, Susan, William P. Banks, and Shaun Gallagher (eds.) (2009), Does Consciousness Cause
Behavior? Cambridge, MA: MIT Press.
Poellner, Peter (1995), Nietzsche and Metaphysics. Oxford: Oxford University Press.
—— (2007), “Affect, Value, and Objectivity,” in Brian Leiter and Neil Sinhababu (eds.),
Nietzsche and Morality. Oxford: Oxford University Press, 227–61.
Popper, Karl (1959), The Logic of Scientific Discovery. London: Hutchinson.
Prinz, Jesse (2007), The Emotional Construction of Morals. Oxford: Oxford University Press.
Railton, Peter (1997), “On the Hypothetical and Non-Hypothetical in Reasoning about Belief
and Action,” in Garrett Cullity and Berys Gaut (eds.), Ethics and Practical Reason. Oxford:
Oxford University Press, 53–80.
Rawls, John (1993), Political Liberalism. New York: Columbia University Press.
—— (1999), “Two Concepts of Rules,” in S. Freeman (ed.), Collected Papers. Cambridge, MA:
Harvard University Press, 20–46.
Raz, Joseph (1999), Engaging Reason. Oxford: Oxford University Press.
—— (2005), “The Myth of Instrumental Rationality,” Journal of Ethics and Social Philosophy 1:
1–28.
Reath, Andrews (2006), Agency and Autonomy in Kant’s Moral Theory. Oxford: Oxford University
Press.
Reginster, Bernard (1997), “Nietzsche on Ressentiment and Valuation,” Philosophy and Phenom-
enological Research 57: 281–305.
—— (2006), The Affirmation of Life. Cambridge, MA: Harvard University Press.
—— (2007), “The Will to Power and the Ethics of Creativity,” in Brian Leiter and Neil
Sinhababu (eds.), Nietzsche and Morality. Oxford: Oxford University Press, 32–56.
Reid, Thomas (1983), Essays on the Active Powers of Man, in The Works of Thomas Reid, D.D., ed.
Sir William Hamilton. Hildesheim: G. Olms Verlagsbuchhandlung, 511–680.
Richardson, John (1996), Nietzsche’s System. Oxford: Oxford University Press.
—— (2004), Nietzsche’s New Darwinism. Oxford: Oxford University Press.
—— (forthcoming), “Nietzsche on Life’s Ends.”
Richardson, John, and Ken Gemes (forthcoming), The Oxford Handbook of Nietzsche. Oxford:
Oxford University Press.
Ridley, Aaron (1998), Nietzsche’s Conscience: Six Character Studies from the Genealogy. Ithaca:
Cornell University Press.
Risse, Mathias (2003), “Nietzsche’s ‘Joyous and Trusting Fatalism’,” International Studies in
Philosophy 35(3): 147–63.
262 REFERENCES
Risse, Mathias (2007), “Nietzschean ‘Animal Psychology’ versus Kantian Ethics,” in Brian Leiter
and Neil Sinhababu (eds.), Nietzsche and Morality. Oxford: Oxford University Press, 57–82.
Robertson, Simon (2011), “Normativity for Nietzschean Free Spirits,” Inquiry 54: 591–613.
Rosati, Connie (2003), “Agency and the Open Question Argument,” Ethics 113: 490–527.
Ross, W. D. (1930), The Right and the Good. Oxford: Clarendon Press.
—— (1939), The Foundations of Ethics. Oxford: Clarendon Press.
Ruben, David Hillel (1997), “The Active and the Passive,” Aristotelian Society Supplementary
Volume 71(1): 229–46.
Sanitioso, R., Z. Kunda, and G. T. Fong (1990), “Motivated Recruitment of Autobiographical
Memory,” Journal of Personality and Social Psychology 59: 229–41.
Sarch, Alexander (2008), “What’s Wrong with Megalopsychia?,” Philosophy 83: 231–53.
Scanlon, T. M. (2000), What We Owe to Each Other. Cambridge, MA: Harvard University Press.
—— (2003), “Metaphysics and Morals,” Proceedings and Addresses of the American Philosophical
Association 77: 7–22.
—— (2009), “Being Realistic About Reasons,” available at: <http://www.philosophy.ox.ac.
uk/lectures/john_locke_lectures/past_lectures>.
Schacht, Richard (1983), Nietzsche. London: Routledge.
—— (ed.) (1994), Nietzsche, Genealogy, Morality. Berkeley and Los Angeles: University of
California Press.
Schapiro, Tamar (2001), “Three Conceptions of Action in Moral Theory,” Noûs 35: 93–117.
—— (2009), “The Nature of Inclination,” Ethics 119: 229–56.
Scheffler, Samuel (1992), Human Morality. Oxford: Oxford University Press.
Schick, Frederic (2003), Ambiguity and Logic. New York: Cambridge University Press.
Schmeichel, B. J., K. D. Vohs, and R. F. Baumeister (2003), “Intellectual Performance and Ego
Depletion: Role of the Self in Logical Reasoning and Other Information Processing,” Journal
of Personality and Social Psychology 85: 33–46.
Schneewind, J. B. (1998), The Invention of Autonomy. Cambridge: Cambridge University Press.
—— (ed.) (2002), Moral Philosophy from Montaigne to Kant. Cambridge: Cambridge University
Press.
Schopenhauer, Arthur (1969), The World as Will and Representation, trans. E. F. J. Payne. New
York: Dover Publications.
—— (1974), On the Fourfold Root of the Principle of Sufficient Reason, trans. E. F. J. Payne. Illinois:
Open Court Publishing.
—— (1999), Prize Essay on the Freedom of the Will, trans. E. F. J. Payne. New York: Cambridge
University Press.
—— (2001), Parerga and Paralipomena, trans. E. F. J. Payne. Oxford: Oxford University Press.
Schroeder, Mark (2007), Slaves of the Passions. Oxford: Oxford University Press.
Schroeder, Timothy (2004), Three Faces of Desire. Oxford: Oxford University Press.
Searle, John (1983), Intentionality. New York: Cambridge University Press.
Setiya, Kieran (2003), “Explaining Action,” Philosophical Review 112: 339–93.
—— (2007), Reasons without Rationalism. Princeton: Princeton University Press.
Shafer-Landua, Russ (2003), Moral Realism: A Defense. Oxford: Oxford University Press.
Shah, Nishi (2003), “How Truth Governs Belief,” Philosophical Review 112: 447–82.
Shah, Nishi, and J. David Velleman (2005), “Doxastic Deliberation,” Philosophical Review 114:
497–534.
REFERENCES 263
Shaw, Tasmin (2007), Nietzsche’s Political Skepticism. Princeton: Princeton University Press.
Sidgwick, Henry (1981), The Methods of Ethics. Indianapolis: Hackett Publishing.
Siemens, H. W. (2009), “Nietzsche’s Critique of Democracy (1870–1886),” Journal of Nietzsche
Studies 38: 20–37.
Singer, Peter (2005), “Ethics and Intuitions,” Journal of Ethics 9: 331–52.
Sinhababu, Neil (2007), “Vengeful Thinking and Moral Epistemology,” in Brian Leiter and Neil
Sinhababu (eds.), Nietzsche and Morality. Oxford: Oxford University Press, 262–80.
Sinnott-Armstrong, Walter (2007), Moral Psychology, Volume I: The Evolution of Morality: Adap-
tions and Innateness. Cambridge, MA: MIT Press.
Sinnott-Armstrong, Walter, Liane Young, and Fiery Cushman (2010), “Moral Intuitions,” in
John M. Doris (ed.), The Moral Psychology Handbook. Oxford: Oxford University Press,
246–72.
Smith, Michael (1994), The Moral Problem. Oxford: Blackwell.
—— (2012), “Beyond the Error Theory,” in Kieran Setiya and Hille Paakkunainen (eds.),
Internal Reasons: Contemporary Readings. Cambridge, MA: MIT Press, 303–28.
Stampe, Dennis (1987), “The Authority of Desire,” Philosophical Review 96: 335–81.
Stanovich, Keith E. (2004), The Robot’s Rebellion: Finding Meaning in the Age of Darwin. Chicago:
University of Chicago Press.
Staten, Henry (1990), Nietzsche’s Voice. Ithaca: Cornell University Press.
Stegmaier, Werner (1994), Nietzsches “Genealogie der Moral.” Darmstadt: Wissenschaftliche
Buchgesellschaft.
—— (2000), “Nietzsches Zeichen,” Nietzsche-Studien 29: 41–69.
Suits, Bernard (2005), The Grasshopper: Games, Life, and Utopia. Orchard Park, NY: Broadview
Press.
Svavarsdóttir, Sigrún (1999), “Moral Cognitivism and Motivation,” Philosophical Review 108:
161–219.
—— (2006), “How Do Moral Judgments Motivate?” in J. Dreier (ed.), Contemporary Debates in
Moral Theory. Oxford: Blackwell, 163–81.
Swann, W. B. (1983), “Self-Verification: Bringing Social Reality into Harmony with the Self,”
in J. Suls and A. G. Greenwald (eds.), Psychological Perspectives on the Self, Volume 2. Hillsdale,
NJ: Erlbaum, 2, 33–66.
Taylor, Charles (1992), The Sources of the Self. Cambridge, MA: Harvard University Press.
Temkin, Larry (1986), “Inequality,” Philosophy and Public Affairs 15: 99–121.
Thalberg, Irving (1978), “Mental Activity and Passivity,” Mind 87(347): 376–95.
Thiele, Leslie Paul (1990), Friedrich Nietzsche and the Politics of the Soul. Princeton: Princeton
University Press.
Thomson, Judith Jarvis (2008), Normativity. Chicago: Open Court Publishing.
Tubert, Ariela (2010), “Constitutive Arguments,” Philosophy Compass 5: 656–66.
—— (2011), “Korsgaard’s Constitutive Arguments and the Principles of Practical Reason,”
Philosophical Quarterly 61: 343–62.
Velleman, J. David (1989), Practical Reflection. Princeton: Princeton University Press.
—— (2000), The Possibility of Practical Reason. Oxford: Oxford University Press.
—— (2004a), “Précis of The Possibility of Practical Reason,” Philosophical Studies 121: 225–38.
—— (2004b), “Replies to Discussion on The Possibility of Practical Reason,” Philosophical Studies
121: 277–98.
264 REFERENCES
Velleman, J. David (2006), Self to Self. New York: Cambridge University Press.
—— (2009), How We Get Along. New York: Cambridge University Press.
Vohs, Kathleen D. (2010), “Free Will is Costly: Action Control, Making Choices, Mental Time
Travel, and Impresion Management Use Precious Volitional Resources,” in Roy Baumeister,
Alfred Mele, and Kathleen Vohs (eds.), Free Will and Consciousness: How Might They Work?
Oxford: Oxford University Press, 66–81.
Wallace, R. Jay (2006), Normativity and the Will. Oxford: Oxford University Press.
—— (2007), “Ressentiment, Value, and Self-Vindication: Making Sense of Nietzsche’s Slave
Revolt,” in Brian Leiter and Neil Sinhababu (eds.), Nietzsche and Morality. Oxford: Oxford
University Press, 110–37.
Watson, Gary (2004), Agency and Answerability. Oxford: Oxford University Press.
Weber, Max (2002), The Protestant Ethic and the Spirit of Capitalism, trans. Peter Baehr and Gordon
C. Wells. New York: Penguin.
Wedgwood, Ralph (2002), “The Aim of Belief,” Philosophical Perspectives 16: 267–97.
—— (2007), The Nature of Normativity. New York: Oxford University Press.
Wegner, Daniel (2002), The Illusion of Conscious Will. Cambridge, MA: Harvard University Press.
Wilcox, John (1974), Truth and Value in Nietzsche. Ann Arbor, MI: University of Michigan Press.
Wilde, Oscar (1998), Lady Windermere’s Fan. New York: Dover Publications.
Williams, Bernard (1981), Moral Luck. New York: Cambridge University Press.
—— (1986), Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press.
—— (1993), Shame and Necessity. Berkeley and Los Angeles: University of California Press.
—— (1994), “Nietzsche’s Minimalist Moral Psychology,” in R. Schacht (ed.), Nietzsche,
Genealogy, Morality. Berkeley and Los Angeles: University of California Press, 237–48.
Wilson, George (2002), “Action,” The Stanford Encyclopedia of Philosophy (Summer 2002 Edition),
ed. E. N. Zalta, <http://plato.stanford.edu/archives/sum2002/entries/action/>.
Wilson, Timothy D. (2002), Strangers to Ourselves: Discovering the Adaptive Unconscious. New
York: Belknap Press.
Witt, J. K., D. R. Proffitt, and W. Epstein (2004), “Perceiving Distance: A Role of Effort and
Intent,” Perception 33: 577–90.
Wolff, Jonathan (2002), Why Read Marx Today? Oxford: Oxford University Press.
Wood, Allen (1984), “Kant’s Compatibilism,” in Allen Wood (ed.), Self and Nature in Kant’s
Philosophy. Ithaca: Cornell University Press, 73–101.
—— (1990), Hegel’s Ethical Thought. Cambridge: Cambridge University Press.
—— (1999), Kant’s Ethical Thought. Cambridge: Cambridge University Press.
Wuerth, Julian (forthcoming), “Sense and Sensibility in Kant’s Practical Agent: Against the
Intellectualism of Korsgaard and Sidgwick,” European Journal of Philosophy.
Young, Julian (2010), Friedrich Nietzsche: A Philosophical Biography. Cambridge: Cambridge
University Press.
Zhu, Jing (2004), “Passive Action and Causalism,” Philosophical Studies 119(3): 295–314.
Index
action: Cohen, G. A. 51 n. 4
attributability to agents 87–9, 93, 96–8, 100–1, compassion 147–8, 186–7, 195–6
111–13 constitutive aims, principles, and standards 1–2,
distinguished from mere behavior 50, 71, 81, 37–46
87–9, 111–13 definition of 39
goodness or badness of 61–7, 98–108, 146 difference between constitutive aims and
Kant’s theory of 16, 18, 35–7, 43–5, 87, 95–6, constitutive standards 45, 86–7
115–36, 143–4, 202, 239, 241–2 as inescapable, see inescapability
Nietzsche’s theory of 156–82, 184–203, normative status of 48–53, 56–61, 197–200
206–9 see also action, goodness or badness of;
active, see agential activity constitutivism; differentially realizable aim;
activity, see agential activity schmagency objection; success, standard of;
affects 34, 166–7, 187 n. 10 value
affirmation, see approval constitutivism 1–2, 37–46, 107–10, 197–200,
agential activity 109–44 238–42
as a constitutive aim of action 143, 145–6, Korsgaard’s version of 86–108
203–4, 206–10, 241–2 Nietzschean version of 109–242
as equilibrium 136–42 objections to 47–67
Anscombe, G. E. M. 50, 69 n., 78 n., 79–81 relation to Humeanism and Kantianism 43–6
approval: and the three challenges for ethical theory 41–2
of a constitutive aim 59–60, 204, 206–9 Velleman’s version of 68–85
of one’s action 138–42, 145–6, 204–9 see also action, goodness or badness of; agential
Aristotle 9–10, 26, 30–2 activity; constitutive aims, principles, and
Arneson, Richard 219 standards; differentially realizable aim;
asceticism 158, 195 success, standard of; value; will to power
Augustine 141–2 Cuypers, Stefaan 131 n., 138 n. 39, 209 n. 48
autonomy 86, 96 n., 98 n. 12, 111, 241–2
Davidson, Donald 50, 88
Bentham, Jeremy 160, 172, 220, 236 n. defective action, see action, goodness or badness of
boredom 172–3 deliberation 111–44, 202–3, 244–5
Bratman, Michael 113 democracy 219–30
Broome, John 44 n. 78, 50 n. 2, 52 desires:
distinguished from drives 169–70, 180
Categorical Imperative 3, 35–7, 43, 45, 87–9, and reasons 32–5, 43–4, 87, 89–95, 164–5
95–8, 99 n. 16, 100–8, 183 n., 188, relation to agential activity 111–44
241–2 relation to happiness or satisfaction 172–6
challenges for ethical theory 25–37, 41–2 role in Kant’s theory 87
epistemological 8–16 role in Velleman’s theory 69–78
metaphysical 17–19 see also motivation
practical 19–25 diachronic stability 93–7, 100–1
chess: differentially realizable aim 75, 100, 107–8
example of constitutive features 1–2, 38–41, distinguished from simple aim 75
49–51, 58–9, 179, 198 in Korsgaard’s theory 100–7
choice: in the Nietzschean theory 200–3
and its relation to activity 116–44 in Velleman’s theory 75–8, 81–4
influences upon 118–19, 125–32 disagreement, argument from 14–16
Kant’s account of 91–8, 117–24, 143 disequilibrium 138–42
Clark, Maudemarie 166, 187 n. 10, 196 n. 33, Dostoevsky, Fyodor 173 n. 46, 242
221 n., 252 n. 18 drives 165–76, 178, 180
Clark, Philip 61, 98 n. 14 dual-process model of agency 124, 133
266 INDEX
egalitarianism 213, 218–31, 240 as a constitutivist 43–6, 86–108

endorsement 209 on normativity 3, 6, 16, 22 n., 35–7, 61 n.,
Enoch, David 16 n. 28, 47, 49–57, 63 n. 14, 188, 202, 228, 241–2
103 n. Kierkegaard, Sren 22, 156
equality 192, 212, 219–24, 229–30 Korsgaard, Christine 2–3, 36 n. 63, 42–3, 45,
see also egalitarianism 51 n. 4, 54, 63–7, 85–108, 110–11, 114,
equilibrium: 117–18, 143, 183 n., 202, 209 n. 48, 241–2
as an analysis of agential activity 136–42
as a constitutive aim 143, 145–6 Lavin, Douglas 99 n. 17
see also agential activity Leibniz, G. W. 117
executive control 123–4 Leiter, Brian 14 n. 25, 16 n. 27, 18 n. 32, 154,
155 n., 223–4, 232 n. 27, 245, 247–52
Ferrero, Luca 55 n. 8 Locke, John 117–21, 131, 136, 239
Frankfurt, Harry 21, 112–13, 125–7, 131–2, Loeb, Paul 148 n. 2
138 n. 39, 209 n. 48, 230–1 n. 24
freedom 90–1, 96, 117–19, 120–4 Mackie, John 14–17, 26 n. 44
Freud, Sigmund 74, 168 Magnus, Bernd 248, 249 n. 12
function argument 30–2, 63–5 maxims 15, 18, 35–6, 95, 118 n. 18
May, Simon 201
Garner, Richard 17 Mele, Al 116 n. 11, 123 n. 24, 124 n. 26
genealogy 8–16 Mill, John Stuart 43–4, 160, 164–5, 172,
genetic fallacy 11, 148 224 n., 236 n.
Goethe, Johann Wolfgang von 185, 213–15, Moore, G. E. 25, 57–60
232–4 moral realism, non-reductive 25–30, 37
good: motivation 109–44
good action, see action, goodness or badness of relation to agential activity 133–42
standards for, see constitutivism; constitutive relation to normativity 19–25, 39–41, 47–67,
aims, principles, and standards 208–9
two models of its relation to reflection 125–32
Hampshire, Stuart 128 see also action; agential activity; constitutive
happiness 160–1, 171–6, 220–1, 236 aims, principles, and standards;
Harman, Gilbert 13–14 constitutivism; drives; reflection; suspension
health 149–52 of motives
hedonism 160–2, 172–3, 195 n. 31
Hegel, G. W. F. 21–2, 37 n., 156 naturalism 17–19, 28–9, 31, 36, 42
Hobbes, Thomas 116 n. 12, 120–1 Nehamas, Alexander 157 n. 18, 247 n. 35
Hollingdale, R. J. 247–50 nihilism 21–5
Hume, David 32–5, 37, 43–6, 60, 244 normativity, see action, goodness or badness of;
Hursthouse, Rosalind 30–1 constitutive aims, principles, and standards;
Hussain, Nadeem 151 n. 11 constitutivism; inescapability; value
Nozick, Robert 138 n. 37
inescapability 40, 47–67, 83, 106, 188, 206–8
instincts, see drives open question argument 57–61
instrumental reason, principle of 44–6, 50–3, overcoming, see resistance; will to power
160–1, 178 Owen, David 224 n. 15, 225, 227, 229
internalism requirement 19–21
intrinsic standards of success 39–40 pain 169–74, 193–6, 205–6, 214–15, 236
intuitions 9–10, 12, 26–8, 35, 42, 85, 185 Parfit, Derek 25–6, 29–30
intuitionism 26–7 particularistic willing 96 n. 10
passivity, see agential activity
James, William 120–1 Pippin, Robert 25 n. 42, 27, 245–6
Janaway, Christopher 18 n. 32, 149 n. 7, 165 n., pity, see compassion
171 n. 40, 228 n. 20 pleasure 160–1, 167, 172–6, 193–4
Jaspers, Karl 23 n. 39, 213–14 see also happiness; pain
power, see will to power
Kant, Immanuel: practical challenge 19–25
on action 18–19, 36, 42, 117–24, 131–2, and Aristotelianism 31–2
136, 143 and constitutivism 41–2
INDEX 267
and Humeanism 33 self-understanding, see self-knowledge

and Kantianism 36 Setiya, Kieran 47, 78 n. 13
and non-reductive realism 29–30 Shaw, Tamsin 197 n. 36, 218 n. 12
Smith, Michael 33, 35, 45 n. 80
queerness, argument from 17–18 success, standard of 39–41, 44–5, 56–7, 60–2,
145–6, 162, 165, 184–91, 204, 208–9
Raz, Joseph 50 n. 1, 113 suicide 217–18 n. 45
realism, non-reductive 25–30 Suits, Bernard 175 n. 50
reason see action, goodness or badness of; suffering, see pain
constitutive aims, principles, and standards; suspension of motives 117–19, 125–32, 133–6
constitutivism; inescapability; value
reflection Temkin, Larry 222–3
in Korsgaard’s theory 89–91
its relation to motivation and agency 111–42 unity, in Korsgaard’s theory 87–107
Reginster, Bernard 157, 159, 171 n. 41, 174,
178 n., 193 n. 25, 194 n. 27, 215 n. 5, value:
234 n. 30, 235 n. 32, 245 n. 3, 249 n. 14 using constitutive standards to assess 185–200
on will to power 157–9 deriving versus assessing 188–90, 240
resistance 146, 156–76 see also action, goodness or badness of;
its connection to drives 168–71 constitutive aims; constitutivism; reasons
its connection to happiness 171–6, 193–4 Velleman, David 2–3, 42 n. 75, 43, 45, 68–87,
its connection to other motives 199 111, 113–14, 115 n. 7, 117, 133 n. 33,
the kinds we seek 176–8 199 n. 40, 241
its relation to reasons 184–91
see also will to power, its normative status Wallace, Jay 113–15, 117
whether all actions aim at resistance 178–81 Wedgwood, Ralph 20, 29 n. 48
and will to power 156–62 will to live 204–6
see also will to power will to power:
ressentiment 13, 233, 235 n. 31 as a constitutive aim 164–76
revaluation 147–56, 163, 186–9, 191, 197 and egalitarianism 211–13, 218–231
Richardson, John 34 n. 62, 161, 221 n., 230 n. 22 and the interpretive problem 152–6, 162–3
Risse, Mathias 131 n., 233 n. 28 its nature 156–62, 176–8
Rosati, Connie 57–9 and Nietzsche’s ethical exemplars 231–7
normative conclusions generated by 183–203,
satisfaction 171–6 211–37
see also happiness its normative status 147–52
Scanlon, T. M. 19 n. 34, 25, 27, 28 n. 45, and the problem of differential
29 n. 48, 30, 187 n. 11 realizability 200–3
schmagency objection, 53–6 its relation to agential activity 203–9
Schopenhauer, Arthur 67, 165n., 166–8, 172–3, as a standard for assessing values 185–200
204–9 and universalism 214–18
self-consciousness 89–92, 96–7, 114–15, whether Nietzsche’s views on it change
116 n. 11, 245 246–51
and its relation to motivation 117–42 whether present in all actions, 178–81
self-constitution 86–108 Williams, Bernard 21 n. 36, 32–3, 245–6
self-knowledge 216
as a constitutive aim 68–85 Young, Julian 247–50

Katsafanas P Agency Foundations of Ethics Nietzschean Constitutivism 13 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Katsafanas P Agency Foundations of Ethics Nietzschean Constitutivism 13 PDF

Uploaded by

Copyright:

Available Formats

Agency and the Foundations of Ethics

This page intentionally left blank

An earlier version of Chapter 5 appeared as “Activity and Passivity in Reﬂective

Reference to Nietzsche’s Works x

Appendix: Is Nietzsche Really a Constitutivist? 243

Reference edition of Nietzsche’s works

List of abbreviations of Nietzsche’s works

succumb to a common problem: each theory can generate substantive normative

Thus, if we accept a Nietzschean account of agency—as I argue that we should—

1. Three challenges for ethical theory

Below, we will ask whether Nietzsche is right.

1.2 The metaphysical challenge

rely on “an imaginary natural science (anthropocentric; no trace of any concept of

This implies that objective goodness is merely something constructed by my conviction,

‘Nothing is worthwhile! You shall not will!’ ” (Z III.12; cf. GS 1, GS 125, GM

2. Assessing ethical theories in light of these challenges

instead appeal to reﬂective equilibrium.45 Rather than seeking to ground morality in

ﬁnd fantastical. Indeed, contemporary realists often acknowledge the counterintuitive

justifying morality through intuition of irreducible normative truths won’t inspire

This argument is notoriously difﬁcult. Chapter 4 will examine a version of it in more

Epistemological Metaphysical Practical

Let’s be more precise. We can deﬁne constitutive aim as follows:

(Constitutive Aim) Let A be a type of attitude or event. Let G be a goal.

(Success) If X aims at G, then G is a standard of success for X.

3.3 Constitutivism’s relationship to Humeanism and Kantianism

1. First objection: the standards of a practice are

(2) Constitutive Aims as Transferring Reasons: If you have reason to participate in A,

1.2 Why the answer to our question doesn’t matter

2. Second objection: agency, schmagency

3. Third objection: inescapability does not

(Success) If X aims at G, then G is a standard of success for X.

4. Fourth objection: the mere fact that we have an aim

constitutivist needs in order to generate normative results. So long as we accept

5. Fifth objection: constitutivism cannot account

Constitutivists derive standards from an account of action’s constitutive features.

1. Velleman’s argument that action constitutively aims

However, Velleman’s claim becomes more plausible when we consider a peculiar

(10 ) If an agent intentionally A’s, then

2. Velleman’s attempt to derive an account of reasons

2.4 Velleman’s theory yields only a simple constitutive aim

constitutive aim of self-knowledge.12 As a result, Velleman’s theory of action does not

3. Does an appeal to improvisational actors or a quest

In other words, an agent performs an action, as opposed to producing mere behavior,

doing-what-would-make-sense-for-his-character will lead him, in conjunction with

4. Can we supplement the constitutivist theory

A more general problem is that if a constitutivist theory requires a supplementary

1. Korsgaard’s Kantian version of constitutivism

2. Stage Two: explication and defense of (B)

2.5 Principles determine the degree of unity

3. Stage Three: from agential unity to the

compromises my diachronic stability—then I am not really uniﬁed at all. (For this

4. Korsgaard’s argument for constitutivism

(Constitutive Principle) A has constitutive principle P iff there is some principle

How can a constitutivist avoid this problem? By incorporating a differentially

(S2) If an agent is committed to performing partial actions, then she is committed to

context, stipulating that we are committed to aiming at maximal unity is circular. So

In short: a constitutivist theory of practical reason can only be as convincing as the

1. Agential activity and its alleged connection

A number of philosophers endorse Frankfurt’s general claim that we must distin-

And the examples could be multiplied.5

2. Teasing apart several distinct claims about

(Choice) Typically, if I am faced with two actions that it is possible for me to

(Non-determination) Choice is not determined by motives. Typically, if we hold