Believing Against the Evidence

The question of whether it is ever permissible to believe on insufficient evi-

dence has once again become a live question. Greater attention is now being
paid to practical dimensions of belief, namely issues related to epistemic
virtue, doxastic responsibility, and voluntarism.
In this book, McCormick argues that the standards used to evaluate
beliefs are not isolated from other evaluative domains. The ultimate crite-
ria for assessing beliefs are the same as those for assessing action, because
beliefs and actions are both products of agency. Two important implications
of this thesis, both of which deviate from the dominant view in contempo-
rary philosophy, are 1) it can be permissible (and possible) to believe for
non-evidential reasons, and 2) we have a robust control over many of our
beliefs, a control sufficient to ground attributions of responsibility for belief.

Miriam Schleifer McCormick is Associate Professor in the Department of

Philosophy at the University of Richmond, Virginia.
Routledge Studies in Contemporary Philosophy

For a full list of titles in this series, please visit www.routledge.com

29 Kant and Education 37 Contemporary Feminist

Interpretations and Commentary Pragmatism
Edited by Klas Roth and Edited by Maurice Hamington
Chris W. Surprenant and Celia Bardwell-Jones

30 Feminism, Psychoanalysis, and 38 Morality, Self Knowledge, and

Maternal Subjectivity Human Suffering
Alison Stone An Essay on the Loss of
Confidence in the World
31 Civility in Politics and Education Josep Corbi
Edited by Deborah S. Mower,
Wade L. Robison 39 Contrastivism in Philosophy
Edited by Martijn Blaauw
32 Philosophical Inquiry into
Pregnancy, Childbirth, and 40 Aesthetics after Metaphysics
Mothering From Mimesis to Metaphor
Maternal Subjects Miguel de Beistegui
Edited by Sheila Lintott and
Maureen Sander-Staudt 41 Foundations of Freedom
Welfare-Based Arguments against
33 Authenticity as an Ethical Ideal
Paternalism
Somogy Varga
Simon R. Clarke
34 The Philosophy of Curiosity
Ilhan Inan 42 Pittsburgh School of Philosophy
Sellars, McDowell, Brandom
35 Self-Realization and Justice Chauncey Maher
A Liberal-Perfectionist Defense
of the Right to Freedom from 43 Reference and Structure in the
Employment Philosophy of Language
Julia Maskivker A Defense of the Russellian
Orthodoxy
36 Narrative Identity, Autonomy, Arthur Sullivan
and Mortality
From Frankfurt and MacIntyre to 44 Civic Virtue and the Sovereignty
Kierkegaard of Evil
John J. Davenport Derek Edyvane
45 Philosophy of Language and 55 Reframing the Intercultural
Webs of Information Dialogue on Human Rights
Heimir Geirsson A Philosophical Approach
Jeffrey Flynn
46 Disagreement and Skepticism
Edited by Diego E. Machuca 56 How History Matters to
Philosophy
47 Philosophy in Schools Reconsidering Philosophy’s Past
An Introduction for Philosophers After Positivism
and Teachers Robert C. Scharff
Edited by Sara Goering, Nicholas
J. Shudak, and Thomas E. 57 The Affordable Care Act
Wartenberg Decision
Philosophical and Legal
48 A Philosophy of Material Implications
Culture Edited by Fritz Allhoff and Mark
Action, Function, and Mind Hall
Beth Preston
58 Realism, Science, and
49 A Philosophy of the Screenplay Pragmatism
Ted Nannicelli Edited by Kenneth R. Westphal

50 Race, Philosophy, and Film 59 Evidentialism and Epistemic

Edited by Mary K. Bloodsworth- Justification
Lugo and Dan Flory Kevin McCain

51 Knowledge, Virtue, and Action 60 Democracy in Contemporary

Essays on Putting Epistemic Confucian Philosophy
Virtues to Work David Elstein
Edited by Tim Henning and David
P. Schweikard 61 Deleuze and Pragmatism
Edited by Sean Bowden, Simone
52 The Ontology of Psychology Bignall, and Paul Patton
Questioning Foundations in the
Philosophy of Mind 62 Mind, Language and
Linda A.W. Brakel Subjectivity
Minimal Content and the Theory
53 Pragmatism, Law, and Language of Thought
Edited by Graham Hubbs and Nicholas Georgalis
Douglas Lind
63 Believing Against the
54 Contemporary Dualism Evidence
A Defense Agency and the Ethics of
Edited by Andrea Lavazza and Belief
Howard M. Robinson Miriam Schleifer McCormick
This page intentionally left blank
Believing Against the Evidence
Agency and the Ethics of Belief

Miriam Schleifer McCormick

First published 2015
by Routledge
711 Third Avenue, New York, NY 10017
and by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Routledge is an imprint of the Taylor & Francis Group,
an informa business
© 2015 Taylor & Francis
The right of Miriam Schleifer McCormick to be identified as author of this
work has been asserted by her in accordance with sections 77 and 78 of the
Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or
utilized in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in any
information storage or retrieval system, without permission in writing from
the publishers.
Trademark Notice: Product or corporate names may be trademarks or
registered trademarks, and are used only for identification and explanation
without intent to infringe.
Library of Congress Cataloging-in-Publication Data
McCormick, Miriam Schleifer, 1970–
Believing against the evidence : agency and the ethics of belief / Miriam
Schleifer McCormick. — 1 [edition].
pages cm. — (Routledge studies in contemporary philosophy ; 63)
Includes bibliographical references and index.
1. Belief and doubt. 2. Evidence. 3. Ethics. I. Title.
BD215.M395 2014
121'.6—dc23
2014022486
ISBN: 978-0-415-81884-1 (hbk)
ISBN: 978-0-203-57914-5 (ebk)

Typeset in Sabon
by Apex CoVantage, LLC
For Hannah and Joseph
This page intentionally left blank
Contents

Preface xi

Introduction 1

PART I
Doxastic Norms

1 Conceptual Defenses of Evidentialism 15

2 Normative Defenses of Evidentialism 37

3 Unity of Norms: A Defense of Pragmatism 52

PART II
Doxastic Responsibility

4 The Puzzle of Doxastic Responsibility 77

5 Responsibility without Voluntary Control 91

6 The Possibility of Doxastic Agency 108

Conclusion 129
Bibliography 135
Index 139
This page intentionally left blank
Preface

“belief is more properly an act of the sensitive, than of the cogitative

part of our natures.”
—David Hume, A Treatise of Human Nature.

When NPR is in the middle of a pledge drive and the battery is too low on
my phone to listen to music, I sometimes listen to a “Christian-oriented”
talk radio station. What I hear there is sometimes disturbing and madden-
ing. But most of what I experience when I listen is fascination.
When we look for examples of societies whose beliefs differ radically
from “ours,” we tend to look to distant places and times: pre-Copernican
Europeans, slave-owning American southerners, isolated hunter-gatherer
tribes. But the beliefs expressed on this talk radio station already represent
a society that seems deeply alien to me.
It is not the religious beliefs that I find disturbing or bizarre; these are
familiar enough when living in one of the most deeply religious countries in
the world. Rather, it is when highly contentious views are presented as facts
that I am amazed. Listeners are told there is no evidence at all for human-
caused climate change, that there are no similarities between the civil rights
movement and the gay rights movement, and that the theory of evolution
contradicts basic laws of nature.
Beliefs held by this “talk radio” society have led to book banning and
curriculum changes that reject the scientific view of the universe and of how
humans came to be in it. When I first had a student tell me that she doesn’t
believe in evolution, I was at a loss for how to respond. To me, that sounded
like someone telling me that she didn’t believe in gravity. It seemed both
irrational and wrong.
Our attitudes and reactions about beliefs are puzzling. Much of what
we believe seems beyond our control, and yet our beliefs also seem open to
normative assessment. I think, for example, that my student ought to believe
something different from what she does believe. But she did not choose
this belief, and it is not obvious that she can change it. So is my evaluation
xii Preface
justified? Is it appropriate for me to hold her responsible for this belief and
even reproach her for holding it?
Questions about how responsibility and agency relate to belief—that is,
questions about doxastic responsibility and doxastic agency—have become
prominent in philosophical discussions in the last few years. When I first
started to work on this project about seven years ago, the literature about
norms for belief and doxastic agency was small and manageable. Now a day
rarely goes by where I do not see something new coming out on related top-
ics. I am happy that there is a lively conversation of which I can be a part,
but it means that I will, inevitably, fail to engage with some important work.
I have tried, at least, to incorporate some of this abundant literature in notes.
I suspect that the increased prominence of beliefs that are unapologeti-
cally anti-intellectual and anti-science is partially responsible for the philo-
sophical attention now being focused on these issues. Normally when I find
myself disagreeing with someone, even about a fairly contentious topic, we
mostly agree about what counts in favor of believing something; for exam-
ple, that there is abundant and extensive scientific evidence supporting it.
But if someone believes that what I call “scientific evidence” does not count
in favor of believing something, it becomes much more difficult to see what
I can say to show what is wrong with that belief.
The journey that led me to an interest in these questions began with
David Hume, and especially with his skepticism. While Hume is mentioned
very little in this book, I see the view of belief I am putting forth as Humean
in spirit. Hume concludes that skepticism is irrefutable, and then turns to
a more interesting question: what does this imply for inquiry and belief?
Skeptical arguments, which date back to antiquity, can reveal the difficulty
of justifying our most basic beliefs. This inability to justify our foundational
principles throws into question whether any of our beliefs are justified. I do
not think that endeavors to refute skepticism can be successful, but I agree
with Hume that much of belief seems immune from skepticism. Even if we
grant that we have no justification for our belief in the external world, we
will, inevitably, continue to believe that an external world exists.
To understand why some beliefs remain even when they cannot be given
a rational justification, one needs to better understand belief. In particular,
one needs to pay attention to the complexity of the phenomenon of belief
and the breadth of its scope and purpose. I argue that beliefs behave much
more like emotions and desires than contemporary epistemologists typically
allow. Believing is not a purely intellectual, or “cogitative” act. Instead,
Hume is right that beliefs typically have an important affective (what Hume
calls sensitive) aspect. And recognizing this aspect can give us a better under-
standing of doxastic agency and help us to figure out how to interact with
those whose beliefs differ vastly from our own.
I can remember when I first decided to write this book, talking about it
to a friend on my kids’ elementary school playground. Now my children,
Hannah and Joseph, to whom this book is dedicated, are both adults. It has
 Preface xiii
been a long process and I have many people to thank for their encourage-
ment, insight, and assistance.
First, I want to thank Sarah Stroud and Alfred Mele, who both were
supportive and encouraging when I first thought about writing this book.
Al has also read versions of a number of the chapters over the years and
I have benefitted greatly from his comments and questions.
A number of people have provided me with extensive comments, without
which I could never have completed this work. Eugene Mills gave me inci-
sive and challenging comments on the first half of the book. I doubt I have
addressed all of his concerns, but I hope I have clarified some of what he
found confusing. Antonia Lolordo read, commented on, and edited the Intro-
duction, Conclusion, and Chapter 3. Her confidence that I could actually
complete this book has helped keep me going in the final stages of writing.
I am very thankful to Berislav Marusic, Daniel Whiting, and Mikhail
(Misha) Valdman; they all read and commented on an entire draft of the
book. Beri convinced me that I could not maintain the distinction of permis-
sible belief without evidence and permissible belief against evidence, and
so he is partially responsible for the title. Daniel’s comments, as well as his
own writing on these issues, helped me see where I needed take more care
in distinguishing between different defenses of evidentialism. Both Beri and
Daniel also helped point me to much of the most important relevant recent
material. I owe a special thanks to Misha. He is both an excellent philoso-
pher and a very good friend, and so his comments were often devastating.
I haven’t been able to deal with all his criticisms, but he has helped me see
the most vulnerable points in my position. Because we live in the same city,
we have had many chances to talk about a lot of the ideas in the book. These
conversations have been invaluable.
I have had a chance to present versions of some of this material and am
grateful to the audiences on those occasions. These include those at Univer-
sity of Virginia, James Madison University, Virginia Commonwealth Uni-
versity, the conference on “Responsible Belief in the Face of Disagreement”
at VU University in 2009, and the 5th Biennial Margaret Dauler Wilson
Conference in 2010. I also benefited from the VCU paper workshop where
we discussed a version of Chapter 2. Parts of the book incorporate material
from two of my published articles, “Taking Control of Belief,” Philosophi-
cal Explorations 14, (2011); and “Compelled Belief,” American Philosophi-
cal Quarterly 42 (2005). I am grateful to the editors and publishers for their
permission to use material from these essays.
I would also like to thank my colleagues at the University of Richmond,
and especially my chair, Nancy Schauber, who has done her best to lighten
my duties as I have been trying to finish this book. I am also grateful to my
students in two seminars on “The Ethics of Belief” and those in my semi-
nar on “Belief, Norms and Responsibility” held in the fall of 2013, where
we worked through most of what would end up being this book. A special
thanks to Benjamin Sales for his insightful questions and comments.
xiv Preface
Thanks to Rachel Atkinson for editing of early versions of this material,
George Wickham for his careful edits of the Introduction and Chapter 1,
and to Josh Kadrich for help in the final preparation of the manuscript, and
reminding me to breathe.
I want to thank some of my wonderful friends who have tolerated my
monomaniacal tendencies, and have been a constant support: Rachel, Anto-
nia, Misha, Tim, Susan, Abigail, Christina, Lee-Anne, George and Mary,
Lenny and Fred, Cat, and my brothers Avrom and Steven. Finally, thanks
to Lance for his love, acceptance, and excellent project management skills.
Introduction

1. THE ETHICS OF BELIEF AND THE NORMS OF AGENCY

The way we think and talk about beliefs reveals that our doxastic practices
are infused with normative judgments. For example, we express disapproval
and approval for each other’s beliefs; we ask in an incredulous tone, “How
can you believe that?” or exclaim, “What a ridiculous thing to believe!”
We seem to think that one’s actual belief can deviate from how one ought
to believe, just as we think one can act in a way that deviates from how one
ought to act. The broad question asked under the heading of “The Ethics of
Belief” is: What ought one believe? The dominant view among contemporary
philosophers is that the only good reasons for believing are evidential, namely
reasons based on evidence. I will call this view “evidentialism.” On this view,
the only legitimate criticism of belief is that it violates evidential norms and
any belief formed against the evidence is impermissible. I will use the term
“pragmatism” to refer to the opposed view that some non-evidentially based
beliefs are permissible and that doxastic norms are not wholly evidential.1
Pragmatists can allow that most beliefs that violate evidential norms are
impermissible but deny that the only relevant considerations when assessing
beliefs are evidential. One central aim of this book is to defend pragmatism
as I have here defined it.
One way of framing the question of what norms guide belief is to compare
them to the norms that guide action, which are often treated as unproblem-
atic; the question is whether the norms that guide belief are the same as,
related to, or wholly different from the norms that guide action. Of course,
the question of what norms guide action is not unproblematic. The entire
field of normative ethics would not exist if it were. But despite deep divisions
and debate about how to evaluate actions, broad agreement exists that if one
engages in practical reasoning, this should include thinking about the dictates
of morality and prudence. We should think about what principles guide our
actions, what the consequences of our actions are likely to be, and what our
actions say about our characters. It is difficult to provide a fully articulated
theory as to which principles matter most, or what the ultimate grounds
are for such principles. Some theorists think the project of articulating such
2 Introduction
general principles is misguided. But we at least know what area, broadly
speaking, to look in when making these practical determinations. I will refer
to the norms that guide action as the “norms of agency.”
If we were wondering what norms governed the game of chess, we would
not turn to the norms that guide action. We would need instead to examine
this specific practice, figure out how one plays the game, how one wins,
consult the rule book, or ask an expert. A particular decision I make when
playing the game might be based on moral or prudential considerations.
For example, one might decide not to take the queen yet because, in doing
so, one’s novice opponent would be embarrassed. But such considerations
are irrelevant in trying to determine what counts as a permissible chess move.
Most contemporary theorists think that the norms of belief are analogous
to the norms of chess; to appeal to the norms of agency in thinking about how
to believe, they think, is to make a category mistake. Beliefs are not actions
and so should not be assessed according to the same criteria. In assessing
a belief, the relevant criteria, it is argued, are alethic or epistemic. In believ-
ing, we seek to gain truth (or, more importantly, avoid falsehood), and so
when we believe for reasons that are opposed to truth-gaining or falsehood-
avoidance, we can be criticized for violating these norms. While there is
some disagreement about the precise relationship between belief and truth,
very few people fundamentally question the view that beliefs require their
own separate ethics.2 The central contention of this book is that they do not;
that, instead, the ethics of belief and action are unified. The norms of agency
apply to both action and belief.
In arguing for a disparity between the norms of action and belief, many the-
orists argue that to understand what norms guide a practice, one must inves-
tigate the aim or purpose of the practice. The norms provide rules that help
one achieve this aim. I assess your skill as a cyclist by appealing to standards
of ideal bicycle riding, for example being able to use your bicycle with maxi-
mum efficiency so that you expend minimal effort to travel far and fast. The
idea that one assesses an x based on x’s function is found clearly articulated in
Aristotle’s Nicomachean Ethics:

For just as the good, i.e., [doing] well for a flautist, a sculptor, and every
craftsman, and in general, for whatever has a function and [character-
istic] action, seems to depend on its function, the same seems to be true
for a human being, if a human being has some function.3

The idea of extending this teleological framework to humans in general is

something most contemporary theorists reject. Yet, extending this framework
to include beliefs has typically not been viewed as problematic. In assessing
whether someone is believing well, it is argued, we must look at the function
(aim, purpose) of belief. While we shall see this aim is characterized some-
what differently depending on the specific account given, most are variations
on the theme that beliefs aim at truth.
 Introduction 3
That belief aims at truth is supposed to be an explanatory hypothesis.4
First, if one thinks that a successful belief is a true one, then the truth-aim
hypothesis will explain why this is so. Just as an archer is successful when
his arrow hits the literal target, so, too, a believer is successful when his
belief hits the target of truth. The truth-aim hypothesis can also explain why
we cannot believe at will. We cannot choose what is true5 and if beliefs, in
some sense, are conceptually tied to truth, then this shows why we cannot
choose what to believe. In recent years, there have been many attempts to
illuminate the conceptual connection between belief and truth. A better
understanding of the truth-aim, it is argued, can make normative statements
about belief less mysterious. If it turns out to be a fact about our cognitive
systems that beliefs aim at truth, then it can also be a fact that false beliefs
are incorrect. Knowing the purpose for which a machine is designed allows
us to make normative claims about the machine. If my car won’t start and so
cannot serve its purpose of transporting me, something is wrong with my car.
So, it is argued, understanding the purpose of the “belief-forming machine”
allows us to assess how well or poorly the “machine” is functioning.
Thinking of believing as analogous to chess playing or bicycle riding is
problematic. What you believe is at least as central to who you are as how you
act (and, of course, how you act is connected in fundamental ways to what
you believe). Even if we eschew talk of a distinctive human function, we can
take from Aristotle that the best (most excellent, virtuous) human will always
act correctly, but such an ideal person will also always believe correctly and
feel correctly.6 The implications of accepting this unity between action, belief,
and feeling is one of the themes I will be exploring throughout this work.
I have been referring to “norms” for belief, but what do I mean when
I claim there are such norms and how, if at all, are these related to rules for
believing or to the aim or goal of belief? What concerns me when thinking
about norms for belief is primarily the criteria of assessment or evaluation
of belief. I am asking what criteria distinguish a belief being good or permis-
sible from a belief being bad or impermissible. One way of approaching this
question is to think about what the aim or goal is and then evaluate beliefs
according to how well they achieve this goal. I will be examining many such
approaches and arguing that they are flawed. This is not to say that such
reflection cannot help deepen our understanding of the doxastic norms, but
one cannot, as some argue, identify one aim or goal that then provides us with
the norm. Thinking about rules that tend to guide us in forming and main-
taining beliefs can also help in furthering our understanding of belief norms,
but they cannot be identified with them. Any rule explicitly articulated will
be an evidentialist one such as, “if one’s current evidence is against a proposi-
tion, one ought not believe it” but, I will argue, that it is possible for a belief
to be permissible even if it violates one of these rules. By contrast, it makes no
sense to say that a move in chess is permissible even if it violates the rules of
chess. Many evidentialists go wrong in thinking that evidentialist rules apply
absolutely, rather than in general.
4 Introduction
2. SOME HISTORICAL BACKGROUND

David Hume is one of the historical figures most commonly invoked in

defending evidentialism.7 Hume’s statement, “a wise man . . . proportions his
belief to the evidence.”8 is often taken to summarize the evidentialist view.
Hume does think, in general, that believing well means believing accord-
ing to evidentialist rules—what can be termed “the rules of the wise”—but
Hume, famously, recognizes that if one were to universally follow these
rules, one could find oneself with no beliefs at all. The belief that one should
proportion one’s belief to the evidence, for example, is not one that can,
without circularity, be evidentially grounded; this is the case with many of
our most basic framework beliefs. So, when Hume puts forth his evidential-
ist dicta, it is within this accepted, though ultimately rationally ungrounded,
framework. And the reason why we should proportion our belief to the
evidence is, I argue, for Hume, ultimately practical.
Some may take Hume’s evaluation of beliefs as purely descriptive. He has
described a prevalent and important human practice, namely the practice of
reasoning. This practice has developed with certain rules so that we can dis-
tinguish good reasoning from bad. We can say, according to the reasoning
game, that this belief is more warranted than that one, and that those who
follow the rules of the game correctly are epistemically responsible. That is,
we can say the “wise,” who play the reasoning game well, proceed in this
way and form beliefs on this basis. But it seems Hume wants to go beyond
mere description. He thinks it is better to follow reason, and strive to be
wise, than to stick with vulgar, unreasonable habits. What is the nature of
Hume’s approval for the wise person?
One answer to the question as to why we should regulate our beliefs
according to evidentialist rules is that doing so can provide us true beliefs or
knowledge. This answer does not take us very far. For it seems we can just
as easily ask the question, “Why should we want true beliefs?” as we can
ask, “Why should we be wise?” Instead, Hume’s preference for reason is
given a moral justification. The wise have the virtues of reasonableness and,
so, excite our moral approbation.9 According to Hume, a person’s virtue
“consists altogether in the possession of mental qualities, useful or agreeable
to the person himself or others.”10 He provides examples of each kind in
considering a paragon of virtue named Cleanthes. Cleanthes’s benevolence
is useful to others, his assiduousness useful to himself, his wit and gallantry
agreeable to others, and his tranquility of soul agreeable to himself.
Hume seems to think that one can locate the main source of approval
for the various mental qualities we call virtues as falling predominantly into
one of these four categories. The approval felt toward the mental quality of
the wise, it seems, stems more from the wise person’s character being useful
to society than from it being agreeable to the possessor or others, or even
useful to the possessor. I think Hume’s preference and recommendation for
following reason is politically motivated. The point is that the world will be
a better place if more people choose reason as their guide.
 Introduction 5
While Hume is taken as the evidentialist’s historic exemplar, he recognized
the limits of evidentialism, a recognition that eludes many contemporary
theorists. They form theories about belief that seem primarily aimed at clos-
ing off all gaps so that no non-evidentially based belief can sneak in as
legitimate. As we shall see, one strategy for such gap-sealing is to argue that
it is conceptually impossible to believe against the evidence.11
For much of the twentieth century, most philosophers seem to have
thought there is no question concerning norms for belief distinct from that
of what constitutes a belief’s justification. During that time when “The Ethics
of Belief” was discussed, it would usually refer historically to the nineteenth-
century debate between W. K. Clifford and William James. In Clifford’s paper
“The Ethics of Belief,” he insists that “it is wrong always, everywhere, and for
any one, to believe anything on insufficient evidence” and that we have a duty
to withhold beliefs for which we do not have evidence. In James’s “The Will
to Believe,” James responds directly to Clifford’s strong evidentialist stance.
He agrees that in many contexts evidential considerations will settle the mat-
ter of what to believe, but when questions cannot be decided by the evidence,
James says it is permissible to let our “passional nature” take over, and for
our beliefs to depend partly on what will help us make sense of ourselves and
our world, on what will provide us with meaning, or even on what will give
us peace and solace. Thus, for James, at least some of the norms governing
belief are practical.
In the past decade, this debate has been revived, and the question of
whether it is ever permissible to believe against the evidence has once again
become a live question. Though it is never simple to account for what brings
a question back into philosophical fashion, one likely reason for this revival is
that there was a perceived need to answer “Reformed epistemologists” who
defend religious belief by saying that beliefs are sometimes justified even if
one has no evidence for them.12 For example, Jonathan Adler explicitly states
that his motivation for his defense of a very strong version of evidentialism
came after engaging with these anti-evidentialist arguments.13 If Adler is
right that the concept of belief guarantees the truth of evidentialism, then
the guiding question of the ethics of belief is misleading. There is no question
about what I ought to believe beyond what I must believe; to say I believe
something though I lack evidence for it, Adler says, is incoherent. But the
price Adler pays for this victory is that he has committed us to widespread
error in many of our doxastic practices.14 An alternative is to allow that
some beliefs are not based on evidence and then figure out when such beliefs
are pernicious and when they are not. So doing would allow us to respond
to the anti-evidentialist arguments Adler considers without committing us to
the view that our doxastic practices are fundamentally confused.15
Just as Hume is seen as the historic founder of evidentialism, those who
argue that there can be good practical reasons to believe independent of
one’s evidence turn for inspiration to Pascal’s claim that the best reason to
form a belief in God was a practical one, namely the possibility of avoiding
eternal suffering.16 Similarly, part of James’s motivation was to defend
6 Introduction
a certain religious perspective. However, opposition to evidentialism need
not be motivated by the desire to defend religion: consider Hume, who,
I’ve argued, should be understood as a pragmatist. For while Hume says
a wise man ought to proportion his belief to the evidence, he also sees that
for some of our most central beliefs, for example our belief in external
objects, “experience is and must be entirely silent.” Yet he does not think
this belief should, or could, be abandoned: “Nature is always too strong for
principle.”17 If we think of belief as isolated, narrow, and purely intellectual
instead of as deeply entwined with our emotions, desires, and well-being,
then we ignore who we are. Recognizing and accepting this complexity will
allow us the proper kind of reflection and, when needed, correction.
I have been discussing evidentialism as if it applies uniquely to one view
although, as we shall see, many different views can be termed evidentialist.
One may think that being an evidentialist does not prohibit one from seeing
evidential norms as grounded in practical or moral ones, and that one may
only mean that we should follow evidence because things will go better
for us. Given that I agree that evidential norms are most often the ones to
follow, perhaps I could be classified as a moderate evidentialist. There are
different ways one could classify these positions. I have termed any posi-
tion “pragmatist” that allows that some non-evidentially based beliefs have
nothing wrong with them. It seems that, despite the differences among con-
temporary evidentialists, they would all reject that view. I will argue that
whereas having true beliefs is extremely important, the truth of a proposition
does not always count in favor of believing it; holding some non-evidentially
based beliefs is possible, permissible, and need not be irresponsible.
The challenge to the pragmatist view I defend is to allow us to distinguish
pernicious non-evidentially based beliefs from those that are permissible.
The challenge, in other words, is to show that abandoning strict evidential-
ism does not simply allow one to believe whatever will make one happy.
I argue that a number of constraints can be placed on when it is permissible
to violate evidentialist rules. Given my view that the norms of agency guide
both action and belief, these constraints will be the similar to those that
prohibit one from acting in any way that makes one happy.

3. A WORD ON METHODOLOGY: WHAT IS

A PHILOSOPHICAL INVESTIGATION?

One of my main criticisms of evidentialism is that the phenomenon evidentialists

call “belief” bears little resemblance to what we ordinarily think of as belief;
its complexity is diminished, its scope and purpose narrowed. But, what do
I mean by “belief?” I will say more about this at the end of Chapter 3, but
I want to make it clear from the outset that my lack of explicit definition is
deliberate. My claim is that to understand the nature of belief, we must care-
fully investigate our doxastic practices.
 Introduction 7
Some philosophers deny that their theories need to match up with our
ordinary practices. Ordinary usage is complex and messy, and one of the
aims of a theory, one may argue, is to offer clarity and precision about the
phenomenon under investigation. That a theory shows that our common
practice is fundamentally flawed is not always thought of as revealing
a problem with the theory.
I agree that if, in the course of our investigation, we discover confusion or
even inconsistency, then some revision makes sense. In a discussion of how
epistemology can be naturalized, Hilary Kornblith discusses the interplay
between pre-theoretical observations and scientific description. He imagines
a rock collector gathering samples of an interesting kind of stone for the
purposes of trying to figure out what they have in common.18 Early on
in this investigation, the collector may have some ideas of what kind of rock
this is but as his theoretical understanding increases, he may find that some
of the rocks he initially thought were examples of the kind of rock in ques-
tion turn out not to be so. Kornblith argues that our pre-reflective intuitions
about knowledge (or belief) are based on a certain amount of understand-
ing but that we can come to revise these views as our understanding of the
phenomena increases.
Although I try as much as possible to avoid entering the debate between
naturalists and their opponents, I do think a better, deeper understanding of
any subject will likely change one’s pre-reflective view.19 But when a philo-
sophical account is revisionist and asks us to restrict usage (as we shall see
is the case with many of the evidentialist views we will look at), we need
a good motivation to do so. If the account, for example, has great explana-
tory power, then the restriction may be worth it. But if the restriction’s
only purpose is to allow one’s theory to be consistent, then I question its
worth. If we end up with a consistent theory that describes a phenomenon
bearing little resemblance to our ordinary practice, what has been illumi-
nated? One of my guiding assumptions when evaluating theories of dox-
astic norms and agency is that they should help to illuminate our doxastic
practices. I take it as a strike against a view if it deviates too much from
our ordinary practices; I realize this is not an assumption everyone shares.

4. STRUCTURE AND CHAPTER SUMMARIES

This book is divided into two main parts, “Doxastic Norms” and “Doxastic
Responsibility.” In Part I, I review and critique a number of defenses of
evidentialism before turning to my argument that the norms for belief are ulti-
mately practical. In Chapter 1, I consider various accounts in which a proper
understanding of the concept of belief reveals the truth of evidentialism.
Despite the differences in detail among these accounts, they all agree that
a belief is correct if and only if it is true, and that it is impossible for us to form
beliefs without good reasons or evidence for these beliefs being true. I think
8 Introduction
both these claims can be questioned, and that none of these defenses succeed
in showing that practical reasons for belief are conceptually impossible.
In Chapter 2, I consider the view that although it is not impossible to
believe for practical reasons, it is always wrong to do so. Following evidential
norms, According to this view, the way to promote epistemic values such as
truth, knowledge, or rationality. Those who offer this kind of defense may
agree that there are times when holding a non-evidentially based belief is
not prudentially or morally wrong, but that evidentialism is concerned only
with what one ought to believe from an epistemic point of view. I argue that
this separation of evaluative domains is problematic, and that the only way to
make sense of epistemic value is to link it or ground it in the practical.
In Chapter 3, I defend my pragmatist view of doxastic norms. The value of
truth and knowledge is instrumental; having true beliefs helps us achieve our
goals, flourish, and be excellent human beings. It is thus possible that some
beliefs can help us achieve these goals independently of their truth-value, or
of their being evidentially based. But truth and knowledge are so highly valu-
able that engaging in practices that lead away from truth and knowledge is
problematic in every sense—prudentially, morally, and epistemically. It will,
thus, only be permissible to hold non-evidentially based beliefs if doing so
does not allow for practices that undermine truth. This chapter also consid-
ers a number of objections and implications of my view, including a discus-
sion of what this view reveals about the nature of belief.
Those who oppose this pragmatist conception of doxastic norms will
point out that, given the involuntary nature of belief, we cannot believe for
practical reasons. This is why a discussion of doxastic norms is intertwined
with the issue of doxastic control and responsibility. Part II focuses on these
issues; I argue that beliefs are products of our agency, something we have an
active role in shaping and maintaining. In Chapter 4, I introduce a tension
in ordinary thinking about belief and consider two responses to what I call
“the puzzle of doxastic responsibility” that I reject. Briefly, the puzzle is
as follows: while much of what we believe is beyond our control, belief
is also open to normative assessment; we hold each other responsible for
our beliefs. But it seems that such lack of control should exempt us from
responsibility and judgment.
One can respond, on one hand, by arguing that we can effectively decide
to believe or, on the other, by arguing that we are, in fact, not responsible for
beliefs and that our common practices of attributions of responsibility are
misguided. I reveal deep problems with both these approaches. Chapter 5
engages with the third, and currently most common, response to the puzzle,
which argues that although we lack voluntary control over our beliefs, we
can nonetheless be held responsible for them. In Chapter 6, I turn to my
own response to the puzzle. I argue that if we want to hold people respon-
sible for their beliefs, then there must be a sense in which we have control
over them. Although we cannot believe at will, neither are we passive in the
 Introduction 9
beliefs we form and maintain. We take responsibility for our beliefs, and
taking responsibility includes taking control of them.
The two parts of the book are two sides of the same coin. That the norms
of agency apply to both belief and action demands that we can make sense
of doxastic agency. And that we can exercise control in the doxastic realm
naturally leads to the view that the same norms guide both action and belief.

NOTES

1. One could be an evidentialist and think there are no positive duties to believe
but instead only norms of permissibility. If this is so, it may seem that one has
no answer to the question “what ought I to believe?” But even if the norms
only dictate how not to believe, this answers the positive question to some
extent. I ought to believe only those propositions that are not ruled out.
2. Jonathan Adler’s book on this topic is titled Belief’s Own Ethics. One of his
central contentions is that it is a mistake to appeal to “normative notions” in
assessing what to believe. He refers to such approaches as “extrinsic,” and he
argues that this notion is based on a faulty assumption, namely that the con-
cept of belief alone does not fix the ethics of belief. Beliefs, thus, have their
own “ethics,” discovered by a clear analysis of the concept of belief. Many
defenses of evidentialism in the last decade have followed Adler in adopting
what he calls the “intrinsic” approach, namely focusing on how we must
believe given what “belief” means. These defenses are the topic of Chapter 1.
3. Aristotle, Nicomachean Ethics, 8.
4. The first extensive discussion of the view that beliefs aim at truth is found in
Bernard Williams’s “Deciding to Believe” in Problems of Self. His discussion
and some more recent accounts will be discussed in Chapter 1.
5. That we cannot choose what is true may be overstating the point. In certain
matters, matters that are up to us, there is a sense in which I can choose what
is true. Berislav Marusic argues for this view in “Belief and Difficult Action.” I
will return to the question of what kind of control one has over belief in Part II.
I will discus, in more detail, how the truth aim is supposed to explain why we
cannot believe at will in Chapter 1.
6. Virtue of character, Aristotle says, is about feelings and actions. A virtuous
person will have “feelings at the right times, about the right things, toward
the right people, for the right end, and in the right way.” Aristotle, Nicoma-
chean Ethics, 24.
7. For example, in the first chapter of Belief’s Own Ethics where he defends his
strong version of evidentialism, Adler begins by saying, “Evidentialism, an
ethics of belief advocated by David Hume, John Locke, W.K. Clifford, and
many others” (5).
8. Hume, Concerning Human Understanding, 170.
9. David Owen argues that Hume’s preference for reason has a moral ground.
He says, “the moral approval we feel towards the wise and reasonable per-
son, on the grounds that the characteristics of that sort are pleasing and
useful to their possessors and others, is the ultimate ground of Hume’s pref-
erence for reason” (Owen, Hume’s Reason, 220). He is one of the few to
engage with the question as to why Hume prefers the ways of the wise. This
is the central question that I engage with in “Why Should We Be Wise?”
10. Hume, Concerning the Principles of Morals, 268.
10 Introduction
11. Stephen Stich notes a very similar tendency when discussing the evaluation of
cognitive processes and the possibility that there might be more than one way
of reasoning that can count as a good one. He started seeing evidence for this
view, the view he calls “cognitive pluralism,” but found a prevailing concep-
tual argument (attributed to Davidson and Dennett) that “serious departures
from good reasoning are conceptually impossible” (Stich, The Fragmentation
of Reason, 15). One of his main worries about this conceptual argument
is that it undermines empirical exploration and data. The conceptual argu-
ments I discuss in Chapter 1 are very similar. If they are correct, beliefs that
deviate too much from the evidential norm are not really beliefs. So actual
cases that seem to undermine their view are not possible.
12. These views are discussed in Plantinga and Wolterstorff, eds., Faith and
Rationality. I will discuss these arguments in Chapter 3.
13. Adler, Belief’s Own Ethics, 3.
14. I will elaborate on the nature of this error in Chapters 1 and 4.
15. This is the main task of Chapter 3. While not all beliefs that can be termed
“religious” will end up on the nonpermissible side according to my view, I do
show that some of the specific religious beliefs discussed by the reformed
epistemologists remain problematic.
16. I do not claim that Pascal should ultimately be understood as rejecting
evidentialism. He seems to think a certain amount of self-manipulation is
needed to get yourself into a position that results in belief. And it is not clear
that once you have the belief you will view yourself as having it for practical
reason. But he, and his wager, are often taken as representing the view that it
can be a good thing to believe for practical reasons. Evidentialists often think
they need to respond to Pascale’s argument.
17. Hume, Concerning Human Understanding, 207.
18. Kornblith, Knowledge and Its Place in Nature, 11–14.
19. It is hard for me to clearly identify with either side in this debate. Many of
my criticisms of the conceptual defenses of evidentialists echo Kornblith’s
complaints against philosophy as a kind of a priori conceptual analysis where
the aim seems to be to come up with a consistent theory that can accommo-
date intuitions. He says that he is interested in knowledge, not our concept
of knowledge, and that philosophical investigations should not be different in
kind than that of the rock collector. Throughout my discussion, I do make use
of empirical studies in cognitive science and social psychology and so it seems
that my sympathies lie with the naturalist. But where I become uncomfortable
with this identification is that I do not want to be committed to the idea that
knowledge, or belief, constitutes a “natural kind” or that in trying to under-
stand belief we should look “outward” at “external” phenomenon rather than
“inward” to our intuitions (Kornblith, 16). The difference between rocks and
beliefs is that beliefs are part of us (the investigators) in a way that rocks are
not. So much of the data collected will include introspection and appeal to intu-
ition. My resistance to identifying as a naturalist may also be because I resist
the sharp division between mind and world that this approach seems to favor.

REFERENCES

Adler, Jonathan. Belief’s Own Ethics. Cambridge: MIT Press, 2002.

Aristotle. Nicomachean Ethics. Translated by Terence Irwin. Indianapolis: Hackett,
1999.
Clifford, William K. “The Ethics of Belief.” In The Ethics of Belief Debate, edited by
Gerald D. McCarthy, 19–36. Atlanta: Scholars Press, 1987.
 Introduction 11
Hume, David. An Enquiry Concerning Human Understanding. Edited by Tom
L. Beauchamp. Oxford: Oxford University Press, 1999.
James, William. “The Will to Believe.” In The Ethics of Belief Debate, edited by
Gerald D. McCarthy, 55–71. Atlanta: Scholars Press, 1987.
Kornblith, Hilary. Knowledge and Its Place in Nature. Oxford: Oxford University
Press, 2002.
Marusic, Berislav. “Belief and Difficult Action.” Philosopher’s Imprint 12, no. 18
(2012): 1–30.
McCormick, Miriam. “Why Should I Be Wise?” Hume Studies 31, no. 1 (2005):
3–19.
Owen, David. Hume’s Reason. Oxford: Oxford University Press, 2002.
Plantinga, Alvin, and Nicholas Wolterstorff. Faith and Rationality: Reason and
Belief in God. Notre Dame: University of Notre Dame Press, 1983.
Stich, Stephen. The Fragmentation of Reason. Cambridge: MIT Press, 1990.
Williams, Bernard. “Deciding to Believe.” In Problems of Self, 136–151. Cambridge:
Cambridge University Press, 1973.
This page intentionally left blank
Part I

Doxastic Norms
This page intentionally left blank
1 Conceptual Defenses of
Evidentialism

When Bernard Williams first introduced the idea of the truth-aim, it was the
first of five features discussed to illuminate the nature of belief, and to show
“how far, if at all, believing something can be related to decision and will.”
Williams actually says of this first feature that it can be “roughly summarized”
or “vaguely summed up” as “beliefs aim at truth.”1 His employment of the
truth-aim in an argument for why we cannot believe at will is as follows:
If truth is the aim of belief, any states that I can achieve at will would not rec-
ognizably be beliefs. For if what I believe were up to me, seemingly, I could
form a belief regardless of whether I thought it true—but if I knew this, then
I would know that there is no reason to think that this “belief” accurately
represents reality. But if to have a belief is to be committed to its truth, then
believing entails viewing the belief as representing reality. Believing at will
is, thus, incoherent, because it entails one viewing a belief as 1) necessarily
representing reality and 2) not necessarily representing reality. Thus, believ-
ing at will is not only a psychological impossibility, but the very concept of
belief also renders the idea incoherent.
It does not seem that my involvement in bringing about a mental state
means that I could not view such a state as representing reality. An obvious
counter-example, and one of those used to show that there is a trivial sense
in which we can control what we believe, is that I can make it be true that
the lights are on by flipping the light switch to turn on the lights. I could,
conceivably, have a belief that the lights are off and then decide to believe
the lights are on, and make this decision effective by flipping the on switch.
Williams does consider the possibility of manipulating oneself in various
ways to get oneself to believe something that one wants to believe through
processes such as taking drugs or hypnosis, but the beliefs he considers
are momentous—such as whether one’s child is alive or dead. Even if such
projects of self-manipulation are conceptually possible, Williams argues
that they must be irrational and, in any case, none of them are a direct result
of a decision to believe.
I will return to the question of whether certain projects of self-
manipulation that result in belief can be counted as decisions to believe.2
What is most important about Williams’s discussion in the context of defenses
16 Doxastic Norms
of evidentialism is that he was the first to clearly argue that one’s inability to
believe at will is not a contingent fact but is rather inconceivable. Thus, he
is the first to argue that the idea that beliefs aim at truth expresses a concep-
tual truth about belief that can help explain other belief-related phenomena.
While Williams was not centrally concerned with defending evidentialism,
a number of theorists in recent years have turned to Williams as inspiration
in thinking that a proper understanding of the concept of belief can reveal
why evidentialism is necessarily true.
One may wonder if it makes sense to talk about belief having an aim at all.
As Ralph Wedgwood has pointed out, beliefs do not literally aim at anything,
as does an archer or a hitman, so this talk is, in a sense, metaphorical.3 There
are, however, many activities that obviously have aims where there is no
literal “taking aim” at anything. It is common, for example, to ask what the
aim of a board game is.4 We shall see in the discussion that follows that when
many theorists posit the truth-aim for belief, they do not think it is necessary
to tie this aim in any way to desires, or conscious goals of agents.5 The aim
of chess is to checkmate one’s opponent, and this is the case independent of
any chess player’s desire to do so. It is even the case if one is aiming to let
one’s child win so as to encourage her. Understanding the meaning of the
game of chess tells us what its aim is. The conceptual defenses of evidential-
ism with which we are concerned in this chapter make an analogous claim.
Once we understand what it means to have a belief, we can know that its aim
is truth and, in turn, appealing to this aim reveals that the norms of beliefs
are strictly evidential.
To say that it is a conceptual truth that beliefs aim at truth or that belief
is subject to a norm of truth can be understood in a many different ways,
and the implications of such a view varies depending on its characterization.
In what follows, I will consider a number of different ways that the truth-aim
is thought to express a conceptual truth about belief. I begin with what
I take to be the most extreme (and ultimately the most implausible) position.
I will then consider progressively less extreme versions of the view and will
argue that the only plausible characterization of the truth-aim cannot show
that evidentialism is true.

1. THE EXTREME VIEW: A BELIEF NOT GROUNDED

IN EVIDENCE IS NOT A BELIEF AT ALL

According to one way of understanding the relationship between belief, evi-

dence, and truth, the entire notion of an ethics of belief is problematic. For,
normally “ought” statements only apply when it is possible to deviate from
what such a statement dictates.6 If believing something for which I lack
evidence is incoherent, then there is no question about what I ought to
believe beyond what I must believe. If it can be shown that it is conceptually
impossible “to maintain a belief in open defiance of the evidence,” then
 Conceptual Defenses of Evidentialism 17
evidentialism quickly follows. On such a view, evidential “norms” are not
really normative; they do not offer direction on how to be a good believer.
Rather, they are descriptive; they merely help elucidate what it means to
believe something. According to this view, most extensively argued for by
Jonathan Adler, for me to really believe p, I must take myself to have adequate
evidence (or epistemic reason) for p. Adler argues for the conceptual neces-
sity of evidentialism mainly by showing that the incoherence of a number
of assertions (inspired by Moore’s paradox) can only be understood if evi-
dentialism, as a conceptual truth, is correct. Adler’s argument is as follows:
If I believe something, then I must be willing to assert what I believe, but to
assert something is to claim that it is true, so to say I believe p is just another
way of stating p. This can be illustrated by the incoherence of the following
assertion: “It is raining, but I do not believe it is raining.” Adler would add
that the following is equally incoherent: “It is raining but I lack sufficient
evidence that is raining,” which (if assertion expresses belief) would also
show that the following assertion is incoherent: “I believe it is raining, but
I lack sufficient evidence that it is raining.” So Adler concludes, “We can-
not recognize ourselves as believing p while believing that our reasons or
evidence are not adequate to its truth and conversely.”7 And the “cannot” is
conceptual, not psychological.
Adler’s “incoherence tests” do not settle the matter of what is conceptually
possible. All they do is tell us that certain assertions sound very strange. Adler
admits that when one leaves the realm of relatively simple beliefs, assertions
of a similar kind may not seem so obviously incoherent. He considers, for
example, the self-acknowledged anorexic who may believe she is overweight,
despite recognizing evidence to the contrary. Adler maintains that any seem-
ing conjunction of one believing p while maintaining that one lacks sufficient
evidence for the belief will either be a case where one is not “fully aware”
of having both beliefs or there will be a temporal distinction between the
two conjuncts; they are not held in “a single consciousness.” He says if this
were not the case, then one would not have to seek “esoteric cases, like
those afforded by thoughts of the mentally disturbed” to refute evidential-
ism. But it seems that if one can generate counter examples (no matter how
esoteric) then what he is deeming “full belief” is only one species of the
genus “belief.” Adler says, “If there is no compelling connection between
the concepts of belief, truth and evidence, then counter examples . . . should
be plentiful. The need to search beyond the simple, blunt cases concedes the
connection even as it tries to refute it.”8 I am not denying that there are such
connections; the question at issue is the nature of the connection specifically,
is it one that is unbreakable?
What then accounts for the incoherence of these Moorean assertions if evi-
dentialism is not true? I think Adler’s diagnosis is largely correct, because beliefs
often do behave the way Adler says they do. If I sincerely assert that I believe
p, then it seems I am committed to acknowledging that I have some evidence
for this belief. To say, “I believe Tom is in the bar but I have no evidence that
18 Doxastic Norms
he is,” seems incoherent because my consciously believing it would normally
entail that I have evidence for it. But it does not seem that beliefs must behave
this way. What if a vicious blow to the head caused me to have the belief that
Tom was in the bar and I was not aware that this was the cause?
It would seem, in such a case, I can recognize myself (and even assert)
that I have a belief while also seeing I have no reason for it. It is possi-
ble that I may take the fact that I believe it to itself indicate that I have
some reason for it, even if I cannot access or remember such reasons. I may
have forgotten the reasons I had for initially forming many beliefs, but I take
the fact that I now have them as a clue that I once had a good reason for
believing them. Maybe my belief about Tom would seem just like my belief
that gold doesn’t decay the way other metals do. The difference with beliefs
for which one is unable to access the reasons and the belief about Tom is
that, in the former case, I can remember having reasons; I just cannot access
their content. I may even be able to remember roughly the context in which
I first formed these beliefs and I can imagine ways to recover and evaluate
the reasons for it.
Even if the simple holding of a belief is some evidence in its favor, it
certainly seems that Adler’s stronger claim doesn’t hold, namely that I regard
my evidence as adequate for the truth of p. This belief may be irrational and
false, but Adler’s theory is about belief, not only about rational belief. We
are charitable to one another, and assume rationality, which is why these
Moorean assertions sound so strange; one cannot be fully rational and make
such assertions. Adler is emphatic that he wants to preclude discussions of
rationality in the ethics of belief. He thinks that once we allow that the ques-
tion of what we ought to believe be “determined by criteria external to belief,
most prominently rationality,” we have gone “onto the wrong track.”9 His
intrinsic approach, which asks what the concept of belief demands, is sup-
posed to reveal why these more traditional, normative discussions of the
ethics of beliefs are misguided.10
To show that something is not a conceptual impossibility, all one needs
is a single counter-example. Is it possible for one to have a belief, think that
the evidence dictates one ought not to have it, and still retain it? In develop-
ing the idea of the truth-aim, Williams says, “If a man recognises that what
he has been believing is false, he thereby abandons the belief he has.”11
As we see, Adler agrees that this is so.12 I think it is possible for a gap to
exist between this recognition and abandonment. In some situations, one
may find oneself compelled to belief against one’s better judgment in a sense
analogous to compelled action when one acts against one’s best judgment.
An action is compelled only if the agent could not avoid performing it even
if he were to believe that, all things considered, he ought to refrain from
performing it. It seems, analogously, a belief is compelled only if the believer
could not abandon it even if he were to believe that, all things considered,
he ought not to believe it. We can thus define a compelled belief (or CB) as
follows:
 Conceptual Defenses of Evidentialism 19
CB: A belief b is compelled for Subject S during time t if and only if:

(1) S’s best judgment during t, all things considered, dictates that he
abandon b and
(2) S cannot abandon b during t.13

What must be better understood in order to assess the possibility of com-

pelled belief is the relationship between the recognition that a belief is not
evidentially supported and its abandonment. What happens after such rec-
ognition takes place? Is it plausible that at times such recognitions will be
entirely ineffectual, where one cannot give up a belief even if one thinks that
one ought to? What could cause a gap between the appreciation of a belief’s
illegitimacy and its abandonment? In John Heil’s discussion of these issues,
he is quite vague on the causes, saying only that there is some deficiency in
the believer’s total psychological state that allows him 1) to fail properly to
integrate his appreciation of certain facts and so 2) to continue to harbor
beliefs that are at odds with his better epistemic judgment.14
Alfred Mele makes use of his discussion of akratic action to say more
about the nature of this breakdown. Mele has pointed out that akratic action
occurs when evaluation and motivation diverge. One’s evaluations of which
desires should be acted on do not always correspond to their motivational
strength. One of Mele’s examples is the following: Someone with a severe
fear of flying may judge that his flying would be better than his not flying on
a particular occasion, and yet be so anxious that he would rather not board
the plane. Similarly, “The assessments of evaluations that ground decisive
better judgments about matters of belief need neither fully determine nor
exactly gauge the causal power of belief-influencing items. This opens the
door to the possibility of a mismatch between determinants of belief (e.g.,
salience) and one’s better judgment.”15 Heil makes a similar point when he
says, “It is one thing to appreciate the evidence, another thing to be moved
by one’s appreciation of it.”16
The cause of the mismatch between appreciation and motivation can be
further diagnosed if one recognizes the diverse determinants of belief, as
well as the multiple reasons why one might judge that one ought not to hold
a belief. One can think that a certain kind of belief-influencing item ought to
be causally efficacious, but this evaluation need not correspond to the actual
strength of this influence. So, perhaps I think that I ought to pay attention
to the mounting evidence that my boyfriend wants to end our relationship
(including his telling me repeatedly that he seriously doubts we have a future
together), but I find myself, against my better judgment, attending to the
way he looks into my eyes, and the tender way he holds my hand and so
continue to believe that we will soon be married and have a family.
It is possible that such a belief is merely akratic and not compelled. If one
lacks the capacity to make the belief-influencing item that is in line with one’s
better judgment causally efficacious—that is, if there is no effective strategy
20 Doxastic Norms
open to an agent so that she can successfully resist believing what she judges
she ought not to believe—then the belief is compelled.17 So consider again
my belief about my boyfriend. There seem to be strategies I could employ to
make my evaluations efficacious. For example, I could make myself repeat
the advice I gave friends who were unwilling to face painful truths, and viv-
idly recall their histories each time I saw that tender look. But if there is no
strategy open to me by means of which I could successfully resist believing
that we have a rosy future ahead, then that belief is compelled.
Is it possible, then, that one’s reflective evaluations about what to believe
are incapable of being efficacious, with the result that one has a compelled
belief? There seems to be nothing that rules out this possibility—why there
cannot be times when the gap between appreciation and abandonment cannot
be closed. Heil says that the incontinent believer is “typified by the psy-
choanalytic patient who has acquired what might be termed an intellectual
grasp of his plight, but whose outlook evidently remains unaffected . . .
he continues to harbour beliefs, desires and fears that he recognizes to be
at odds with his better epistemic judgment.”18 Heil is confident that, after
repeated reminders and the passage of time, the import of what the patient
has all along recognized will begin to take hold so that he “can be restored
to wholeness.” There are strategies open to this patient, with his therapist’s
help, that can be employed so that the “apprehension can be made to sink
in.” But if we remember that a belief is compelled only during a period of
time, then it seems quite plausible that there is a period of time during which
this patient’s therapy is incapable of having the apprehension sink in, or that
even once it has sunk in, certain events may trigger lapses.
However, it is not only severe mental illness that can lead to this incapac-
ity. A belief’s centrality or the painfulness of its loss may make it impossible,
even for perfectly healthy people, to abandon it—even while appreciating
its illegitimacy. Consider someone who has grown up in a fundamentalist
tradition and believes that the Bible is literally true. This person may come
to conclude that this belief is suspect, and that he should give it up. Now
we can imagine that it would be very difficult to break the habit of believing
in something inculcated at a very early age and reinforced throughout one’s
life—just as it is very difficult to give up smoking after twenty years even
when one decides one should.19
Psychological research on a phenomenon termed “belief perseverance”
shows that compelled belief may be even more widespread than has been
here suggested. This research shows that people’s beliefs tend to persist even
after all the evidence from which they were derived has been discredited—
particularly if these beliefs are central to an explanatory structure that has
been adopted. For example, in one study, subjects read case studies suggest-
ing that firefighters were either cautious types or risk-takers, and then were
asked to generate an explanation for this correlation. They were later told
that the information they were given was totally false but the subjects clung
to their newly created beliefs even though the evidentiary basis for them was
 Conceptual Defenses of Evidentialism 21
invalidated. These studies seem to indicate that it can be difficult to give up
beliefs even when one sees that one should.20
Now, if one is convinced that there are actual cases of what I have termed
“compelled belief,” then we know Adler is wrong, given that actuality
implies possibility.21 But the contingent fact of nonexistence is clearly not
sufficient to establish conceptual impossibility. Adler tries to bolster his case
by illustrating how completely bizarre it would be to, in “full awareness,”
acknowledge one believes without evidence. But these illustrations, if any-
thing, tell us about how attributions of rationality necessarily behave, not
that beliefs must necessarily conform to the dictates of such rationality.

2. A LESS EXTREME VIEW: IT IS A CONCEPTUAL TRUTH

THAT A BELIEF NOT GROUNDED IN EVIDENCE
CANNOT BE A CORRECT BELIEF

(i) Appeals to Design and Proper Functioning

If “beliefs aim at truth” is understood as a claim about the actual desires
and goals of believers, then evidential (truth-directed) norms only apply
contingently; the imperative to seek truth would be only hypothetical
and, thus, escapable. While many who endorse evidentialism would not
endorse the claim that it is conceptually impossible to hold beliefs without
sufficient evidence, they still seek to show, following Williams, that a proper
understanding of belief shows that it is a conceptual, and not a contingent,
truth that only true beliefs are correct. If following evidential norms tends to
provide us with true beliefs, then, if it is conceptually true that belief aims at
truth, it is also conceptually true that they are the norms governing beliefs,
thus revealing the truth of evidentialism. I will now consider a number of
ways of thinking about the truth-aim as revealing a conceptual truth about
the norms guiding belief.
David Velleman offers a way of understanding the truth-aim that he
thinks can show why only true beliefs are correct.22 Beliefs are not the only
cognitive attitudes whose content we regard as true. When we make an
assumption or imagine something, we regard the proposition assumed or
imagined to be true. What is distinctive about belief, according to Velleman,
is that we regard a proposition as true with the aim of getting the truth-value
right. He thus characterizes belief as follows: “Believing involves regarding
a proposition as true with the aim of so regarding it only if it really is. Thus,
to believe a proposition is to accept it with the aim of thereby accepting
a truth.”23 In expanding on this definition, Velleman makes it clear that this
aim need not be consciously recognized or pursued. Instead, he is thinking
of these cognitions being regulated for truth independently of how the agent
within whom they reside thinks about them. He conceives of there being
“cognitive systems” that regulate an agent’s cognitions “in ways designed
to ensure that they are true, by forming, revising and extinguishing them in
22 Doxastic Norms
response to evidence and argument.” The aim is, thus, constitutive; it is part
of what makes a belief a belief, and so, he says, it is a “conceptual truth that
beliefs are correct when true and incorrect when false.”24
Like Adler, Velleman’s discussion contains some slippage between appeals
to conceptual truths on one hand and empirical claims about human
cognition on the other. When Velleman uses the designer metaphor, it is
clear from the context that he means evolutionary design. We can then ask
why are human cognitive systems so designed? Eyes that work well will
allow us to see, cognitive systems that work well will lead us to form true
beliefs. Both our eyes and our beliefs are regulated so that we can survive
and perhaps flourish. Imagine if there were some contexts in which blurry
vision would better serve our purposes; then, an ocular system that allowed
for clear sight when needed and blurred when that was called for would be
the optimal design. The same holds for beliefs. If it turns out that at times
the truth of beliefs is irrelevant to our survival and flourishing, then the best
system would be one that is, in general, regulated for truth in the manner
described by Velleman, but can be suspended at times.25 Both our respiratory
and circulatory systems operate in this way. Breath and blood flow regularly,
but at times their deviation is a good thing. So, when we are in dangerous
situations, our breath quickens as adrenaline surges. When we have a cut,
blood flows quickly to the area and swells so that the cut can heal.
Velleman admits that it is possible that some psychological mechanisms
were designed to cause beliefs that happen to diverge from the truth. One
of his examples is that we have been adapted to have dispositions to be
cautious in perceiving predators so we are apt to believe, for example, that
the coiled rope is a snake. His second example is that we have adapted to
overestimate our own popularity. But the less responsive a state is to “cor-
rective influences,” the less likely Velleman is to call it a belief; a biased
belief can be corrected but a fantasy cannot. So, for example, he says that
someone who is not Napoleon but who professes that he is does not actu-
ally have the belief that he is Napoleon. We normally describe such people
as believing they are Napoleon but, according to Velleman, we are mistaken
in this usage:

Aren’t there people who believe that they are Napoleon? (People other
than Napoleon, I mean.) Don’t such people have a belief that isn’t
regulated for truth? I think the answer is that it isn’t literally a belief.
I suspect that we tend to apply the term “belief” in a figurative sense to
phantasies for which the subject doesn’t or cannot have countervailing
beliefs . . . The phantasy of being Napoleon is thus what he has instead of
a belief about his identity; and in this sense it is his belief on the topic, just
as a cardboard box on the sidewalk may be his house by virtue of being
what he has instead of a house. If you ask me, however, a cardboard box
on the sidewalk isn’t really a house. And a phantasy of being Napoleon
isn’t really a belief.26
 Conceptual Defenses of Evidentialism 23
But what justifies this restriction of usage? Even if it were a fact that
human cognitive systems have evolved so that beliefs track truth, deviation
from the norm is not conceptually impossible.27 If someone is born with an
extra finger, we do not then say it is not really a finger. Velleman’s account
can admit that the system can break down and we can end up with false
or irrational beliefs. But if someone has a mental state that misrepresents
reality, and there is no way for this state to be corrected, then, he says, this
mental state should not be categorized as a belief. An uncorrectable belief,
on his view, is a conceptual impossibility.
It is unclear how Velleman’s argument is supposed to support a conceptual
truth. Despite his claims about the concept of belief, his view is much more
compelling if it is taken as a more traditional (normative) defense of eviden-
tialism. We can say we ought to believe in a way that tracks truth (namely by
following evidence) because that is the way we believe when our cognitive
systems function properly. This is not a conceptual claim but an empirical
one and one that I will question.

(ii) Appeals to the Nature of the Reasoning Practice

Ralph Wedgwood’s characterization of the truth-aim is similar to Velleman’s,
but his account does not rule out the possibility of incorrigible beliefs.28 He
argues that the claim “the aim of belief is truth” should be understood as a
normative claim which he characterizes roughly as “that a belief is correct if
and only if the proposition believed is true.” He argues that this constitutes
the fundamental epistemic norm because it can explain all other normative
concepts relating to belief, such as the norms of rationality and knowledge.
To say that a concept is normative means that it plays a “regulative role” in
a certain practice. He elaborates on what this means as follows:

Suppose that a certain concept “F” is normative for a certain practice.

Then it is a constitutive feature of the concept F that if one engages in
this practice and makes judgments about which moves within the prac-
tice are F and which are not, one is thereby committed to regulating
one’s moves within the practice by those judgments.29

For example, Wedgwood says that the concept of a “legal chess move” is
normative for the ordinary practice of playing chess. If one judges that a cer-
tain move is not legal, then one is committed to not making it. To judge it
to be an illegal move and to do it anyway is irrational; it involves having an
incoherent set of mental states. This judgment commits one to accepting that
one “ought not” to make that move. Two concepts that Wedgewood argues
are normative in this sense, for the practice of reasoning (which is made
up of the “moves” of forming, maintaining and abandoning beliefs) are
“correct” and “rational.” If one judges a belief to be incorrect or irrational,
one is committed to not believing in this way and one is committed to
24 Doxastic Norms
accepting the view that one “ought not” to hold incorrect or irrational
beliefs. Wedgwood says, “If you make judgments about what it is correct
for you to believe and what is not, you are thereby committed to accepting
that you (in some sense) ‘ought not’ to hold beliefs that are not correct.”
What exactly is the nature of commitment here? If these norms are constitu-
tive of belief in the way that legal chess moves are constitutive of the game
of chess, then this commitment is only escapable if one ceases to engage in
the practice of belief formation and maintenance, something that might well
be impossible to escape.
Returning now to the truth-aim, to say that it is only correct to believe
a true proposition entails that one ought to hold only true beliefs and no false
ones, further believing any true proposition is correct and any false proposition
is incorrect. Wedgwood is aware that one can judge it rational to hold a false
belief in certain circumstances and so needs to accommodate this in his expla-
nation of how this fundamental norm explains the norms of rationality. He
does so by conceding that the fundamental norm as he has described leaves it
indeterminate when belief counts as rational. It dictates that if a certain propo-
sition p is true, then it is always better to believe it than to suspend judgment,
but if p is false, then it is better to suspend judgment than to believe it. But,
“The fundamental epistemic norm of correct belief . . . does not determine any
unique way of balancing the value of having a correct belief about p against
the disvalue of having an incorrect belief about p.” There are a number of con-
textually determined precisifications of “rational belief” that are still regulated
by the fundamental truth norm.30
This norm directs one to believe p only if p is true, but this end can only
be achieved if one uses certain means to achieve it. The secondary norms
(norms of rational belief) will direct one to follow rules that are reliable
means to achieve this end. Even though a rational belief can be incorrect and
irrational one can be true, these norms are still explained by the fundamen-
tal norm, because the ultimate reason we deem them rational is because they
are guided by rules that tend to be the appropriate means to achieve truth:
“Even though irrational beliefs can be correct, the only way it makes sense
to aim at having a correct belief is by means of having a rational belief.”
As it turns out, the norms of rational belief are evidential norms because,
again, proportioning one’s belief to the evidence tends to be the best means
to achieving true beliefs and knowledge.31
A worry with this kind of account of the belief aim is that it is overly restric-
tive in its view of rationality or what can count as a reason for belief. Perhaps
a belief’s being true is one criterion of assessment among others.32 Ernest
Sosa, for example, contrasts the norms that guide epistemic performances
with ones that guide other performances. An epistemic performance is
deemed successful when one believes truly because one manifests epistemic
virtue or competence.33 Beliefs aimed at comfort or some other pragmatic
objective, Sosa says, cannot be guided by epistemic competence and so can-
not count as knowledge. But Sosa does allow that one can be motivated to
 Conceptual Defenses of Evidentialism 25
believe for practical reasons and that, all things considered, it can be rational
to do so. Still, he thinks it makes sense to say that “a belief can be epistemi-
cally irrational though rational all things considered, where this last is to be
understood as rational all practical considerations considered, including the
desire for truth on the question at hand.”34
For Wedgwood, what it means for a belief to be rational is that it is
epistemically rational; the norms of rational belief are determined by the
fundamental epistemic goal of achieving true beliefs and knowledge. We can
have practical reasons for choices or intentions but not for beliefs. If, for
example, one were to determine that having a certain belief would have
disastrous consequences, this can serve as a reason for intending to try to not
bring it about that one believe the proposition in question. But if one none-
theless ends up with the belief, then, according to Wedgwood, “there is noth-
ing irrational” about having this belief.
This view is more plausible than the alternatives I have considered
because it does not require us to say certain mental states that we are
pre-reflectively inclined to call beliefs are not really beliefs. It does, however,
preclude the possibility of a belief held for non-evidential reasons being
rational. And it does so not by arguing for the merits of believing according
to evidential norms, but by appeal to what is claimed to be a conceptual
truth about belief.
As was the case with Velleman’s view, only a weaker conclusion can
be supported by Wedgwood’s account. He has described a certain prac-
tice that is governed by a specific norm. And while he has argued that it is
somehow conceptually built-in to the reasoning practice that it is governed
by this norm, the only way he argues for this is by means of the explana-
tory power it has in explaining the other norms of that very practice. One
needs to have already accepted a view of rationality to be convinced by
his view that the truth-aim can explain why we ought to hold “rational”
beliefs. It seems then that we are only left with a hypothetical imperative;
one ought to follow evidential norms if one wants to engage in a certain
kind of reasoning practice.35 It may be that all believers have a goal that can
be summarized as aiming at truth but Wedgwood, like others, is trying to
show that the concept of belief itself imposes norms on believers regardless
of what they want. But for this to follow, one must accept a substantive
view of rationality and reasons, one that requires defense. I will return to
the question of whether an account of the truth-aim that appeals to actual
goals of believers can be used to show that evidentialism is true. But first,
I will consider one more attempt to show that a proper understanding of the
concept to belief shows that it must be true.

(iii) Appeals to the Nature of Deliberation and Reason

Nishi Shah’s defense of evidentialism builds on David Velleman’s discussion
of the truth-aim and tries to correct some of the more problematic elements
26 Doxastic Norms
of Velleman’s account. His view also has a lot in common with Wedgwood’s
account, but he is more directly concerned with defending evidentialism.
Shah argues that Velleman’s account fails to distinguish between beliefs
formed in a deliberative context and those formed in other contexts; he
labels this problem the “teleologist’s dilemma.” Shah emphasizes that when
we reason, or deliberate, about what to believe, then only truth-related
questions matter; we are concerned only with evidence. Shah calls this phe-
nomenon transparency; the question whether to believe that p collapses into
the question of whether p is true.36 We shall see that many discussions of
belief accept this as a kind of platitude and then the question is how to best
explain it. But in many contexts in which we form beliefs, or are caused to
have beliefs, non-evidential processes such as wishful thinking are respon-
sible. If the teleologist, such as Velleman, weakens the disposition to form
true beliefs to allow for cases of wishful thinking and other non-evidential
processes, then he cannot explain why evidence plays an exclusive role in
reasoning about what to believe. To account for this, the teleologist would
have to strengthen the aiming-at-truth disposition so as to exclude influence
of non-truth-regarding considerations.
While we have seen that this seems to be the route Velleman is tempted to
take, as he says a non-correctable belief is not really a belief, to say that no
states that result from wishful thinking, for example, are beliefs might be even
more revisionist than Velleman would like. Shah’s problem is summarized as
follows: “We need an account that explains why deliberative belief-formation
is regulated solely by a disposition to be moved by alethic considerations, but
doesn’t require non-deliberative instances of belief-formation be also solely
regulated by such a disposition.”37
Shah’s way out of the dilemma is to emphasize, as does Wedgwood, the
conceptual necessity of truth being the standard of correctness for belief;
he thinks the concept of belief is a normative one; built into the concept of
belief is the idea that a correct belief is a true one.38 Here is a clear state-
ment of what he takes that to imply: “To say that it is a conceptual rather
than merely metaphysical matter that truth is the standard of correctness for
belief is to say that a competent user of the concept of belief must accept the
prescription to believe p only if p is true for any activity that he conceives
of as belief-formation.”39 As I stated, I think this is what Wedgwood argued
for as well; the imperative to seek only truth is categorical and inescapable
for anyone who believes. This understanding of the connection between
belief and truth offers a way out of the teleologist’s dilemma. According
to Shah, when one applies the concept in one’s reasoning, truth-relevant
considerations must be applied; but in non-deliberative contexts where the
concept is not exercised, one’s cognitive activity need not be regulated by
truth-relevant dispositions.
Shah further argues that reflection on transparency can help to show the
truth of evidentialism.40 Evidentialism, he argues, is “built in” to the nature
of doxastic deliberation. Although transparency does not immediately imply
evidentialism (namely, the view that only evidence can serve as a reason for
 Conceptual Defenses of Evidentialism 27
belief), it is so implied when coupled with what Shah calls “the deliberative
constraint on reasons.” This constraint tells us that something can be a reason
to X only if it is possible for it to function as a premise in deliberation to X.
When this constraint is applied to belief, the following holds:

R is a reason for X to believe that p only if R is capable of disposing X

towards believing that p in a way characteristic of R’s functioning as a
premise in doxastic deliberation.

Given that transparency shows that questions related to the truth of p are
the sole focus of our attention in doxastic deliberation, when it is combined
with the deliberative constraint, pragmatic considerations for believing are
shown to be impossible. Pragmatic reasons focus on the attractiveness of
doing something but nothing about the attractiveness of believing (aside
from whether is true), if we accept transparency, can serve as reason for
believing from the perspective of the believer. But is transparency as obvious
as Shah thinks it is, and is it true that I must always see my own beliefs
as evidentially based? I will argue that it is possible to consciously form
beliefs even when one recognizes one’s evidential reasons do not support it.
Shah, and many others, take as an unquestionable fact that “the deliberative
question whether to believe that p inevitably gives way to the factual ques-
tion whether p.”41 And, it is further assumed that the only considerations that
will enter into my deliberation have to do with evidence for p’s truth. This
characterization does not exhaust the ways in which we consider the question
whether to believe p. It seems non-alethic considerations can be part of even
first-person doxastic deliberation. Further, even if there is a sense in which the
question whether to believe p ends up collapsing into the question whether p,
it is not clear that all the considerations opposed to or in favor of p are strictly
“evidential.”
The kinds of examples of believing that put this evidentialist defense in ques-
tion are those put forth in defense of doxastic voluntarism. The connection
between doxastic voluntarism—the view that deciding to believe is possible—
and pragmatism—the view that it can be permissible (and possible) to believe
for non-evidential reasons—is not accidental. One way to defend evidential-
ism, we have seen, is to point out that the evidentialist thesis explains why
we are unable to decide what to believe. If we cannot decide what is true,
and if beliefs in some sense aim at or are governed by or are conceptually
tied to truth, then this shows why we cannot decide to believe. One way to
counter evidentialism is to deny that the phenomenon it purports to explain
is genuine. Perhaps it is possible to decide to believe when the evidence is
inconclusive. Carl Ginet offers a number of examples of beliefs that he thinks
depict agents deciding to believe.42 Here is one:

Sam is on a jury deliberating whether to find the defendant guilty as

charged; if certain statements of a certain witness in the trial are true, then
the defendant cannot have done what he is charged with; Sam deliberates
28 Doxastic Norms
about whether to believe those statements, to believe the prosecutor’s
insinuations that the witness is lying, or to withhold belief on the matter
altogether. He decides to believe the witness and votes to acquit.

I am not here concerned with whether this and the other cases Ginet
considers are, indeed, cases of deciding to believe.43 What I want to stress is
that it does seem that in these cases the subjects would regard themselves as
believing but they may not regard themselves as believing for truth-conducive
reasons. If I asked Sam why he believes the witness, he may have nothing
to say and it could be that he came to his decision partly based on feelings
of sympathy for the witness or based on his own need to for peace of mind.
While Shah can allow for such factors to enter into the causes of Sam’s
belief, what is precluded by his account is that Sam could see such non-alethic
considerations as reasons to believe the witness. But why can’t such practical
considerations be reasons for belief? The deliberative constraint on reasons
says a consideration can only be a reason to x if it is capable of “function-
ing as a premise” in deliberation to x. Now, if deliberation is characterized
as a kind of deductive argument with premises and conclusions, it would
certainly be very odd for a practical consideration to function as a premise
in whether to believe something. To say I am hungry and tired and, there-
fore, the witness is innocent is very bad reasoning, but is it conceptually
impossible? Again, does the nature of reasoning necessarily preclude it or
is only precluded if one accepts certain standards of reasoning that can, in
principle, be rejected? Shah seems to say that practical reasons are excluded
because belief is (conceptually) truth-governed, but part of his argument for
why it is a conceptual matter that truth governs belief appeals to the nature
of doxastic deliberation. The fact that truths about belief that are discovered
based on truths about deliberation are needed to explain these truths about
deliberation seems somewhat problematic, but I will set that aside for now
and turn to another question: is deliberation really best understood as an
argument with the conclusion being an action or belief?
In thinking about practical deliberation, it seems we deliberate when it is
not immediately clear what to do; it is usually when there are reasons sup-
porting different, often conflicting, actions. I have to decide whether I should
stay home and grade, or go see my friend’s band play. What goes on when
I deliberate about this? It seems I make a kind of list of considerations in favor
and opposed to each course of action. Some people even transfer this mental
list on to actual paper to assist in their deliberation. If, in the end, I decide to
stay home and grade, it seems anything that came up in that list can be a rea-
son for my staying home and grading. But did it function as premise? Would
it make sense to think of my deliberative process along these lines:

If I don’t grade tonight it will just make things worse for me tomorrow.
Things being worse for me tomorrow is something I should avoid.
Therefore, I should grade.
 Conceptual Defenses of Evidentialism 29
I suppose one can reconstruct practical reasoning in such a way, though it
bears little resemblance to what I think actually goes on in such deliberation.
And its conclusion is not an actual action but a normative statement.
If weak-willed actions are possible, I may go through that process and still
go out to the show. This way of thinking about deliberation fails to capture,
for example, all the considerations that were rejected that supported
another course of action. Now Shah is not committed to saying something
can only be a reason if it actually functions as a premise in deliberation; it
must only be capable of doing so. But it seems all considerations that arise
during the course of deliberation, even if they are rejected or overshadowed,
should count as possible reasons, though it is hard to see how to reconstruct
such complex, and somewhat messy, thoughts into argument form.
So again, consider our juror. In his process of deliberation, he is thinking
about how good he would feel if he believes the witness and can be done
with this process, that he would be able to get something to eat and go
home to his family. These are practical considerations in favor of believing
the witness. What Shah, and many others, either argue or assume is that
such considerations cannot be reasons for believing the witness. But now
it seems like this is only be ruled out if we are committed to a particular
substantive view about the nature of reasons. If these considerations can be
causes of Sam’s belief and Sam can even recognize them as such, it does not
seem impossible (or even implausible) for Sam to say I believed the witness
because I was fed up and exhausted. This sounds a lot like saying it was at
least one of the reasons for his belief.
In addressing the question of what is going on when it seems as if practical
considerations function as reasons in deliberation about whether to believe
something, Shah considers a number of explanations. One is that an agent
can be mistaken about what counts as evidence. If someone takes it as a gen-
eral principle that if something is good for him, it is probably true, then it
is possible that when thinking about whether to believe something, such
a person would think that facts about his good provide him with reasons to
believe. Shah says such a person is not mistaking a practical consideration
for an evidential one, but is accepting an unwarranted evidential principle.
So, though a third-person perspective can indicate that his belief is not based
on evidential reasons, from a first-person perspective, the agent mistakenly
sees desirability as an indicator of truth.
What Shah thinks these examples usually reveal, however, is a conflation of
the question of whether to bring about the belief, and the question of whether
to believe. So we could say of Sam either that practical considerations (say the
need for peace of mind or to return home) led him to attempt to act in ways to
bring about a belief, or that he perceived some feeling he had as an indicator of
truth. These do not seem to be the only ways of describing of what is going on
in this kind of a case. We can imagine him forming this belief, largely based
on his need for peace of mind, even if he does not subscribe to any general
principle that whatever provides him with peace of mind is true.44
30 Doxastic Norms
Probably if we asked Sam if he thought it was true that the witness was
honest, he would say “yes.” Williams is right that if one admits to having
a belief, one would also admit that what he believes is true. There is a trivial
sense in which believing p entails believing p to be true in that if I take myself
to believe something and someone asks me if I think it is true, then I will
say “yes.” I could not consciously hold a belief that I would assert is false.
But this connection between belief and truth cannot defend evidentialism.
For the truth-aim or truth-governedness of belief to defend evidentialism,
when I hold my belief to be true I must also mean that my reasons for it
are evidentially based. But it is possible, as I have shown above, to believe
something and take it to be true without taking myself to have any evidence
or reason for it. I have also suggested it is possible to believe something on
the basis of a non-evidential consideration.45

3. THE PRAGMATIST’S DILEMMA

In Shah’s defense of evidentialism, he lays out a dilemma for the pragmatist.

Again, the pragmatist is defined as one who is “committed to the existence
of at least some non-evidential reasons for belief.”46 The pragmatist must
either deny transparency, or if she does not deny transparency must deny
the deliberative constraint. For as we have seen, the two together imply
evidentialism. Again, this is because if the question whether to believe p
always gives way to the question about the truth of p (transparency), and if
for something to function as a reason for belief, then it must be capable of
functioning as a premise in deliberation (deliberative constraint), then only
reasons taken to be related to the truth of p, (namely evidential reasons for
belief) are possible. Yet, Shah contends no explanation of transparency can
succeed once one eschews this constraint. What horn of the pragmatist’s
dilemma, then, am I falling on? I am denying transparency as a conceptual
truth and so it may seem as if I am falling on the first horn, namely denying
transparency. Yet, Shah is careful to begin his discussion by remaining open
as to whether the described phenomenon is something that holds necessarily
or merely contingently. I do not deny that deliberating about whether to
believe something is often the same as questioning whether it is true. This
is a familiar and prevalent phenomenon, and thinking about it can help
tell us something important about the nature of belief. Shah thinks that
the best explanation of this common phenomenon is to think of belief as
governed by a single norm, and so, deliberation, which focuses on this con-
cept, can only attend to reasons that derive from this single norm. I think
this is mistaken and the challenge is then to explain transparency in a way
that allows for the possibility of non-evidential reasons for belief.
My answer is a version of one Shah considers and rejects. A pragmatist
can explain transparency by focusing on the interest we all have in our beliefs
being true. In asking myself whether to believe p, I focus on the question of
whether p is true because I have an interest in having a true opinion about p.
 Conceptual Defenses of Evidentialism 31
Shah’s problem with this account is that while it can explain why truth is
relevant to doxastic deliberation, “it cannot account for the fact that truth is
hegemonic with respect to doxastic deliberation.”47 The only way this could be
explained is if everyone always had an overriding interest in having true beliefs.
Shah finds this implausible because we can certainly imagine cases where it
would be more in one’s interest to have comforting false beliefs than to have
discomforting true ones. In answering this objection, I would like to consider
a suggestion from Hilary Kornblith. Kornblith argues that for epistemic norms
to have the force they do, they must be grounded in some universal desire. He
recognizes, however, that no desire or goal one posits (e.g., truth, knowledge,
rationality) will be sufficiently universal, and we would, thus, have different
norms, depending on our desires. “Human beings are a very diverse lot; some
of us are quite strange. It is hard to imagine making a plausible case for any
particular goal or activity which is genuinely universally valued.”48
Kornblith’s view is that we all—that is, anyone who has any goals at all—
have a reason to favor a cognitive system that is effective is generating truths
and avoiding falsehoods. So, we should care about beliefs being justified or
reasonable because these are the norms that, in general, will help us achieve
our goals. So, anyone with any goals must care about their beliefs being true.
The reasons I care about truth, he says, are ultimately pragmatic reasons.

I have argued that epistemic evaluation finds its natural ground in our
desires in a way that makes truth something we should care about
whatever else we may value. This provides us with a pragmatic account
of the source of normativity, but an account that is universal and allows
truth to play a central role.49

This general interest is sufficient to explain transparency. If, as a general

rule, avoiding false beliefs and trying to have true ones is of paramount
importance to my well-being, it would make sense that when I am self-
consciously attending to a particular belief, the truth of it will absorb
me. All the cases of doxastic deliberation Shah considers are of this self-
conscious type. It is not even clear if on Shah’s account Sam (the juror in
Ginet’s case) is engaged in doxastic deliberation, because he is ultimately
concerned with what to do—acquit or not. Coupled with this decision to
acquit comes a belief in the witness’s testimony. It seems here, where the sole
focus is not simply a particular belief, practical considerations can enter in
the deliberation. Ultimately, then, I do not think the connection between
belief and truth excludes practical reasons for believing, because the reason
why truth is so central is due to its practical value. So, although the fact that
a proposition is true usually counts in favor of believing it, it is possible for
practical considerations to override the evidential in favor of believing it.
I will expand on this defense of pragmatism and respond to a number of
objections to the view in Chapter 3. But first, I will discuss what Adler has
termed the “traditional” versions of evidentialism, namely those that go
beyond reflection of what the concept of belief demands to defend the view
32 Doxastic Norms
that one ought only believe on evidential grounds. In these views, believing
on non-evidential grounds is doing something wrong or impermissible.

NOTES

1. Williams, “Deciding to Believe,” 136.

2. In Chapters 4 and 5.
3. Wedgwood, “The Aim of Belief,” 267. In the introductory paragraph, he
says: “Beliefs are not little archers armed with little bows and arrows; they
do not literally ‘aim’ at anything.”
4. Owens, “Does Belief Have an Aim?” 283–305. David Owens has argued that
beliefs do not have an aim in any meaningful sense. It is our supposed lack
of control we have over beliefs that leads him to this conclusion. Because he
thinks beliefs cannot be controlled by reflection, the way many actions can,
it is misleading and unhelpful to think of beliefs as having aims at all; he
ultimately concludes that we should not try to force beliefs “into the mould
of the teleological norms which govern action.” While I ultimately agree with
Owens that appeals to the truth-aim cannot do the work to which various
theorists try to put it, given that the aim is characterized in so many ways, one
cannot rule out, in advance, that none of these are meaningful.
5. Whiting, “Does Belief Aim (Only) at Truth,” 279–300. In a recent, and very
helpful, discussion about different ways to understand the truth-aim, Daniel
Whiting says he prefers to “view the aim as one which subjects possess at
a personal level.” I am not sure if the goal must be consciously recognized or
pursued. He also states that this is not a contingent goal but one that is neces-
sary for having beliefs at all. It seems then, as long as one believes anything,
one has this goal whether or not one recognizes it. If this is the case, then
I am not sure how viewing the aim at personal level differs substantially from
viewing it from a sub-personal level.
6. Adler, “Beliefs Own Ethics,” 3. As I discussed briefly in the introduction,
Jonathan Adler presents the most extreme version of such a defense. He uses
the language of extremism himself. In discussing the evolution of his view
he says: “My position started to shift to the extreme stance defended here
in reaction to discrepancies I observed between sophisticated, contemporary
anti-evidentialist pronouncements and their actual detailed claims and
arguments.” Again the anti-evidentialists of concern are those defending the-
ism and religious belief in general.
7. Ibid., 32.
8. Ibid., 35.
9. Ibid., 26.
10. These defenses are the subject of Chapter 2. While I ultimately think they fail
as well, I think the questions posed in these more traditional approaches are
more pertinent to what matters most when we are evaluating our own and
others’ beliefs.
11. Williams, “Deciding to Believe,” 137.
12. Ralph Wedgwood makes a similar claim: “The judgment that it is not
rational for you to believe p commits you to not believing p.” One of the
many questions about Wedgwood’s account is what he means by commit-
ment here. We will return to his discussion below (“The Aim of Belief,” 269).
13. There are beliefs that may seem compelled, in certain ways, that do not
conform to this definition. Am I not compelled to believe the conclusion of a
 Conceptual Defenses of Evidentialism 33
valid argument or to believe there is a tree in front of me when I see a tree?
There are many beliefs that are irresistible, beliefs that it seems I cannot give
up. An irresistible belief can be defined just as is CB above, absent condition
(1). So, compelled beliefs are a subclass of irresistible beliefs. They are the
type in which I am interested because if beliefs of this kind exist, they clearly
show that Adler’s conclusion is false. They also lend support to the view that
there is a place for control in the doxastic realm, a point I will return to in
Chapter 6. Cases of compelled belief seem to reveal a lack of control that we
normally do have. I provide a full account of the nature and possibility of
such beliefs in “Compelled Belief.”
14. Heil, “Doxastic Incontinence,” 56–70.
15. Mele, Irrationality, 119.
16. Heil, 70.
17. Mele, Springs of Action, 102. The account given here of compelled belief has
much in common with Mele’s account of irresistible desire provided in Chap-
ter 5 of his Springs of Action. Mele there defines a desire D as being literally
irresistible for S at t if there is no strategy for resisting D open to S. What it
means for a resistance strategy to be “open to S” is given careful elaboration.
For our purposes, we can summarize it thus: For a resistance strategy to be
open, it must be “both representationally and motivationally open” to the
agent and the agent must have the “physical and psychological skills and
capacities necessary to execute the strategy in such a way that he intention-
ally brings it about that he does not perform an intentional action on the
basis of D.” Consider a weak-willed pie-eater, Fred, who had successfully
resisted desserts for a month. He has strategies of self-control at his disposal;
he could have refused to focus his attention on the yumminess of the pie, he
could have promised himself a reward for resisting. And if he had employed
such a strategy, he could have augmented his motivation to the point where
he might have refrained from the akratic action. If the action resulted from
an irresistible desire, then this is not the case; he had no means at his disposal
that would allow him to resist the desire.
18. Heil, 69.
19. In Chapters 5 and 6, I discuss examples of this kind to help understand
doxastic responsibility and why it is sometimes mitigated.
20. Kassin and Studebaker, “Instructions to Disregard and the Jury: Curative and
Paradoxical Effects,” 413–434. These studies do not tell us that the agents in
question judge that they ought to give up these perseverant beliefs, but they
do seem to show that epistemic considerations, even when made apparent,
do not always lead to the abandonment of the suspect beliefs.
21. Adler, “Akratic Believing?” 18. All one really needs as a counter-example
to Adler’s strong evidentialism is a clear case of an akratic belief because,
according to Adler, no gap between appreciation of lacking evidence and
abandonment of belief is possible (even a gap that can be overcome.) Not
surprisingly, in “Akratic Believing?” Adler argues that akratic belief is not
possible. His central argument is that once one has achieved full belief, it
is not possible for there to be any belief-determinants left to motivate as
they have all been taken into account in the forming of the belief. While the
yumminess of the pie can still serve as motivation for Fred even after he has
determined it is best not to eat the pie, Adler says once one reaches full belief
“contrary or undermining evidence is nullified. So there is no evidence to play
the role of conflicting desires in drawing the agent away from his better judg-
ment.” We see again that if we are still being drawn by any of the evidence
we have deemed unworthy, according to Adler, we are not really believing.
34 Doxastic Norms
22. Velleman, Practical Reason, 243. He says, “If it can be a natural or scientific
fact that belief aims to be true, then it can also be a natural or scientific fact
that false beliefs are wrong or incorrect.”
23. Velleman, 251.
24. Ibid., 253.
25. In Chapter 3, I will discuss some examples of what are claimed to be
“adaptive misbeliefs.”
26. Ibid., 289.
27. It is not at all clear that human cognitive systems have evolved to maximize
achieving truth. I will be discussing some of the theories of cognitive science
on this matter in Chapters 2 and 3. We shall see that it is argued that we have
evolved to avoid gross errors rather than gain truths.
28. Hamid Vahid discusses and critiques both Velleman and Wedgwood in
“Aiming at Truth: Doxastic vs Epistemic Goals.” Some of his objections to
these are similar to mine.
29. Wedgwood, “The Aim of Belief,” 268.
30. Ibid., 275.
31. Wedgwood devotes a lot of time to explaining the nature of the rules guiding
rational belief formation. While his main purpose in this article is not to
defend evidentialism, the only way that the fundamental truth-norm as he
describes it, can explain the norms of rationality is if these norms are ones
which tend to be truth-directed, which in turn means they are evidentially
based. He summarizes what it means for a belief to be rational: “rational
beliefs are beliefs that either result from, or amount to, one’s following a rule
or set of rules that is rational for one to believe to be reliable.”
32. See, Jose Zalabardo “Why Believe the Truth? Shah and Vellman on the Aim
of Belief” on the view of whether beliefs can only be assessed according to
their truth value.
33. Sosa, Knowing Full Well.
34. Ibid., 34.
35. See footnote 11. Stich argues extensively for what he calls “cognitive
pluralism,” that there is more than one kind of reasoning practice. There
have recently been some defenders of what has been termed “permissivism,”
which accepts that two different doxatic attitudes can be rational given the
same body of evidence. But this requires making sense of there being multiple
permissible epistemic standards. For a defense of this view, see Miriam
Schoenfield, “Permission to Believe: Why Permissivism Is True and What It
Tells Us about Irrelevant Influences on Belief,” and her note 5 for others who
defend permissivism.
36. Moran, Authority and Estrangement, 60–61. Many theorists have endorsed
some thesis about belief for which they use the term “transparency,” and
they do not all mean the same thing. Richard Moran offers two formulations,
which both seem quite similar to Shah’s use. He says, “from within the first-
person perspective, I treat the question of my belief about P as equivalent to
the question of the truth of P,” and, “a first-person present-tense question
about one’s belief is answered by reference to (or consideration of) the same
reasons that would justify an answer to the corresponding question about
the world.”
37. Shah, “How Truth Governs Belief,” 467.
38. In a later article co-written by Shah and Velleman, Velleman agrees with
Shah on this point. See “Doxastic Deliberation,” 497–534.
39. Shah, “How Truth Governs Belief,” 470.
40. Shah, “A New Argument for Evidentialism,” 481–498.
 Conceptual Defenses of Evidentialism 35
41. Shah, “A New Argument,” 481. This is Shah’s most recent description of
transparency.
42. See Ginet, “Deciding to Believe.”
43. In Chapter 4, I will return to Ginet’s examples.
44. I am not even sure that the principle Shah considers (namely, whatever is
good for me is true) is coherent. Many things that are good for me are not the
type of things that admit of truth or falsehood. It is good for me to exercise;
is my exercising true? I do not understand this question.
45. Whiting, “Does Belief Aim (Only) at the Truth?” 279–300. In his discussion
of the truth-aim, Daniel Whiting says the proposition, “For any p, believing
p is correct if and only if it is true that p,” is the “merest platitude.” This is
only the case in light of certain assumptions. First, that one takes “correct,”
as does Whiting, to mean “acceptable,” not “required.” Second, that one
means acceptable from an epistemic perspective. It may be acceptable to hold
a false belief from a practical perspective. Even if one cannot consciously
hold a belief that one recognizes as false, nothing in the proposition as stated
limits believing to conscious believing. But even if one can construe the
belief-aim in such a way that it can account for this correctness-condition
that is a long way from showing that the aim reveals that the norms of belief
are evidential norms. Whiting’s discussion shows how this connection is not
obvious.
46. Shah, “New Argument,” 482.
47. Ibid., 490.
48. Kornblith, “Epistemic Normativity,” 367. See also Kornblith, Knowledge
and Its Place in Nature, 150.
49. Kornblith, “Epistemic Normativity,” 373; and Kornblith, Knowledge and
Its Place, 161.

REFERENCES

Adler, Jonathan. “Akratic Believing?” Philosophical Studies 110, no.1 (2002): 1–27.
———. Belief’s Own Ethics. Cambridge: MIT Press, 2002.
Ginet, Carl. “Deciding to Believe.” In Knowledge, Truth, and Duty: Essays on
Epistemic Justification, Responsibility, and Virtue, edited by Matthias Steup,
63–76. New York: Oxford University Press, 2001.
Heil, John. “Doxastic Incontinence.” Mind 93 (1984): 56–70.
Kassin, S. M., and Studebaker, C. A. “Instructions to Disregard and the Jury:
Curative and Paradoxical Effects.” In Intentional forgetting: Interdisciplinary
Approaches, edited by J. M. Golding and C. M. MacLeod, 413–434. Hillsdale,
NJ: Erlbaum, 1998.
Kornblith, Hilary. “Epistemic Normativity,” Synthese 94, no. 3 (1993): 367.
———. Knowledge and Its Place in Nature. Oxford: Oxford University Press, 2002.
McCormick, Miriam. “Compelled Belief.” American Philosophical Quarterly 42,
no. 3 (2005): 157–169.
Mele, Alfred. Irrationality: An Essay on Akrasia, Self-Deception, and Self-Control.
New York: Oxford University Press, 1987.
———. Springs of Action. Oxford: Oxford University Press, 1992.
Moran, Richard. Authority and Estrangement. Princeton: Princeton University
Press, 2001.
36 Doxastic Norms
Owens, David. “Does Belief Have an Aim?” Philosophical Studies 115 (2003):
283–305.
Schoenfield, Miriam. “Permission to Believe: Why Permissivism Is True and
What It Tells Us about Irrelevant Influences on Belief.” Nous 48, no. 2 (2014):
193–218.
Shah, Nishi. “A New Argument for Evidentialism.” The Philosophical Quarterly
56, no. 225 (2006): 481–498.
———. “How Truth Governs Belief.” The Philosophical Review 112, no. 4 (2003):
447–482.
———, and J. David Velleman. “Doxastic Deliberation.” The Philosophical Review
114, no. 4 (2005): 497–534.
Sosa, Ernest. Knowing Full Well. Princeton: Princeton University Press, 2010.
Vahid, Hamid. “Aiming at Truth: Doxastic vs Epistemic Goals.” Philosophical
Studies 131, no. 2 (2006): 303–335.
Velleman, David. The Possibility of Practical Reason. Oxford: Oxford University
Press, 2000.
Wedgwood, Ralph. “The Aim of Belief.” In Philosophical Perspectives 16 (2002):
267–296.
Whiting, Daniel. “Does Belief Aim (Only) at the Truth?” Pacific Philosophical
Quarterly 93, (2012): 279–300.
Williams, Bernard. “Deciding to Believe.” In Problems of Self, 136–151. Cambridge:
Cambridge University Press, 1976.
Zalabardo, Jose. “Why Believe the Truth? Shah and Velleman on the Aim of Belief.”
Philosophical Explorations 13, no. 1 (2010): 1–31.
2 Normative Defenses of
Evidentialism

We have seen that conceptual defenses of evidentialism are not concerned

with the question of why it is good to follow evidential norms. If these are
the norms that necessarily guide belief formation, then there is little to say
about why they should be followed. It would be like asking why it is good
to move a bishop diagonally rather than horizontally. It is constitutive of the
game of chess that bishops only move diagonally. If a child learning the game
were to ask why do bishops only move this way, no response could be given
beyond, “That’s just the way the game works . . . that’s the bishop’s job.”
Of course, if the child keeps learning and then asks, “Why did you move
your bishop one space instead of five spaces,” a lot can be said in response.
One can explain why it was a better move, and why it is likely to put one
in a better position to win the game. Historically, defenses of evidentialism
treated the question of “why believe according to evidential norms” as more
akin to the second kind of question. There was no assumption that this was
the way one must form beliefs, but it was argued that this was the better way
to believe. When Hume said, “A wise man . . . proportions his belief to the
evidence,” he was contrasting the way the wise form beliefs with the way
the vulgar (or most people) do most of the time. Ultimately, he argues that
the ways of the wise are the better ways, but, as discussed earlier, explaining
why this is so requires going beyond an explication of the concept of belief.1
Part of what I argued in Chapter 1 is that conceptual defenses collapse
into normative ones. Inevitably, some reference is made to the proper way
of believing, or they appeal to an implicit understanding of good reasoning
or the nature of rationality. I did not there, however, consider the merits of
explicitly normative defenses. Again, these defenses claim that although it is
not impossible to believe without evidence, or on the basis of practical con-
siderations, it is always wrong to do so. One common way of arguing for
this is to claim that following evidence leads to the promotion of something
of value, such as truth or knowledge. Promoting such a value is some-
times characterized as the aim or purpose of believing, though not all such
teleological accounts refer specifically to belief’s aim. Though they vary in
how they characterize the aim of belief, or of the value to be promoted by
believing virtuously, most of these defenses of evidentialism assume that the
38 Doxastic Norms
relevant criterion when assessing a belief is epistemic. Following eviden-
tial norms will provide one with beliefs that are epistemically valuable; the
norms governing proper belief-formation occupy a domain independent of
other normative domains.
I will argue that one cannot make sense of a point of view that is distinctly
and exclusively epistemic while at the same time retaining the normative
force that those endorsing evidentialism seem to take epistemic value to
have. I will begin by discussing an account that represents the widely held
contention that epistemic value is autonomous from other kinds of value.
I will then consider why one may think that epistemic goods—such as truth,
knowledge, and rationality—are goods to be promoted. I argue that the
only way to make sense of epistemic value as a good to be promoted is to
link it, or ground it, in the practical or moral. Thinking about why we value
truth and knowledge reveals that the norms guiding us in what is called
the epistemic realm are not isolated from other normative domains. If this
is true, I will argue, then to address the question of what is permissible to
believe from a purely epistemic perspective is overly narrow and ultimately
unhelpful.

1. PROMOTING EPISTEMIC VALUE

In Richard Feldman’s discussion of the ethics of belief and argument for

why one should believe in accordance with the evidentialist’s dicta, he uses
language of believing well and of doing so to promote a certain value. His
account thus seems to be illustrative of this more traditional, Cliffordian
defense of evidentialism. If one takes it as a starting point, as does Feldman,
that following one’s evidence is the “proper way to achieve something of
epistemic value,” then the question arises: which epistemic value does it
promote?2 From our discussion so far, it would seem an obvious candidate is
“truth.” But if one is in unfortunate circumstances in which evidence leads
to falsehoods, following the evidence as we usually think of it is not the best
way to the truth. Feldman still thinks that one ought to follow one’s evi-
dence in such circumstances. He says, “A person who irrationally believes a
lot of truths is not doing well epistemically. In contrast, a person who forms
a lot of rational but false beliefs is doing well epistemically.”3
If not truth, then what epistemic value is evidentialism in service of? What
we want, Feldman argues, is rationality; we want beliefs that would be true
and provide us with knowledge in ideal conditions. Feldman’s principle of
epistemic value is as follows: “When adopting (or maintaining) an attitude
toward a proposition, p, a person maximizes epistemic value by adopting
(or maintaining) a rational attitude toward p.” And it is rational to follow
the evidence. Following the evidence will give us knowledge when knowl-
edge is available; if you follow the evidence and other conditions hold, then
you will get truth and knowledge.
 Normative Defenses of Evidentialism 39
On this account, as with many others, it is important that the epistemic
occupies a distinct normative realm, that the category of “epistemic
value” is independent of other value. Feldman refers to people doing well
“epistemically” as “achieving epistemic excellence.” One of Feldman’s main
conclusions is that the epistemic “ought” is entirely distinct from the moral
or practical. He contends that for each “ought” there is an associated value
and we “ought, in the relevant sense, [to] do the thing that maximizes, or
perhaps something that does well enough in achieving that kind the value.”4
So we morally ought to do what produces enough moral value, prudentially
ought to do what produces enough prudential value, and epistemically ought
to do what produces enough epistemic value. But if these “oughts” conflict,
there is no way to adjudicate between them, no meaningful question about
what I ought to do or believe all things considered. He says: “We’ve disam-
biguated ‘ought’ and we can’t put the various senses back together again.”5
This separation of evaluative domains is problematic. When I wonder
whether I should believe something or when I criticize someone for holding a
belief, my question or criticism does not seem limited to one of these evalua-
tive perspectives. There may indeed be a source of normativity that provides
force to our practical, moral, and epistemic judgments. Feldman wonders
what value would be associated with this “just plain ought.” This is a good
and difficult question, but not one that is meaningless or unanswerable. It
seems that each of these “oughts” can be thought of as connected to a dimen-
sion of agency. Beyond adopting one of these evaluative perspectives when
asking what I should do or believe, I can ask what I should do or believe qua
agent. If thinking, feeling and acting are all dimensions of agency then the
values associated with each dimension can be combined and compared. On
such a view, epistemic value is not autonomous from other values.
Many epistemologists who think about what makes a belief permissible
agree with Feldman and assume this question can be answered by thinking
about what is valued epistemically, and they distinguish epistemic value
from other kinds of value. They may disagree about how to characterize this
value, but they share the assumption that it is a value that is entirely distinct
from other non-epistemic values. It is that assumption that I question. I will
now consider some possible ways of explicating epistemic value and argue
that the most plausible way of understanding why we value true belief and
knowledge shows that epistemic value is not autonomous in the way it is
often assumed to be.

2. WHY ARE TRUTH AND KNOWLEDGE VALUABLE?

It is generally assumed that knowledge is good and that it is something

we want and care about. A number of theorists have recently argued that
it is important for any theory of knowledge to explain why knowledge is
valuable.6 For example, in introducing the central thesis of his latest book,
40 Doxastic Norms
John Greco says that his account tells us what knowledge is but “more
specifically and more importantly, it is a thesis about the normative status
that knowledge requires.”7 It seems uncontroversial, he says, that when we
say someone knows something, we are making a value judgment and, if this
is so, that is, if epistemology is a “normative discipline,” then “a central
task of epistemology is to provide an account of the normativity involved.”
In Ernest Sosa’s most recent book, he claims that any successful account of
the nature of knowledge must also explain its value. Whatever condition
must be added to belief and truth to constitute knowledge “must add
normatively positive content, moreover, sufficient to explain how it is that
knowledge, which must satisfy this further condition, is as such always better
than would be the correspondingly merely true belief.”8 Linda Zagzebski
claims that “any acceptable answer to the question ‘What is knowledge?’
must be compatible with a reasonable answer to the question ‘What makes
knowledge good?’ ”9
One obvious answer to why true beliefs and knowledge are valuable is
that they are extremely useful. One’s life may depend on knowing which
drink is the poison. Less dramatically, our ability to meet friends for coffee,
drive to the grocery store, or cook dinner all require us to know a lot.
More lofty knowledge has led to eradication of disease, building of cars
and planes, computers and the Internet, all of which are extremely useful.
An immediate worry some may have about the pragmatic explanation of
epistemic value is that the value of knowledge would be relativized to what
people want and value. It seems no sense could be given to knowledge being
a universal value. I may value knowing something that you have no reason
to care about at all. And if the value of knowledge is grounded in the care
and desires of humans, then it seems that epistemic value will be as fleeting
and fickle as those desires. I have already mentioned (in Chapter 1) Korn-
blith’s answer to this worry. His view is that anyone who has any goals at
all has a reason to favor a cognitive system that is effective is generating
true beliefs. We should care about knowledge because it allows us achieve
our goals.
Kornblith expresses some dissatisfaction with his pragmatic characterization
of epistemic value. If someone could give substance to an account that made
the value of truth non-instrumental and the injunction to seek truth categori-
cal, then he “would not be hostile to such an account,” but he cannot see
how it can be done. Those who take up Kornblith’s challenge argue that the
value of knowledge can be completely divorced from its practical benefits.
On such a view, though obtaining knowledge may not be, all things consid-
ered, valuable, it always possesses some prima facie value. I will show that
such accounts fail, but I do not think anything is lost because of this failure.
How does one argue that the value of a particular good is independent
of the value of other goods? One way is to argue that its value can never be
lost, that at least part of what makes it a good is simply that it is the kind
of thing that it is. Something that, by its very nature, possesses value has
 Normative Defenses of Evidentialism 41
been called an intrinsic, inherent, basic or final good. While these terms
may be pulled apart, what they share is the idea that these goods are not
good because of another good they lead to or are grounded in.10 I will
use the term “non-instrumental” to refer to this kind of value. The most
uncontroversial good of this kind is happiness, particularly when happiness
is thought of as equivalent to the Greek eudaimonia. A happy life, on this
view, is a flourishing life, a life lived well. There is not one single kind of
flourishing life, no single way of exhibiting excellence. Without any specifi-
cation or qualification, one can say that a flourishing human life, whatever
its particularities, is good. It is not good because of any other good. One
prominent way of arguing for knowledge’s non-instrumental value is to con-
nect it to the value of flourishing and to then show that the nature of the
connection is such that the non-instrumental value of flourishing entails the
non-instrumental value of knowledge. These accounts all situate themselves
in a broadly Aristotelian framework.11 I will now discuss some ways in
which one might think about the role of truth and knowledge in a flourish-
ing life. I contend that no plausible view of this connection can support the
conclusion that epistemic value occupies an independent normative realm.
Thinking again about the huge practical benefit of knowledge, we can
see it would be very difficult for one to flourish without it. At least some
knowledge is required for flourishing. In this view, knowledge is valuable, in
a way that is similar to friendship or social interaction, in that they all have
an important role in flourishing human life.12 But if this is the connection
between knowledge and flourishing, it does not seem to confer normativity
on the epistemic that is isolated from the practical. It is important to have
some knowledge, just as it is important to have some social interaction, but
this does not then show that every kind of social interaction has some value,
or every kind of knowledge. Even if we do not view knowledge or social
interactions simply as a means to having a flourishing life, it seems the most
a view of this kind will do is to show that flourishing is not possible without
some knowledge. But this does not take us any further than Kornblith’s prag-
matic account. True beliefs and knowledge are needed to reach whatever our
goals may be and it would seem impossible to flourish if one were incapable
of reaching one’s goals. And, so, some knowledge is needed to flourish.
Again, taking inspiration from Aristotle, one can argue knowledge exhibits
virtue and given that virtue is constitutive of human flourishing so is knowl-
edge.13 For Aristotle, one performs one’s human function excellently when
one engages in virtuous activity. As discussed earlier, the best (most excellent,
virtuous) human will always act correctly, and such an ideal person will also
always believe correctly and feel correctly. Eudaimonia differs from other
goods because only it is complete and self sufficient; nothing can be added to
it to make it any better. All other goods are valuable because of their contri-
bution to eudaimonia. Even if knowledge has a central place in a flourishing
life, and one could not live an excellent life without some knowledge, this
does not show that knowing something, whatever that thing is, has some
42 Doxastic Norms
value simply in virtue of it being knowledge. When Aristotle catalogues the
various kinds of virtues of thought, we see that most kinds of knowledge
are clearly valuable because of their practical use. Having knowledge of
medicine, for example, helps to keep people healthy and cure disease, and
craft knowledge leads to the creation of all kinds of useful objects. Even
practical wisdom (or what is sometimes translated as prudence) that is
required for virtuous action is said to be good because of what it brings
about: “It seems proper to a prudent person to be able to deliberate finely
about things that are good and beneficial for himself . . . about what sorts of
things promote living well in general.”14 Aristotle does, however, distinguish
between those goods that are a means to happiness and those that are a part
of it. He says of philosophical wisdom that it produces happiness “not in
the way that medical science produces health, however, but in the way that
health produces health . . . For, since wisdom is a part of virtue as a whole, it
makes us happy because it is a state that we possess and activate.”15
It is this distinction that becomes the focus of contemporary “neo-
Aristotelian” accounts of non-instrumental epistemic value. Again, on this
view there is always some good in possessing knowledge and the value is not
instrumental.16 The contention is that truth and knowledge are always valuable,
always worth caring about for their own sake. One way to defend this view
is to argue that caring about truth and knowledge in this way is constitutive
of a flourishing life. To make this argument, we need to think about what
is required for excellence and how that these requirements necessitate that
epistemic goods possess some non-instrumental valuable. Two candidates that
have been suggested for what these constituents might be are authenticity and
intellectual integrity.17 To live authentically requires having a sense of self that
requires knowing what one cares about. If having such a sense of self is part
of a happy and flourishing life, then it seems some true beliefs are valuable
not because of their instrumental worth but because of what they are part of,
namely a happy and flourishing life. But again, the most this argument would
show is that a very specific set of true beliefs, namely true beliefs about what
I care about, are always valuable. But the view being argued for is stronger;
it is the view that any true belief, just in virtue of being true, has some non-
instrumental value. The value of authenticity can perhaps more productively
be invoked to explain the disvalue of false beliefs. If I have a lot of false beliefs
and fail to have an authentic grip on reality then it may seem my life is a kind
of sham and such a life cannot be a flourishing life. But even if is possible to
show there is always disvalue in avoiding false beliefs, this does not entail that
there is always some value in holding true ones.18
Perhaps thinking about intellectual integrity’s connection to happiness
can help support this stronger claim. A person with intellectual integrity will
be open to the truth and so not cling to beliefs when they are shown to be
false, and will stand up for the truth even when it is inconvenient, unpopular,
or maybe even dangerous. Why might this seem like a necessary constitu-
ent of a flourishing life? It has been argued that someone without it will
 Normative Defenses of Evidentialism 43
lack self-respect, and it may be that a flourishing life requires a certain
amount of self-respect. But it seems that what intellectual integrity requires
is to be willing to stand up for one’s own best judgment on matters of
importance. When we think about examples of those who either possess this
virtue or lack it, the examples all seem to concern weighty moral matters
like homophobia, racism, or the death penalty.19 So, even if we grant that
intellectual integrity is a necessary constituent of a flourishing human life,
the most that has been shown in that possessing certain traits, dispositions,
and virtues that will help one attain some truth and knowledge are con-
stitutive of “the flourishing life, the life lived well.” Appeals to the value
of integrity may do a better job than the argument from authenticity of
explaining that if one is to have a belief about something, one should try to
ensure that it is true. But more needs to be said to show that there is always
some non-instrumental value in holding a true one, even one about the most
trivial matters.
As I have been suggesting, the most obvious objection to the view that all
truth and knowledge is valuable is that content matters in deciding whether
a true belief is good to have. Some true beliefs, it seems, lack value or even
possess disvalue. One can point to many standard examples of what some
may call “dangerous knowledge.” These can include knowledge that has
dangerous content—such as knowing how to make nuclear weapons—or
knowledge related to cloning and genetic engineering, as well as knowledge
obtained through evil means—such as the knowledge attained through Nazi
experimentation.20 A response to this worry is to distinguish between all
things considered good and prima facie good. Something that is prima facie
good is defeasibly good. One can endorse the view that being true always
makes a proposition prima facie good to believe, but it may be better, all
things considered, not to believe it.21 The effort it would take to attain some
trivial truths may override the value of having them; the potential benefit
of self-deception may override the good of believing a truth; the dangerous
consequences of obtaining a truth may defeat its value.
But how can thinking about intellectual virtues and their connection
to flourishing support the view that even the most trivial true belief or
knowledge (like knowing how many threads there are in my carpet) is even
a prima facie good? Perhaps thinking about analogous morally good actions
that have prima facie value but that one might all things considered have
good reason not to do can help here. There is an abundance of morally good
actions I could be doing right now, but that would not be worth doing, for
all kinds of reasons. Helping somewhat frail people cross streets is always
prima facie good, but the time alone it would take me at this moment to find
someone suitable would be prohibitive and, given all the other facets of my
life, it is not the best thing for me to do, all things considered.
I think this analogy, rather than supporting the view that even the most
trivial true beliefs have value, actually highlights problems with it. If I had
superhuman powers and could help anyone who would benefit from crossing
44 Doxastic Norms
the street without sacrificing other things of importance, I would do it. If the
same powers allowed me to acquire beliefs about numbers of threads in
carpets, or grains of sand on the beach, it is not at all clear that I would be
motivated to employ my powers in such a way.22 Why is it even prima facie
good to believe the most trivial truth? The broadly Aristotelian framework
employed by those who argue for the non-instrumental value of knowledge
can explain why helping others is good and why one should develop traits
that will lead one to act in helpful and generous ways at the appropriate
times. The virtuous person who knows how to act rightly in the right way
at the right time will be able to see when helping one cross the street is the
virtuous action to take. It is rather easy to see how the act of helping a per-
son across the street is connected to the virtue of generosity or beneficence.
But how is this analogy supposed to help connect the believing of a trivial
truth with the intellectual virtues associated with integrity? The virtuous
person will believe rightly, will seek knowledge and wisdom, will cultivate
virtues of openness and carefulness. So, if there were a matter of importance
that required knowing how many threads were in the carpet, the virtuous
person would be particularly adept at finding the truth. But why any true
belief has some non-instrumental value has still not been made clear.
What does it mean for something to be instrumentally good? It could
mean that it is good simply as a means to an end. The most obvious
candidate of an instrumental good of this kind is money. Its value is solely
determined by what other goods it can obtain, and that a particular piece
of paper is deemed valuable is wholly arbitrary. If the arguments appealing
to flourishing were just meant to show that the value of truth and knowl-
edge is not conventional, arbitrary and so potentially fleeting, in this way,
then their arguments would succeed. But the claim being defended is much
stronger: the truth, simply in virtue of being the truth (the truth as such)
is good. But if we think about Kornblith’s discussion of epistemic value,
we see that a good can be considered instrumental in a sense that extends
beyond the simple means/ends relationship. If we care about achieving any
goals at all, we have a reason to care about truth and knowledge. Kornblith
contrasts this with accounts that argue for what he terms the “intrinsic”
value of truth, but it seems those claiming to establish the value of truth or
knowledge as such end up with an account very similar to that of Kornblith.
Kornblith thinks his account does not allow for the non-instrumental
value of truth, because truth would not be valuable but for its connection
with a kind of pragmatic value. It seems that this relationship allows for
another way of thinking about what it means for a good to be instrumental.
He is pointing to an asymmetrical relationship; one value depends on the
other, but not vice versa. It does not seem that appeals to flourishing succeed
in establishing that epistemic value exists independently of other values; it
would not be valuable but for its connection with the value of flourishing.
What is supposed to preserve the non-instrumental value of truth and
knowledge is that they are “constitutive goods” essential and necessary for
 Normative Defenses of Evidentialism 45
a flourishing human life. Somehow, their status is supposed to become equal
to that of the flourishing life of which they are a part. The only way this
could be is if it the connection between a flourishing life and valuing truth
and knowledge for their own sake cannot be broken, that part of what it
is to flourish, is that one values truth and knowledge for their own sake.
In Kornblith’s view, it is possible to break the connection between the value
of truth and flourishing. If a world existed where true beliefs and knowledge
ceased to be at all helpful in achieving one’s goals they would cease to have
value. They would also cease to have value if humans ceased having goals.
If we lived in a world where true beliefs had no benefits, then, in my view,
a proposition being true would not count at all in favor of its being believed.
Ultimately, I think that those who argue for truth’s non-instrumental value
aim to offer an account of truth’s value that is transcendent, viewing its
value as transcending the human world. My account of epistemic value
does rule this out for, ultimately, the value of truth is tied to what believers
value. And so if it were possible to conceive of a world where the truth was
detached from these values, then it would lose its value. But can we imagine
such a world? What would we—the world—and our relationship to the
world need to be like for this to be possible?
If we lived in a world where we were radically deceived by Descartes’s
evil genius or by the agents of the Matrix, almost all our beliefs would
be false. But such a world does not rule out that if one could come to have
a true belief that it would possess value. Some would take this to support the
view that true beliefs have some non-instrumental valuable. But the reasons
why someone in the Matrix world would want to take the red pill, or why,
in general, one may want to possess painful knowledge rather than remain
in blissful ignorance may well be reasons grounded in prudential and moral
values. One kind of life I can think of where true beliefs would cease to have
any use at all is one of complete inactivity. If nothing were ever done, if
nothing were ever thought about, then truth would serve no purpose. And,
in my view, if truth serves no purpose, it ceases to have value. So, if a pause
button were pressed on the world right now, no sense could be given to true
beliefs having value.
What these discussions of “epistemic value” or “cognitive goods” reveal
is that the value of believing what is true is dependent on the value of both
individual and collective flourishing. Having true beliefs is good, because
true beliefs are likely to be life enhancing. If a world existed where having
true beliefs and knowledge were not required for flourishing, then they
would cease to have value. The response here must be that such a world is
not possible, that it is like trying to conceive of a world where water is not
H2O. But anything that is not H2O is not water. Similarly, the contention
goes, any “flourishing” that does not contain valuing truth and knowledge
for its own sake is not really flourishing. But, of course, when urging the
importance of truth and knowledge, the examples given all appeal to this
world, to us, to humans, to what we care about.
46 Doxastic Norms
We can see that the position that epistemic goods possess non-instrumental
value can lead to some very strong (and, I think, intuitively implausible)
claims. In this view, any piece of knowledge—no matter how trivial—possesses
value, and that part of what it means to live a life well is that one, in principle,
would want to have true beliefs even about the most trivial matters. It seems
the motivation for making such strong claims is the worry about what the
alternative entails. Michael Lynch begins his book in which he argues for the
non-instrumental value of truth by pointing to the dangers of a “pragmatist”
view. He uses Stanley Fish as the representative of the view against which
he is arguing, calling it “antitruth.”23 Lynch seems to equate the view “what
makes a belief a good one is that it is useful” with the view that “truth has no
value.” Perhaps Fish holds both of those views but it is possible to separate
them. One can think truth matters enormously, that having true beliefs is
extremely valuable, while still thinking its value is instrumental. Truth mat-
tering and being valuable does not mean it is always matters and always has
value. When Lynch explicitly argues against the view that the value of truth
is instrumental, in a chapter titled “Truth and Happiness,” he acknowledges
that the most obvious reason why truth matters and why true beliefs are bet-
ter than false ones is that believing truly gets us other things we want. True
beliefs, for example he says, “keep us from getting run over by buses.” I have
been urging that the immense instrumental value of true beliefs is sufficient
to explain their importance. But to Lynch, and others, this picture fails to
capture something deep and important about truth and knowledge. He says,
“Most of us value truth as more than just a means to an end. We care about
having true beliefs” and we care about them “as such.”24 Further, as we have
seen, he argues that caring about them as such is constitutive of leading a full,
flourishing life.
The suggestion here is the care we have for things that we value instrumentally
is not enough to explain the way in which we care about truth; that is a kind
of “second class” caring, that somehow real caring can only be for something
we value for its own sake. We have seen that Kornblith explains why we care
so deeply about truth by showing how it is connected to other things we care
about. It seems I care very much about which color wire is the one that needs
to be pulled to diffuse a bomb; I care about whether there are peanuts in the
cookies if I am allergic. So when Lynch says here we care about true beliefs,
he means it in the restricted sense of caring, a caring that is more akin to the
way we care about our children. We don’t care about them because of any-
thing else or for any reason; we care about them for their own sake. But, in
explaining our care for truth and knowledge, other things we care about tend
to be invoked; for example, self-respect, authenticity, and happiness. None of
those values are epistemic but without them the value of truth and knowledge
evaporates.
If the value of truth and knowledge stands or falls with these other values, it
seems, according to Kornblith’s taxonomy, the value is not non-instrumental.
Perhaps calling it “instrumental” is misleading if we take that to mean only
 Normative Defenses of Evidentialism 47
that it is a means to an end, an “instrument” used for attaining the good of
flourishing. But we also say of something or someone that it is “instrumental”
when we are emphasizing its importance, the key part that it plays in achieve-
ment or success, for instance, “that defender was instrumental in protecting
the 4–3 lead.” This is not to say that the defender was a means to the end of
winning the game; I am not even sure that makes sense. But the value of the
defender’s actions are certainly dependent on the value of winning the game.
Similarly, it seems, epistemic value is dependent on non-epistemic value, and
the relationship is asymmetrical. Knowledge and truth, in general, have no
capacity for independent value. This is so, even if all flourishing human lives
require some knowledge. So, according to one prominent way of trying to
explain the value of truth and knowledge, epistemic value cannot stand on its
own; it cannot occupy a distinct normative realm.
How else might one argue for the autonomous nature of epistemic value?
One might just say the value of true belief is basic or final, that questions
about why it is good are misplaced. This view resists the idea discussed above
that a central task of epistemology is to explain why truth and knowledge
are good. For if it is just a basic good, then there is nothing to explain. Many
prominent teleological accounts take this approach. Their views parallel the
structure of consequentialist ethical theory. Just as act-utilitarianism begins
with the idea of happiness being the final value and then defines right action
by relation to the promotion of this final value, one defines right belief by
relation to what is taken to be the final epistemic value.25 While there are
many ways in which consequentialism can be critiqued, if a theory posited
as the final value to be promoted something that seemed obviously to have
only instrumental value, for example, money, then its conclusions about
how to act permissibly or rightly would be implausible.
It seems here we are faced with an issue regarding burden of proof. The
instrumental value of true belief and knowledge is clear and noncontentious.
No one denies that they have instrumental value. Those who claim these
are basic values think it is problematic to view them as only instrumental
values. It seems the burden of proof is on those who deny that epistemic
value is only instrumental to explain what is problematic about such a view.
What would motivate thinking that there is some value to truth and knowl-
edge beyond their instrumental value? I think what we find again is the
concern that if epistemic value is only instrumental, then it fails capture its
import and significance, and that the dictates to believe in a way to promote
it will lose their force. I have been urging, and will continue to argue, that
one can preserve the central importance of truth and knowledge even if this
importance is ultimately derived from their practical import. I will argue
that this pragmatic view of epistemic norms does a better job of explaining
which beliefs are permissible than the view that says it is always better to
have a true belief than a false one simply because true beliefs possess some
non-instrumental value and false beliefs possess some non-instrumental
disvalue.
48 Doxastic Norms
Of course, in a certain sense, one can carve out a particular, specified
normative domain in all kinds of areas where the values and norms are
specific to that area. Sartorial norms can tell one how one ought to dress,
and gustatory norms can say that one crème brulée is better than another. Or
any game one plays will dictate both whether a move is acceptable as well
as whether it was a good one. But epistemic value and norms are supposed
to differ from such “conventional” norms. They are norms that any rational
agent should follow, not just one who happens to have an interest in knowl-
edge. The point is that one ought to, as a rational agent, value knowledge,
whereas valuing food or clothes, or chess is a choice; you are not flawed if
you fail to follow or care about these norms. Further, it seems you are blame-
worthy if you fail to believe in accordance with these epistemic norms.26
And those who endorse the kind ethics of belief that so closely ties belief
with an epistemic aim or to the promotion of epistemic value think that this
domain imposes its dictates categorically upon believers. For example, in try-
ing to explain what kind of “ought” is meant in saying that one “ought” to
believe a certain way, Feldman argues that it should be understood as a kind
of “role ought”; the norms are dictated by the particular role one plays. For
example, given one’s role as teacher, there are certain ways one ought to
behave that would not apply to one who is not playing that role. But all of
us play the “believer role.” Feldman realizes this: “Forming beliefs is some-
thing people do. That is, we form beliefs in response to our experiences in the
world. Anyone engaged in this activity ought to do it right . . . it is plausible
to say that the role of a believer is not one that we have any real choice about
taking on. It is our plight to be believers. We ought to do it right.”27 Similarly,
in all the broadly teleological accounts of belief norms, we are flawed and fail-
ing if we do not achieve the proper aim associated with successful believing.
Whether one cares about clothes or coffee varies from person to person
and so the evaluations, norms, and judgments only apply contingently. But
there is a prevailing sense that one ought to care about morality, about one’s
own welfare, and about knowledge. I agree, but the reason one should care
about them is all the same; they are unified. If these normative realms are uni-
fied, then the standard evidentialist picture of doxastic norms is flawed in that
it assumes one can make sense of a distinctly epistemic realm that has nor-
mative force. Our beliefs serve the purpose of providing coherence, meaning
making, prediction, and navigation, both individually and collectively. Having
true beliefs and knowledge help us achieve our goals, flourish, and be excel-
lent human beings. In general, forming and maintaining beliefs in accordance
with evidential norms that produce knowledge is the best way to achieve these
goals. But even though evidential norms generally govern belief formation, it is
possible that some beliefs can help us achieve these goals independently of their
truth-value, or of their being evidentially based. It can, thus, be permissible to
hold some non-evidentially based beliefs.28 I will now turn to the question of
which non-evidentially based beliefs are permissible, if one accepts the view
that the value of true beliefs is ultimately linked to their pragmatic value.
 Normative Defenses of Evidentialism 49

NOTES

1. See footnote 9 in the Introduction for references that discuss this issue in
Hume.
2. Feldman, “The Ethics of Belief,” 682. Feldman says that he prefers to defend
evidentialism in this way rather than showing that following evidence will
help achieve a goal, the goal of true belief or knowledge, because “if the
oughts in question are supposed to be means to goals that people actually
have, then it seems that only people who do have [such] epistemic goals
would be subject to the relevant epistemic requirements.” But Feldman views
evidentialist requirements as unrestricted, applying to everyone regardless of
their goals: “All people epistemically ought to follow their evidence, not just
those who have adopted some specifically epistemic goals.”
3. Feldman, “The Ethics of Belief,” 683.
4. Ibid., 694.
5. Ibid.
6. Some are centrally concerned with explaining this normativity in such a way
that it also explains why it is more valuable than merely true belief. One
may think that post-Gettier, it must also say why it is more valuable than
justified true belief. In a sense, this distinction collapses because most theo-
rists explain why knowledge is more valuable than true belief in a way that
thereby explains why Gettier cases fall short and so explain why the true
beliefs in these cases are not as a valuable as knowledge even if they are, in
some sense, justified. Our concern here is not with whether one can explain
why knowledge is more valuable than true belief but, rather if one can make
sense of the normative category of the epistemic that both isolates it from
other normative domains and explains why it is of such central importance.
This category, as we have seen, can include truth, knowledge, rationality.
7. Greco, Achieving Knowledge, 4.
8. Sosa, Knowing Full Well, 3.
9. Zagsebski, On Epistemology, 5.
10. Although many people will use the “intrinsic” as equivalent to “non-
instrumental,” it is widely accepted that the intrinsic vs. extrinsic value
distinction is different from the final vs. instrumental value distinction. An
intrinsic value is something that has value by virtue of its intrinsic properties.
It does not seem that all those who see the value of truth and knowledge as
non-instrumental need be committed to the view that they are so because of
certain intrinsic properties.
11. This is the case with the discussions of knowledge found in Greco, Sosa,
and Zagezebski quoted earlier; these theorists all see themselves as espous-
ing a kind of virtue epistemology and all refer, at some point, to the role of
knowledge in the good life. In what follows, I will be drawing mostly on
Michael Lynch’s discussion of the value of true belief found in his True to
Life: Why Truth Matters. His book is the most extensive discussion centrally
concerned with defending the view that that truth and knowledge possess
some non-instrumental value.
12. Sosa, Knowing Full Well, 66. This is Sosa’s view of the value of knowledge.
He says it is a valuable commodity because knowledge of certain important
matters “normally make an important positive contribution as part of a life
that flourishes individually, or as part of the flourishing of a community.”
13. Greco, Achieving Knowledge, 97–98. To solve the “value problem,” Greco
explicitly turns to Aristotle. On Greco’s view, in cases of knowledge, someone
gets things right because of his own virtuous abilities, as opposed to getting
50 Doxastic Norms
it right through luck. It is evident, he says, that virtuous activity is “both
intrinsically valuable and constitutive of human flourishing” and so intellectual
activity as exhibited in cases of knowledge is one such manifestation of the
human good, or “the activity of the soul exhibiting excellence.”
14. Aristotle, Nicomachean Ethics, 89.
15. Ibid., 97.
16. Throughout this discussion, I will often refer to that value of “truth and
knowledge” as if they are the same. At times, Lynch talks about the value
of knowledge rather than truths. We shall see this is the case when he
considers the importance of having true beliefs about oneself; he calls this
the “argument of self-knowledge.” I think his point in focusing mostly on
the value of true beliefs is that, if he can show they all possess value, then
knowledge—a special case of true believing—will clearly be valuable.
17. This is how Lynch argues for the truth possessing non-instrumental value.
See Lynch, Why Truth Matters, Chapter 8.
18. Daniel Whiting argues that it is important to distinguish the view that false
beliefs are always bad from the view that true beliefs are always good and
that the first view is more defensible than the latter. He argues, “while
believing the truth does not have content-independent value, believing the
false has content-independent disvalue” (Whiting, “Does Belief Aim,” 222).
See also Whiting, “The Good and the True,” 219–242. I will argue that there
are times when the disvalue of false beliefs can potentially be overridden by
the practical advantages they provide.
19. These are all the examples Lynch uses in his discussion.
20. Lynch’s Chapter 4 is called “Truth Hurts” and begins with these examples of
“dangerous knowledge.”
21. Lynch, 46–47.
22. I find, in conversation, that people have conflicting reactions to this hypothetical
possibility. Some say, “Well, why wouldn’t I?” while others say, “Why would
I?” Now, perhaps one could find the same division considering the ethical
case, but it seems that one can respond with all kinds of reasons why one
should help others if it costs them nothing. For a helpful discussion of why
we might be attracted to the idea that if it did not incur costs, having trivial
true beliefs is valuable. See Whiting, “The Good and the True,” 228–232.
23. Lynch, 2.
24. Lynch, 119–120.
25. Selim Berker offers an extensive and illuminating discussion about what he
calls “epistemic teleology” and how closely it models standard consequen-
tialist views in ethics. His discussion convincingly reveals that this view of
how to approach the question, “What should I believe?” is almost universal
among diverse epistemological views; that one finds disagreement about the
value to be promoted but not that one believes well when the value of choice
is so promoted. He thinks this approach leads to an insurmountable prob-
lem and that a different non-teleological way of approaching the question
would do better. See Berker, “Epistemic Teleology and the Separateness of
Propositions.”
26. Stephen Grimm makes a similar point when discussing Sosa’s attempt to
carve out an isolated epistemic domain. Sosa suggests that one can identify
numerous “critical domains.” Each domain has fundamental values and we
can appraise or assess (hence, criticize) the derivative value of other items in
the domain in terms of how well they promote the domain’s fundamental
values. The fundamental values within a given domain, therefore, serve as
the goal around which the critical domain is structured. We have seen Wedg-
wood takes a similar approach in his explanation of truth as the fundamental
 Normative Defenses of Evidentialism 51
epistemic norm. While taking this approach can explain why even the most
trivial truths have value of a certain kind, Grimm makes the following point:
“By remaining agnostic about the domain-transcendent value of true belief,
Sosa seems to introduce a new problem—seems to, indeed, lose sight of one of
the most important aspects of our epistemic appraisals. For notice: when we
judge a belief to be unjustified or irrational, we seem to be doing more than
just evaluating (in this case, in a negative way) the skill or virtuosity of the
believer’s performance. In addition, we seem to be in some sense criticizing,
perhaps even reproaching, them for believing in this way” (“Epistemic Nor-
mativity,” 243–264).
27. Feldman, The Ethics of Belief, 676.
28. I am deliberately avoiding saying that it can be rational or justified to
believe in violation of evidential rules. My point is that, in some cases, one
can believe despite lack of evidence without having done anything wrong.
Depending on one’s theory of rationality or justification, these cases may be
described as cases where it is permissible to be irrational or hold unjustified
beliefs. Ultimately, I think such theories are flawed because their views rule
out beliefs that ought not be ruled out, but I am trying to avoid, as much as
possible, entering debates concerning the nature of justification or rational-
ity. Those who defend permissivism (see note 35 in Chapter 1) will be more
sympathetic to my view that more than one doxastic attitude can be permis-
sible given the same evidence but they will not describe these cases as believ-
ing against the evidence. Instead, they will say that whether the evidence
supports belief depends on the standards of rationality being employed, and
that differing standards can both be permissible.

REFERENCES

Aristotle. Nicomachean Ethics. Translated by Terence Irwin. Indianapolis: Hackett,

1999.
Berker, Selim. “Epistemic Teleology and the Separateness of Propositions.” Philosophical
Review 122, no. 3 (2013): 337–393.
Feldman, Richard. “The Ethics of Belief.” Philosophy and Phenomenological
Research 60, no. 3 (2000): 667–695.
Greco, John. Achieving Knowledge: A Virtue-Theoretic Account of Epistemic
Normativity. Cambridge: Cambridge University Press, 2010.
Grimm, Stephen. “Epistemic Normativity.” In Epistemic Value. Edited by Adrian
Haddock, Alan Millar, and Duncan Pritchard, 243–264. Oxford: Oxford
University Press, 2009.
Lynch, Michael. True to Life: Why Truth Matters. Cambridge: MIT Press, 2005.
Sosa, Ernest. Knowing Full Well. Princeton: Princeton University Press, 2010.
Whiting, Daniel. “Does Belief Aim (Only) at the Truth?” Pacific Philosophical
Quarterly 93, (2012): 279–300.
———. “The Good and the True (or the Bad and the False).” Philosophy 88, no. 2
(2013): 219–242.
Zagsebski, Linda. On Epistemology. Belmont: Wadsworth, 2009.
3 Unity of Norms
A Defense of Pragmatism

1. WHEN IS IT PERMISSIBLE TO HOLD A NON-EVIDENTIALLY

BASED BELIEF?

I have argued that defenses of evidentialism fail to show that all non-
evidentially based beliefs are impermissible. Conceptual defenses do not
show that it is impossible to form a belief for practical reasons. Normative
defenses reveal the value of following evidentialist rules: doing so helps us
to attain truth and knowledge, and, in general, truth and knowledge con-
tribute to both individual and collective flourishing. But if the value of truth
and knowledge is merely instrumental, then if a belief helps us flourish with-
out being evidentially based, it can be permissible to hold that belief. Thus,
I argue that even though evidential rules generally govern belief formation,
there are times when it is permissible to believe despite a lack of evidence.
For shorthand, I will refer to “non-evidentially based beliefs” as “practical
beliefs.”
Some people worry that if truth is not the only factor in assessing whether
one should believe a proposition, then it will be permissible to believe what is
false though comforting. Should the importance of truth be so dispensable?
The challenge to the pragmatist is to explain what distinguishes pernicious
practical beliefs from ones that are not. Given that truth and knowledge are
so valuable, engaging in practices that lead away from truth and knowl-
edge is problematic in every sense—prudentially, morally, and epistemically.
It will, thus, only be permissible to hold practical beliefs if doing so does not
rely on practices that undermine truth. If one has evidence that one’s belief
is false, and maintains the belief by deliberately ignoring that evidence, then
one’s practical belief is impermissible.
I will argue that it is sometimes permissible to violate evidentialist dicta
when faced with neutral evidence or no evidence at all. According to the
evidentialist the proper epistemic state in such a case is suspension of judg-
ment. I will then go on to consider times when it is permissible to believe
when the evidence is inconclusive.
 Unity of Norms: A Defense of Pragmatism 53
(i) “Meaning Making” and Belief When the Evidence
Is Silent or Neutral
We can begin to distinguish impermissible practical beliefs from permissible
ones by reflecting on the debate between Clifford and James. Here is Clif-
ford’s robust defense of evidentialism:

It is wrong always, everywhere, and for any one, to believe anything

on insufficient evidence . . . Belief, that sacred faculty which prompts
the decisions of our will, and knits into harmonious working all the
compacted energies of our being, is ours not for ourselves, but for
humanity . . . Every hard-worked wife of an artisan may transmit to her
children beliefs which shall knit society together, or rend it in pieces. No
simplicity of mind, no obscurity of station, can escape the universal duty
of questioning all that we believe.1

In many cases of believing against evidence, I think Clifford is right.

Believing against the evidence tends to be harmful, both to individuals and
to the collective. Thus, wise reasoning habits should be cultivated, and
acceptable deviations from evidentialist norms should be rare.
In James’s “The Will to Believe,” he responds directly to Clifford’s strong
evidentialist stance. He agrees that in many contexts, evidential considerations
will settle the matter. But on questions that cannot be decided by the evidence,
he says the following:

Our passional nature not only lawfully may, but must, decide an option
between propositions, whenever it is a genuine option that cannot by
its nature be decided on intellectual grounds; for to say, under such
circumstances, ‘Do not decide, but leave the question open,’ is itself
a passional decision—just like deciding yes or no—and is attended with
the same risk of losing the truth.”

James is right that when we have no evidence or it is equally balanced,

belief is sometimes permissible.
The evidentialist will say the only permissible doxastic attitude in such a
case is suspension of judgment. For this is the only way one can be sure to
avoid false beliefs and, as previously discussed, there is disvalue in having
false beliefs. But what James is pointing to here, and goes on to discuss
further, is that by choosing this option, one is deciding that the potential
disvalue of having a false belief outweighs the value of holding the belief.
Certain kinds of beliefs about the significance or meaning in life events—
even when there is no evidence to support them—enrich our lives and help
us thrive. There is some evidence from cognitive psychology that suggests
that not only have humans evolved to be able to decode and interpret mean-
ing from our fellow creatures, but also that our brains have also adapted
to see meaning in life events.2 To see both other people and life events as
54 Doxastic Norms
meaningful is a distinctive human capacity and one lacking in those with cer-
tain cognitive disorders. This meaning making could translate into a theistic
or religious perspective but it need not.
In a recent article, Ryan McKay and Daniel Dennett examine a number
of cases of “misbelief”—where one has a belief “that to some degree parts
from actuality”—and argue that some of these are actually adaptive, mean-
ing that they have been evolutionally favored because they contribute to the
propagation of the species. In summing up their findings, they echo what
I have been arguing about beliefs, namely that their purpose is not simply
to track truth; “because our belief states have complex effects beyond sim-
ply informing our deliberations—they flavour our attitudes and feed our
self-images—and complex causes that can create ancillary effects, such as
triggering emotional adjustments and immune reactions, the dynamics of
actual belief generation and maintenance create a variety of phenomena that
might be interpreted as evolved misbeliefs.”3
Feldman, and others, would clearly see such beliefs as irrational, because
they violate evidentialist injunctions. That evolution has disposed us to
form “meaning-making” beliefs does not settle the question of whether is it
permissible for us to do so. Some traits that evolution favored, perhaps like
“natural” male aggression when faced with rivals, are ones that we may not
evaluate positively. It may even seem that that the evolutionary story is irrel-
evant to the question of permissibility. I take it as relevant. Recall that one of
the ways of defending evidentialism is by appeal to the proper functioning of
our cognitive system. In explaining the nature of belief and its aim, Velleman
argues that human “cognitive systems” regulate an agent’s beliefs “in ways
designed to ensure that they are true, by forming, revising and extinguishing
them in response to evidence and argument.” But McKay and Dennett’s find-
ings suggest that if we strictly followed evidentialist rules, we would not do as
well as if we deviate from them in special circumstances. For example, in one
study of AIDS patients at a time when life expectancy was not long, having a
realist acceptance of death was found to be a significant negative predictor of
longevity, with high scorers typically dying months earlier than low scorers.4
Now if someone is convinced that the practical value of belief is irrelevant
to its permissibility, then these arguments and studies will seem beside the
point. The response will be that even if such beliefs are practically valuable,
they are epistemically impermissible. I have argued (in Chapter 2) that such
a response, which attempts to isolate the epistemic as a distinct normative
domain, is problematic. The reason why true beliefs are valuable is that they
help us to flourish and contribute to our overall good, and so these kinds of
meaning-making beliefs may be another way of serving this greater good.
The evolutionary story does not tell us which beliefs are permissible and
which are not. But it does support the idea that the purpose, aim, or goal of
belief is more complex and varied than truth.
Can any evidentialist rule be overridden if the practical stakes are high
enough? Some obviously false beliefs, or ones that go against all evidence,
 Unity of Norms: A Defense of Pragmatism 55
can be very useful. Imagine a plane crashes in the middle of winter, high in
the Rocky Mountains, and some people survive. It seems that even if all
evidence points to the likeliness of their imminent demise, it would be good
for the survivors to believe—against overwhelming evidence—that they can
live. There are abundant examples of people ignoring the evidence of the
doctors who tell them they only have a few months to live and such ignoring,
perhaps, allows them to live longer. As we have seen, some of the beliefs that
McKay and Dennett deem adaptive are “positive illusions” of this type.
In cases like this, I think we should excuse these believers. However, some
of these cases may still be cases of pernicious believing. It is important that,
in general, we do not violate the rules that tell us to attend to and not ignore
evidence. The situation is analogous to the value of truth-telling. Telling
a lie in a particular case can be very valuable, but this does not undermine
the general moral rule that says lying is wrong. I think the clearest case of
permissible practical belief is when the evidence is neutral; in such a case,
there is nothing pernicious about believing in a way that contributes to your
well-being (and perhaps collective well-being). But why am I so complacent
about violating this particular evidentialist rule? Isn’t violating this general
rule akin to violating other epistemic rules? Isn’t what makes it a rule that
following it ultimately contributes to the good of humanity, as Clifford says?
It is significant that all of Clifford’s examples of believing on insufficient
evidence involve ignoring, suppressing, or deliberately not attending/failing
to attend to evidence; these are all vices closely tied to potentially problematic
actions and policies. In one case, people drown because a shipowner ignored
the evidence that his boat was not sea-worthy. In another, hastiness and
prejudice leads to innocent people being falsely accused. But if one is in
a situation where the evidence is neutral or silent, then one is not actively
ignoring, suppressing, or failing to attend to evidence.
Why does it seem that following the neutrality rule serves our epistemic
goals? Well, clearly, if one withholds belief in such contexts, one will
avoid believing something that is not true. Of course, if one suspended all
judgment, one would also not believe anything false. But most evidentialists
are not radical skeptics and they would surely point out that such a position
is not practical. If one has no beliefs, then one avoids holding false beliefs,
but one also fails to gain true ones and cannot achieve many of the practical
goals that having such beliefs could help achieve. These cases help support
the view that doxastic norms are not wholly evidential. We do not do well
as believers if we strictly and universally adhere to evidentialist rules.
In his discussion of the neutrality rule, James points out that knowing truth
and avoiding error “are not two ways of stating an identical commandment,
they are two separable laws . . . Believe truth! Shun error!—these, we see,
are two materially different laws; and by choosing between them we may
end by coloring differently our whole intellectual life.”5 Clifford, and other
evidentialists, view the avoidance of error as paramount: one must suspend
judgment so as not to “incur the awful risk of believing lies.”6 James thinks
56 Doxastic Norms
that such a position reveals Clifford’s “own preponderant private horror of
becoming a dupe.” James says he cannot follow Clifford, for although he
also has “a horror of being duped; [he] can believe that worse things than
being duped may happen to a man in this world: so Clifford’s exhortation
has to [his] ears a thoroughly fantastic sound. It is like a general informing
his soldiers that it is better to keep out of battle forever than to risk a single
wound.”7 For James, one of those worse things would be missing important
and essential truths about life’s meaning and value.
Whether we follow James or Clifford here is a matter of choice, one that
is an expression of our “passional life,” namely an expression of our desires,
hopes, and fears. Thus, when the evidence is neutral (or close to it), there
is some freedom involved in what we believe. And it does not seem to be
irresponsible to fail to suspend belief, in such contexts, when the practical ben-
efits are high. Rather than try to force beliefs of this kind into an evidentialist
framework, I think it is better to expose this framework as impoverished.
When it comes to questions about, for example, what, if anything provides
meaning or significance in life, or what happens after death, we have some
flexibility in what we can permissibly believe.
How much flexibility? Is any belief permissible if its practical advantages
outweigh the disvalue of holding a potentially false belief? Beliefs that are
very closely tied to immoral actions are not permissible, but this is not because
they violate evidentialist rules. In a recent discussion, John Bishop attempts
to articulate criteria for when a “doxastic venture” is permissible under
conditions of “evidential ambiguity.” Conditions of evidential ambiguity
are defined as “conditions where our total available public evidence neither
shows the proposition’s truth nor its falsehood to be significantly more
probable than not, and where the total evidence is systematically open to
viable overall interpretation, both on the assumption that the proposition
is true and on the assumption that the proposition is false.”8 As we have
seen, the evidentialist would say the only appropriate attitude in such cases
in to suspend belief. Of course, many people do believe in such cases—they
engage in what Bishop calls a “doxastic venture.” The question, then, is
when is such a venture permissible. James argues that such ventures are
permissible only under quite severe constraints. One such constraint is that
it must matter whether one takes the proposition under question to be true
or false—it must make some practical difference.
Why? What is wrong with believing something utterly trivial when the
evidence is neutral, like for example, that the first person who flipped a coin
in Berlin today got heads (assuming one has no evidence about this mat-
ter)? First, it seems like this is one of those beliefs I may well be unable to
form because I would find no reason to think it is true, and would think it
is very likely false. One cannot believe something while thinking it false; this
connection between belief and truth holds. Central, meaning-making beliefs
are not only possible, but seem to be little affected by arguments and evi-
dence. Second, if I were asked why I believed it, it seems I would have nothing
 Unity of Norms: A Defense of Pragmatism 57
to say, no reason to point to. There is an important difference between hold-
ing a belief for no reason and holding a belief for very important practical
reasons.
What if you had a good practical reason to form this belief, say some-
one offered you a huge amount of money to do so? That we are unable to
form a belief against the evidence when offered money to do so is often
taken to show decisively that we cannot believe for practical reasons.9 While
I think there are some beliefs that one cannot believe for some practical
reasons, I do not think we can generalize from examples of this kind to the
conclusion that practical reasons are always the wrong kind of reasons for
belief. It is quite likely that there are many actions one could not perform
no matter how high the monetary incentive like, for example, killing an
innocent person or jumping out the window, but this would not tell us that
one can never act for reasons of this kind. To object that one could perform
these actions but one chooses not to begs the question that we have a choice.
In both cases—that of believing and that of acting—one is being asked to
do something that goes against a deeply entrenched view of who one is and
what one values.
To these Jamesian constraints, Bishop adds two kinds of moral constraints:
both one’s motivation for believing, as well as the content of the belief, must
be morally acceptable. When we blame someone for having a racist belief,
although part of this judgment is owing to the belief being ill-founded, part
of the blame stems from knowledge of how closely tied such beliefs are to
treating others unjustly. Similarly, some beliefs formed under conditions of
evidential ambiguity are closely tied to questionable moral practices. If your
belief about the afterlife entails that I will suffer eternal damnation, this will
impact the way you think about me and treat me. Or, perhaps even more
worrisome, beliefs about the afterlife might seem to condone suicide and
murder for a God-serving cause. It is likely that the strong evidentialist dictum
that supports suspending judgment when the evidence is neutral is motivated
by the worry that permitting freedom here will permit morally problematic
beliefs. Yet beliefs about a greater power or in something transcendent can
also bring one to care more about others, to recognize a connection among
all humans and all nature, to view the world in a more positive and beauti-
ful light than if one chose to suspend judgment. In these moments, it seems
James is right that one is faced with some freedom regarding what to believe.
As we have seen, many philosophers deny that we can choose what to
believe; they think that suspension of judgment is simply what happens
when we think that our evidence is indecisive, because they think that what
it means to believe is to take oneself to have evidence. Such philosophers
will think that James must have come to see the practical value of his
religious belief as evidence for its truth. I have argued (in Chapter One)
that it is possible to believe even if one takes oneself to have no evidence for
the proposition believed. Beliefs are not wholly passive.10 If we think about
cases of “doxastic ambivalence” when we are really unsure what to believe,
58 Doxastic Norms
it does not seem like the only factors that will tip us over to one side or the
other are factors beyond our control. In trying to decide what to believe in
such a case, one can think carefully about what kind of person one wants
to be, to recognize that the beliefs we adopt both express and shape who we
are. And if one believes because one thinks that doing so better expresses the
person one wants to be, even while recognizing the evidence is insufficient,
then one does not thereby hold an impermissible belief.

(ii) Inescapable “Framework” Beliefs

Another set of permissible practical beliefs includes those that are indispens-
able and perhaps inescapable. Our belief that there is an actual world of
external objects (and persons) is likely one of those. It is not a belief that is gen-
erated by any argument; we just find ourselves with it. And just a little reflec-
tion and exposure to standard skeptical arguments show that it is a belief that
cannot be justified in any standard evidentialist way. This is because when
the evidentialist argues that following one’s evidence will lead to truth, it is
assumed that many of these particular truths are about this external world.
Of course, many philosophers think that this belief can be justified or sense
can be given for evidence supporting it. Some argue that the “external-world
hypothesis” is a more probable explanation for our sense experience than are
other possible hypotheses involving some kind of deception.11 Yet, it seems
problematic to appeal to probability without already assuming the existence
of an external world. Hume sums up the problem in the following passage:

It is a question of fact, whether perceptions of the senses be pro-

duced by external objects, resembling them: how shall this question be
determined? By experience surely; as all other questions of a like nature.
But here experience is, and must be entirely silent. The mind has never
anything present to it but the perceptions, and cannot possibly reach any
experience of their connexion with objects. The supposition of such a
connexion is, therefore, without any foundation in reasoning.12

One can take the Moorean response—this is perhaps the currently

most-favored response—and say that my senses provide me with all the
evidence I need and that I need not find any way of ruling out skeptical
hypotheses to rely on my senses as evidence.13 In one way, I agree with this
response. I don’t think the permissibility of this belief depends on the ruling
out of these hypotheses. But I am not convinced that our senses provide us
with evidence of its truth. The belief is about the nature of existence inde-
pendent of our senses; how can our senses alone tell us anything about such
existence? I am not here going to consider all the possible ways that one can
respond to skepticism to show that this belief is, in fact, evidentially based.
What I want to urge is that whether there is evidence or not is irrelevant to
its permissibility. Its permissibility does not depend on the success of these
arguments against skepticism.
 Unity of Norms: A Defense of Pragmatism 59
Nothing about my view depends on there being an absence of evidence
for belief in the existence of an external world. If there is evidence, then such
belief is permissible for both the evidentialist and the pragmatist. But if there
is no such evidence, I can still allow this belief, whereas the evidentialist
would deem it impermissible. This again points to an advantage of recogniz-
ing that doxastic norms are not wholly evidential.
Belief in the external world, perhaps more than any other, is one required
for coherence, and meaning making. It is not even clear if it is possible
to follow the neutrality rule regarding this belief. Perhaps in moments of
deep reflection or under the influence of powerful drugs, we can manage
brief suspension of this belief. But those moments will likely induce anxiety
and feelings of disassociation. We would not think of a mind lacking this
belief as a healthy mind. And just like the beliefs about life events hav-
ing significance, the belief that our perceptions are caused by and fairly
accurately represent a mind-independent reality seems distinctly human. It
doesn’t seem that such a complex belief is required for survival. But it seems
we do require the belief for some other reason, given that it is one of the
most universal and persistent of all beliefs. One may take its universality
and persistence as a sign that it is true. Indeed many of Hume’s contempo-
raries assumed that if “Nature” had seen fit to provide all humans with such
a persistent belief, then it must be true. But Hume does not say this. He says
instead, “Nature has not left this to . . . choice, and has doubtless esteemed
it an affair of too great importance, to be trusted to our uncertain reason-
ings and speculations. We may well ask, What causes induce us to believe
in the existence of body? but it is in vain to ask, Whether there be body or
not? That is a point which we must take for granted in all our reasonings.”14
There is an important difference between this kind of belief, or a belief
that one’s children are not automata, and more ordinary beliefs, such as,
“Those animals in the zoo are zebras rather than cleverly painted mules,” or
“The train leaves at 5 o’clock.” Even if all competing epistemologists agreed
that I know these more ordinary empirical propositions—they are based on
sufficient evidence, they are the output of a reliable cognitive mechanism,
they are formed in an environment where no relevant alternatives would
undermine claims to know—it could turn out that I am wrong, and nothing
terrible would happen. We are used to ordinary knowledge claims being
defeated. Our epistemic position regarding the falsity of sceptical hypotheses
(like the hypothesis that we are brains in vats or that our children are actu-
ally automata) is importantly different from our position regarding things
we ordinarily claim to know.
In his discussion of how ordinary knowledge is possible, Robert Nozick
says even if all the evidence available to him would be the same if his children
were automata, so that he cannot know that his children are not automata,
this does not undermine his belief that his children are not automata. Further,
he says, he would stake his life on his children not being automata.15 This
is dramatic language but I think it points to something important; there is a
sense in which our lives depend on holding certain beliefs, at least the kind
60 Doxastic Norms
of life that makes sense to us. Even if such a belief cannot be grounded in
evidence, it is not faulty in any way.
Again, if the evidentialist agrees that the evidence cannot support such
beliefs, he would say that the right cognitive attitude in such a context is
suspension. So, if it turned out that arguments showing evidence can support
such beliefs fail, then I ought to, according to the evidentialist, suspend judg-
ment about whether or not my children are really automata. What I think
Nozick’s discussion implies, is that these kinds of belief are different from
most others; their permissibility does not depend on the success of an argu-
ment showing that they are based in evidence.
Instead of trying to figure out how we actually can have sufficient evidence
for these fundamental beliefs, the evidentialist may instead insist again
that, from a purely epistemic perspective, suspension is the right cognitive
attitude. We do not require this belief to do well epistemically. This is
especially true if the epistemic value we promote through evidentialism is, as
Feldman argues, rational belief rather than truth or knowledge. Remember,
for Feldman, following evidentialist rules will allow for truth or knowledge
when possible. In a world where some skeptical hypothesis holds, knowl-
edge isn’t possible, but one can still do well epistemically if one’s beliefs are
based on the evidence available. The evidentialist may say that the consider-
ations I’ve just put forward may show that it it’s permissible in some sense
to not suspend judgment about the existence of other minds, but they are
not epistemic considerations. I have shown (in Chapter 2) how hard it is for
even the evidentialists themselves to keep these evaluative domains distinct.
I have further argued that to make sense of the value of truth, knowledge,
or rationality, we must go beyond the mere epistemic.

(iii) Beliefs about Loved Ones

I have been arguing that it is permissible to violate evidentialist rules if
they require us to give up beliefs that seem essential to our lives. One may
wonder if, in the context of friendship and family, one can also be permitted
to believe in a non-evidentialist way. First, consider how you respond to evi-
dence that seems to impugn your good friend’s character. Sarah Stroud has
argued that it would be wrong in such a case to weigh the evidence as you
would in any other situation. It seems the demands of friendship are such
that your beliefs about your friend should be somewhat unresponsive to the
evidence. Stroud suggests that you should interpret what you hear in a less-
damaging way than would a stranger, looking for alternative interpretations
to the obvious, and damning, ones. Just as a certain way of behaving con-
curs with our ideas of what a good friend would do, so, she argues, does
a corresponding manner of believing.16 Perhaps being a good, loving parent
also permits one to be similarly unresponsive to evidence.17
These cases often involve suppressing or ignoring evidence, and so—
like in the case of the plane crash survivors—would count as cases of
 Unity of Norms: A Defense of Pragmatism 61
pernicious, though, in many cases, excusable, believing. But what about
cases where the evidence is close to evenly balanced and so does not require
radical and problematic self-deception? Must one, in such cases, suspend
judgment about whether one’s child is a drug-dealer or whether one’s friend
acted maliciously? These are difficult cases and how we ought to evaluate
these beliefs will depend on the context. If we think about the constraints
discussed earlier about when a doxastic venture is permissible, we must ask
both what the motivation for belief is and what the likely consequences of
believing would be. Presumably, the motive to believe that one’s child is
not a drug-dealer or that one’s friend did not act maliciously stems from
feelings that are generally good ones—feelings of love and generosity. If the
motive comes from one’s fear for one’s own reputation, then we may assess
the belief differently. While the content of these beliefs are not themselves
problematic, it seems the same motives can easily lead one to keep believing
even if the evidence started to tip in favor of the unwanted belief. In cases
like this, belief rather than suspension is permissible, but one needs to take
extra care to ensure that one does not end up ignoring evidence so as to
maintain the wanted belief.

(iv) Religious Belief

Finally, I turn to belief in God and other religious beliefs. The way I have
described many of the beliefs that I think permissible to hold despite lack of
evidence is similar to the way those with religious beliefs describe them—as
central and foundational, providing meaning and coherence. Indeed, many
defenses of the legitimacy of religious beliefs argue that these beliefs are basic
much in the way that belief in the external world is, and that these beliefs
essentially stand and fall together. Again, it is such a defense of religious belief
found most prominently in the writings of Alvin Plantinga and Nicholas Wolt-
erstorff that led Adler to formulate his very strong version of evidentialism.18
In examining the arguments of the Reformed epistemologists, we find
many different claims being made, some with which I concur and many
which I question. One theme in these writings is a critique of classical foun-
dationalism, the view that all our beliefs ultimately rest on some basic beliefs
that provide justification, without their own justification stemming from
any other beliefs. Classic foundationalism further holds that the only beliefs
that can legitimately serve this foundational function (and so are properly
basic) are beliefs that are either incorrigible or self-evident. Plantinga argues
that these criteria are problematic; his critique is not unique, stemming
back to insights one finds in, among others, the writings of Wilfrid Sellars.19
Briefly, it seems that many of the candidates for proper basic beliefs actually
require, to maintain their special status, that we also accept a number of
other propositions to be true. Suppose that I take a perceptual belief as basic,
for example my belief that the table is red. I only take this to be self-evident
because I take certain generalizations, such as “my visual perceptions are
62 Doxastic Norms
reliable,” as true. If one really restricts properly basic beliefs to those that
are self-evident or incorrigible without recourse to other background beliefs,
it becomes increasingly difficult to see how they will provide the justificatory
support for most of our ordinary empirical beliefs.
Many view such criticisms as revealing that foundationalism is untenable.
Plantinga, however, retains the foundationalist structure of justification but
argues that the criteria for a belief being properly basic should be altered.
He does not offer an alternate set of conditions for when a belief is properly
basic. Instead, he thinks that one should begin by determining if the belief is
grounded in experiential conditions rather than being formed on the basis of
any other belief.20 If so, then the belief is basic. It is properly basic, meaning
that a “rational noetic structure” can have such beliefs as foundations, if
these conditions justify the belief. According to Plantinga, these conditions
are justifying if one is under no obligation to defend them as legitimate.
For example, although, the belief “I had breakfast this morning” is neither
self-evident nor incorrigible, it still counts as properly basic, given that it is
not grounded in any other belief, and, if you have no reason to doubt your
memory, the conditions that ground your belief justify it as well.
On this view, Plantinga claims it is perfectly possible for belief in God to be
properly basic, that one’s “rational noetic structure may very well contain
the belief in God among its foundations.”21 This is even more plausible
when we realize that what counts as properly basic varies according to
context and persons. So the belief “I see a tree” is properly basic when
certain conditions hold, when I have a certain kind of “being appeared
treely to experience.” And so it is possible, according to Plantinga, that for
some people, belief in God is properly basic. This does not mean such a
belief is groundless. It is grounded in certain experiences, just as perceptual
beliefs or memory-beliefs are so grounded. If certain conditions cease to
hold, then these beliefs cease being properly basic.
Plantinga recognizes that his view seems to be open to the objection that
just about any belief can count as basic in the right circumstances. He calls
this concern “The Great Pumpkin Objection.” How, then, does Plantinga
address this objection? He says the following: “The Reformed epistemolo-
gist can properly hold that belief in the Great Pumpkin is not properly
basic, even though he holds that belief in God is properly basic even if he
has no full-fledged criterion of proper basicality.”22 He does, of course,
think that these beliefs are relevantly different, so that one is excluded as a
candidate for being properly basic. The difference, according to Plantinga,
stems from the conditions that ground and justify belief in God, condi-
tions that one cannot find to ground or justify belief in the Great Pumpkin.
The conditions that ground and justify someone’s belief in God are certain
experiences and feelings, such as having a “deep sense that God is speak-
ing to him,” feeling guilty “in God’s sight,” feeling forgiven by God, ask-
ing God “for protection and help,” and having “a spontaneous sense of
gratitude.”23
 Unity of Norms: A Defense of Pragmatism 63
It also seems integral to his distinguishing belief in God from belief in the
Great Pumpkin that such experiences are widespread, that “there is a natu-
ral tendency to see his hand in the world around us.” As we have seen,
some philosophers think that our natural tendency to believe in an external
world is evidence that there’s an external world. I, like Hume, do not think
we can take our natural tendency as evidence, but I do think it is permis-
sible to believe in the external world in the absence of evidence. So it seems,
if we have a natural tendency to believe in God, I should accept that it is
permissible to believe, despite not having evidence for this belief. What I have
to say about this depends on what is meant by “belief in God,” a question
that the Reformed epistemologists do not answer consistently. It is this lack
of consistency, I will argue, that undermines their argument.
Again, a belief is basic if it is not grounded in another belief. But it is
only properly basic if one is not violating any “epistemic duties and is in
one’s epistemic rights in accepting it.”24 Consider a case where one’s belief
in the Great Pumpkin is based on some visual perception and, so, counts as
basic. Retaining this belief would require one to suppress or ignore a great
deal of evidence. I have argued that any belief that needs such ignoring
and suppressing is one that, though perhaps excusable in certain circum-
stances, is still pernicious. Plantinga argues that belief in God requires no
such ignoring. So, now we are back to discussing whether there is evidence,
that if attended to, undermines this belief, a discussion it seemed that the
Reformed epistemologists wanted to forestall by arguing that one could
legitimately believe without evidence.
Now, Plantinga does argue that these “justifying-conferring circumstances,”
such as certain experiences and feelings, are defeasible. So, one’s belief in
God may cease to be properly basic if one confronts arguments that lead
one to question the authority of those experiences. But he does allow that it
is possible for even a reflective believer to confront these arguments, coun-
ter them, and violate no duties in holding on to this belief. This possibility
only seems plausible if the content of the belief “in God” is kept extremely
vague, meaning something like what Williams James terms the “religious
hypothesis.” Here is how James describes it: it is the idea that “the best
things are the more eternal things, the overlapping things, the things in the
universe that throw the last stone, so to speak, and say the final word.”
What James means here is not transparent and is open to interpretation,
but it seems he is thinking of the feeling of something greater or beyond
what can be measured materially.
If there is “a natural tendency” to believe in God, it would, again, have
to be of this very vague variety. But Plantinga and the other Reformed
epistemologists do not mean this when they talk of belief in God. They are
talking about a specific Christian belief and, for them, belief in God has a
very specified content, content that includes some quite specific doctrinal
beliefs. Plantinga begins his argument by making this clear: he says that by
belief in God he means the “belief that there is such a person as God” and
64 Doxastic Norms
he specifies which God: “the God of Abraham, Isaac, and Jacob, the God
of Jewish and Christian Revelation: the God of the Bible,”25 and he appeals
to Christian theologians of the Reformed tradition to support his argument.
Further, he actually scoffs at those who think of this belief in a less-restricted
sense, stating emphatically that he is concerned with a belief that “there
exists a person of a certain sort—a being who acts, hold beliefs and has aims
and purposes. This person . . . is immaterial, exists a se, is perfect in good-
ness, knowledge, and power, and is such that the world depends on him for
its existence.”26
But the more definite and restricted the belief becomes, the more difficult
it becomes to distinguish it from a belief in the Great Pumpkin. It seems
extremely implausible that the Christian believer will be able to support his
belief without ignoring or suppressing a lot of evidence. And just as the
Great Pumpkin believer must contend with the vast amount of non-believers,
so, too, the person who believes in the Christian God must contend with the
millions of people who do not share his belief. When it comes to this belief,
we are certainly not in the same situation as we are with respect to our belief
in the external world. I have not claimed that any evidence supports the
belief in external objects, but we also have no evidence that undermines it.
There do not exist, for example, millions of people who hold beliefs directly
contrary to the view that there is an external world. The only plausible
candidate for a universal tendency to believe in God, given the diversity of
cultures and faith, is some very general one with very vague content. As
Plantinga’s argument proceeds, it seems more and more like this is the kind
of belief he has in mind. But that is a very different belief from the Christian
one that is supposed to be his concern.
I have tried to steer away from talk of rationality or justification, given
the diversity of views of what is meant by such concepts. Is our belief in the
external world rational? Here, I am tempted to turn to Wittgenstein, who
does not think the certainty we have about these proposition is unjustified.
He says, “I want to conceive of it as something that lies beyond being justified
or unjustified; as it were, as something animal.”27 Interestingly, according to
Wolterstorff’s conception of rationality, the belief is rational. He argues that
one is rational in holding a belief if one is properly governing one’s belief
dispositions, that is, the various mechanisms that are involved in belief for-
mation.28 This idea of belief governance is quite extensively discussed and
elaborated, but one of the central ideas is that one can be rational in holding
a belief if one has “no adequate reasons to surrender it.”
Some beliefs, though not inferred from any other, and so believed
“immediately,” can still be rational in this sense. It seems right to say that the
remote possibilities that the skeptic appeals to do not give us adequate reason
to surrender our belief in the external world, especially given how central and
valuable this belief is. So, for Wolterstorff, the question to ask when assessing
someone’s belief in God is the following: “Might a person’s being in the situ-
ation of believing immediately that God exists represent no failure on his part
 Unity of Norms: A Defense of Pragmatism 65
to govern his beliefs as well as can rightly be demanded of him with respect
to the goal of getting more amply in touch with reality?”29 Though it is not
hard to think of a context where this is conceptually possible, the likelihood
of there being an actual person in this state becomes less and less likely the
more we specify the content of the belief that “God exists.”

2. WHAT DO I MEAN BY “BELIEF”?

The evidentialist might object that what I am calling “belief” is not really
belief but something closer to desire, or to hope. One common way of
differentiating beliefs and desires is in terms of “direction of fit,” a distinction
first suggested by Elizabeth Anscombe.30 These states are essentially different
because we want our beliefs to “fit” the world, but we want the world to
“fit” our desires. Here is a standard formulation of the distinction between
belief and desire:

The distinction is in terms of the direction of fit of mental states to the

world. Beliefs aim at being true, and their being true is their fitting the
world; falsity is a decisive failing in a belief, and false beliefs should
be discarded; beliefs should be changed to fit with the world, not vice
versa. Desires aim at realization, and their realization is the world fitting
with them; the fact that the indicative content of a desire is not realized
in the world is not yet a failing in the desire, and not yet any reason to
discard the desire; the world, crudely, should be changed to fit with our
desires, not vice versa.31

I think this description mischaracterizes both beliefs and desires, placing

them in completely separate domains when, really, they are much more inter-
dependent and overlapping. What leads to an alteration or elimination of
desire; what counts as “a failing” in desire or as a reason to discard it? This
is a difficult question and I cannot here offer anything like a complete or
proper answer, but it does not seem like facts about the world are irrelevant
in assessments of desire. Suppose that after a break up, I have the desire to
get back together with my ex. I will do what I can to “change the world”
for it to fit with my desire, but as time goes on and it becomes clearer that
the world is such that my desire cannot be fulfilled, I will likely judge that
I should not have that desire. The desire will not disappear immediately
upon recognition of the world’s failure to “fit” it—if it did heartbreaks
would not take so long to get over—but neither do beliefs disappear the
moment we see they do not fit the world. If a belief is deeply held, one that
is central to one’s identity, one may well spend a lot of mental energy seeking
ways of understanding the world so that one’s belief does fit it. If the world
begins to reveal that my belief in my partner’s fidelity is not fitting, I will
seek out aspects of the world that will allow it to fit. At a certain point
66 Doxastic Norms
this may become impossible, perhaps at the point when I actually witness,
through direct sense perception, the world thwarting my belief. Then I will
discard it on pain of being delusional. But this seems quite similar to the case
of desire. I will try to make the world fit my desire, but if I realize there is
no way to do so, I will give up the desire, again on pain of being delusional.
Of course, there are important differences between beliefs and desires.
Beliefs, like intentions, involve one’s making up one’s mind on an issue,
adopting something like a commitment, taking a stand.32 While achieving
clarity about the nature of such commitment is difficult, whatever it amounts
to, it is not included in having a desire. But, to treat beliefs as purely cognitive
states, on par with, for example, supposing does not seem right. At least in
ordinary practice, we find people using “I feel” and “I believe” interchange-
ably. I have often told my students when they use the expression “I feel” to
convey what they believe that I don’t care about their feelings. But I think
this conflation points to something important. It is not simply metaphoric
in the way saying that I “see” the answer to a problem or I “see” what
someone means is. Believing something feels different than just supposing it
for the sake of argument and part of this might have to do with the feeling
of commitment. Jonathan Cohen argues that to have a belief that p is a
disposition to have certain feelings, in particular a disposition to “feeling it
true” that p. He contrasts “credal feelings” with “affective feelings,” such
as anger and desire, where the former “share the distinctive feature of con-
stituting some kind of orientation on the ‘True or false?’ issue,” whereas the
latter constitute some kind of orientation on the “good or bad?” issue.”33
While I take it as an overstatement to say that beliefs simply are a disposi-
tion to have certain feelings, I do think that having beliefs includes having
such dispositions, at least according to ordinary practice.
We have seen that some evidentialists do not mind if their accounts
of belief deviate from ordinary practice. If we think about Velleman’s
discussion of what we would normally describe as falsely believing that
one is Napoleon, for example, we can see that he is not bothered by such
deviations. Such a person would have many of what we normally think of
as dispositions associated with belief. Given Velleman’s account of belief
as a cognitive state that is designed to capture truth, if a state is completely
insensitive to evidence, Velleman says it is not a belief. Why do we normally
describe someone who claims to be Napoleon as believing he is Napoleon?
Well, if we asked him if he believed it, he would say “yes” and it seems
he would feel about this proposition the way he would about others he
believed. According to Velleman, we are mistaken in this usage. But it seems
the only reason he provides for why this cannot legitimately be called a
belief is that his account of belief rules it out.
Now, if he and other evidentialists claimed that they were only interested
in a certain kind of belief, namely evidence-sensitive kinds, such restriction
would be acceptable. But none of them are overt about such a restriction,
and it is unlikely they would see their views as applying only to a particular
 Unity of Norms: A Defense of Pragmatism 67
category of believing.34 Further, most defenders of evidentialism do appeal,
at times, to ordinary usage in their accounts. For example, in defending his
very strong version of evidentialism, Adler asks the reader to think, from a
first-person perspective, if he would ever claim to have a belief without also
claiming to have adequate reasons (which usual means evidence) for that
belief. He often appeals to what we ordinarily do or say and even says that
when some terms remain unexplicated he is usually deferring to “ordinary
understanding.” Adler realizes that counterexamples will come at him,
particularly since he has been clear that he means to include beliefs with
all kinds of content: “empirical, mathematical, moral, emotional, intui-
tive.” In considering a supposed case of believing against evidence, he says
that he needs a way to “explain away such declarations,” and one way of
doing this is to point out that “ordinary usage of the term ‘belief’ and its
cognates . . . will be very loose.” So, it seems he appeals to ordinary usage
when it supports his view and dismisses it when it does not.
What I have said about norms for belief remains neutral between most
theories about the nature of belief. It does, however, rule out a traditional
dispositional view of belief, where what it means to believe something is sim-
ply that one is disposed to act in certain ways. It does not commit one, how-
ever, to the idea that one can only believe something at the moment one is
attending to it, that beliefs must only be occurrent. Any theory allowing that
having a belief includes having some kind of feeling about what is believed
when the belief is occurrent is compatible with my view. Eric Schwitzgebel’s
account that he calls “a phenomenal, dispositional account of belief” is one
that seems to match up well with what I deem a belief. In his view, “to
believe that P . . . is nothing more than to match to an appropriate degree
and in appropriate respects the dispositional stereotype for believing that P,”
and so it allows for intermediate cases where it is not entirely clear whether
one believes or not. The dispositional properties belonging to the belief ste-
reotype include behavioral, phenomenal, and cognitive dispositions. So, my
feeling of surprise upon discovering no beer in the fridge is one of the dispo-
sitions associated with the belief that there is beer in the fridge.35
Evidentialists privilege one of the many dispositions associated with
belief, namely that of “ceasing to have most belief-related dispositions upon
discovering good reasons for its falsity,” over all the others. They think that if
you don’t have this disposition, then you don’t have a belief. Why think this?
Some evidentialists have a functionalist view of belief and, if belief is a state
with the function of attaining truth, any state that is incapable of realizing
this function should not be called a belief. I have argued (in Chapter 1) that
such a view fails to take into account that when something (a tool, machine,
organ, organism) deviates (even dramatically) from its normal function, it
need not thereby stop being that thing. As we have seen, it is possible that
such malfunctions are adaptive.
Here is a definition of belief given by McKay and Dennett, which they
say is general enough to cover most representationalist and dispositional
68 Doxastic Norms
accounts: “A belief is a functional state of an organism that implements
or embodies that organism’s endorsement of a particular state of affairs as
actual.”36 Unlike Schwitzgebel’s, this definition does not require endorsement
to include a feeling. But it does not rule this out either, so it is compatible with
my view. And it is clear that in McKay and Dennett’s view, a “misbelief,”
that is, a state that endorses another state of affairs as actual when this state
does not obtain, is a belief nonetheless. This definition helps to explain why
the permissible practical beliefs are not really “hopes.” Belief seems best
understood as a mixed state, as having a cognitive/representational aspect
as well as a feeling-aspect. Hoping that something is true does not require
any kind of endorsement or commitment.

3. TRANSITION TO PART II: WHAT KIND OF

“OUGHT” IS THE DOXASTIC “OUGHT”?

As I mentioned earlier, some philosophers agree that non-evidentially based

beliefs can be prudentially or morally valuable, but still insist they are
incorrect from an epistemic perspective. I accept that for certain purposes it
can be useful to distinguish epistemic value from prudential or moral value.
But the question remains: What is the nature of the doxastic ought, namely
the ought in question when I ask, “What ought I to believe”? I have argued
that the source of normativity concerning beliefs is the same as that concern-
ing actions. When we say one ought to act a certain way and when we say
one ought to believe a certain way, these “oughts” are not distinct. It seems
there is an “ought” associated with all our activities as agents, whether these
result in beliefs or in actions. So it may well be that the actions performed by
the survivors of the plane wreck are praiseworthy, even though their beliefs
are not. The same thing can be said of the loyal friend. Her sympathy and
non-abandonment may be praiseworthy (or at least permissible), while her
beliefs may not be. What we are assessing is the competence (or excellence)
of agency. One may think that it is impossible to act as a loyal friend if
one does not dismiss the evidence pointing to one’s friend having behaved
maliciously. But it may just be very difficult—as Aristotle has told us, “it is
hard work to be excellent.” If one can remain loving and caring without
any need for ignoring evidence or self-deception, then that person would be
doing better overall than one who needed to choose which aspect of agency
(believer or friend) to favor.
One reason many philosophers resist assimilating the norms of belief to the
norms of agency is that they think there is no such thing as doxastic agency.
As Shah puts it, practical considerations can determine what is desirable to
believe and so speak to the question of what belief to bring about, but such
considerations have nothing to say concerning the question what to believe:
“Conflating the doxastic question whether to believe that p with the practi-
cal question whether to bring about belief that p has misled philosophers
 Unity of Norms: A Defense of Pragmatism 69
into thinking that evidential and practical considerations issue in competing
answers to the same question, and that therefore we must decide which has
priority.”37 According to Shah, only evidential considerations are possible
in deciding what to believe; the world determines what to believe, and so
there is no room for agency in the doxastic realm. This is similar to the point
that Williams made. We cannot decide what to believe in any way that is
analogous to how we can decide how to act, and so the norms guiding these
realms must be distinct.
Matthew Chrisman has argued that the kinds of “oughts” that are
applied to beliefs should be seen as akin to rules of criticism, as distinct
from rules of action.38 Rules of criticism apply to states of things, of how
things ought to be rather than to what one ought to do. Still, these rules of
criticism can be related in various ways to rules of action, in that they imply
that some agents ought to act in various ways so as to promote things being
as they ought to be. Chrisman thinks doxastic “oughts” tell us, in general,
truths about beliefs and believing. They would be of the form:

X ought to have doxastic attitude A toward proposition p under

conditions C.

This rule does not specify how to bring it about that one comes to have
the attitude one ought to have. The kind of rule of action that is implied by
a doxastic ought, Chrisman argues, may well apply to individuals beyond
the believer. For example, when we say, “One ought to disbelieve the earth
is flat,” this could well imply the “interpersonal” rule of action: “Parents
and teachers ought to teach young people that the earth is not flat.” Rules
of criticism are usually applied to states of things, not to states of agents.
However, Chrisman points out that “none of this implies that believers
cannot be agents. We just have to appreciate that they do not exercise agency
in believing what they believe.”39 This is a clear statement of the view I am
opposing; I think one of the central ways that we exercise agency is in believ-
ing what we believe. When I say that one “exercises agency” in believing or
acting, I mean we are not passive in regard to such beliefs or actions; they do
not just happen to us but we are active in making them happen. Chrisman
comes to this conclusion, largely, because we do not exercise the same kind
of voluntary control over beliefs as we do over many actions. But to say
that “exercising voluntary control” and “exercising agency” are equivalent
is a view that needs defending and, ultimately, we will see, implies that even
many of our actions for which we hold each other responsible are not those
in which agency is exercised.
The second part of this book explores various conceptions of doxastic
agency and responsibility. In this chapter, I have argued that we have some
control over what we believe and have, thus, opened up room for doxastic
agency. I will argue that we have much more control in the realm of belief than
many theorists allow; our expectations of ourselves and others as agents reveal
70 Doxastic Norms
that we do have a fairly robust kind of control over our beliefs. My conclusion
will reinforce the possibility of doxastic agency and so give further support to
the view that we should allow for a unity of norms guiding action and belief.

NOTES

1. Clifford, “The Ethics of Belief,” 24.

2. Bering, “The Existential Theory of Mind,” 3–24.
3. McKay and Dennett, “The Evolution of Misbelief,” 493–561.
4. Ibid., 36. The evidence for what is called “depressive realism” may also sup-
port my view. Lyn Abramson and Lauren Alloy first introduced this theory
in “Judgment of Contingency in Depressed and Nondepressed Students: Sad-
der but Wiser?” (1979). A series of experiments seemed to support the view
that depressed people better grasp the control they have over events than
do non-depressed people. It is further argued that non-depressed people’s
appraisals tend to be positively biased. Some further research has supported
the theory, whereas others have questioned the findings. For a fairly recent
critical review of the literature and one that ends by questioning the scope of
the theory, see Allan, Siegel, and Hannah, “The Sad Truth about Depressive
Realism,” 482–495. It could be that having beliefs that always properly cor-
respond to the evidence could have serious practical costs.
5. James, “The Will to Believe,” 64.
6. Ibid.
7. Ibid.
8. Bishop, “How a Modest Fideism May Constrain Theistic Commitments,”
388. See also Bishop, Believing by Faith.
9. See, for example, Daniel Whiting, “Reasons for Belief, Reasons for Action,
the Aim of Belief, and the Aim of Action.” I will return to this discussion
about the connection between practical reason and belief in Chapter 6.
10. In Part II, I will discuss in more detail the ways in which one can be active in
the doxastic realm.
11. When developing an argument for why we are justified in believing that a
belief system that retains long-term coherence would be one that also cor-
rectly describes reality, Lawrence Bonjour tries to develop an argument
that reveals that the external world hypothesis is a priori more likely than
any other hypothesis. When he first introduced this idea in The Structure
of Empirical Knowledge (181), he admits philosophers are likely “to have
qualms about” the idea of a priori probability. We are just supposed to “see”
that skeptical hypotheses are “antecedently less likely to be true than the
correspondence hypothesis.” Of course, this is how it seems given that we
believe one and not the other. I am not sure if other’s qualms have been
assuaged over the last thirty years, but I still cannot make sense of this idea.
12. Hume, An Enquiry Concerning Human Understanding, 119.
13. The revival of the Moorean response is due, in large part, to James Pryor’s
influential discussion first articulated in his “The Skeptic and the Dogma-
tist” (517–549). While Pryor’s view is sometimes construed as saying that our
perceptual experiences provide us with evidence for our perceptual beliefs,
this is not actually what he says. Instead, he says that for the dogmatist “your
experiences give you justification for believing p, but it would be misleading
to call these experiences your “evidence” for believing p . . . The dogmatist
thinks the mere having of an experience as of p is enough for your percep-
tual justification for believing p to be in place” (519). When Pryor defends
 Unity of Norms: A Defense of Pragmatism 71
and clarifies his own view, he is clear that this kind of justification is prima
facie justification. Even if perceptual experiences provide us with prima facie
justifications for corresponding beliefs about them, it is not clear that they
provide us with anything like justification for our belief in the external world.
Moore’s “proof of the external world” relies on my knowing that there is
some external object (e.g., that I have hands), which then entails that I know
the skeptical possibilities that would undermine my ordinary belief do not
hold. It, thus, seems to be an argument for showing we know they do not
hold, which somehow justifies our belief in the external world (see Moore,
“Proof of the External World”). Pryor’s point, however, is we do not need
to provide any argument to show that these possibilities do not hold. It isn’t
even clear that the dogmatism Pryor defends has anything to say about the
belief about the cause of our perceptions, which is, ultimately, what our belief
in the external world amounts to. Even if my experience of a tree provides
me with prima facie justification for my belief that there is a tree, it seems
that this justification can be defeated by the skeptical possibilities that would
seem to undermine it. The question is whether one is justified in taking it for
granted that our perceptual experiences can also justify the belief that, in
general, these experiences reliably connect to the world as it really is. Some
arguments for how dogmatism cannot provide us with this justification are
found in Matthew McGrath’s “Dogmatism, Underminers, and Skepticism”
forthcoming in Philosophical and Phenomenological Research and Roger
White, “Problems for Dogmatism,” 525–557.
14. Hume, A Treatise of Human Nature, 187.
15. Nozick, Philosophical Explorations, 220. Nozick’s main concern is to offer
a theory of knowledge that explains how one can know despite not knowing
the sceptical hypotheses don’t hold. Roughly, if my belief tracks the truth
of a proposition, then it is sensitive and I know it. If I have a true belief but
would still believe it if it were false, then that belief is not sensitive. Suppose
I believe that a particular bird is a nuthatch and it is, in fact, a nuthatch.
If I were to believe it is a nuthatch even if it is a chickadee, I do not know
it is a nuthatch. I have many beliefs that are insensitive and, therefore, do
not know a lot. But with many such beliefs, we can imagine that if I were
in a stronger epistemic position, these beliefs would become sensitive and
so I would acquire the knowledge I now lack. So, for example, I can imag-
ine increasing my ornithological credentials to the point where my nuthatch
belief would be sensitive. But with sceptical hypotheses, it seems, my epis-
temic position can never be strong enough for my belief that they are false
to be sensitive. Tracking accounts of knowledge are not widely endorsed at
present. Many think that it is fatal flaw for Nozick’s view that he denies a
kind of epistemic closure. It seems that if I do not know I am brain in a vat
and I know that being a brain in vat entails not having hands then I do not
know I have hands. I am attracted to a Nozickian kind of view but the details
of such a view and whether it can overcome objections are not relevant to
my discussion here. All I want highlight here is that a belief in the falsity of
a sceptical hypothesis is very different from other beliefs, in that it doesn’t
respond to evidence, cannot be abandoned, and cannot be strengthened.
16. Stroud, “Epistemic Partiality in Friendship,” 498–524. For arguments about
other times it might be rational to believe against the evidence, see Maru-
sic, “Belief and Difficult Action,” as well as his “Promising against the Evi-
dence,” especially pages 311–315.
17. McKay and Dennet, “The Evolution of Misbelief,” 506. Among the misbeliefs
that McKay and Dennett deem adaptive are viewing those one loves as possess-
ing more positive traits than others, in particular one’s children and one’s partner.
72 Doxastic Norms
18. In a recent paper, Yuval Avnur considers the issue of how to distinguish the
belief in the external world from belief in God. Although there are some
similarities in our approach, ultimately Avnur thinks we can only be excused
for our belief in the external world and this is because we are stuck with it,
and stuck with it in away in which we are not stuck with a belief in God. I am
not happy with this view, because it still makes it look like there is something
wrong with the belief. See Kvanvig, “In Defense of Secular Belief.”
19. Sellars, “Empiricism and the Philosophy of Mind,” 253–329 (originally pre-
sented at the University of London Special Lectures in Philosophy for 1956
as “The Myth of the Given: Three Lectures on Empiricism and the Philoso-
phy of Mind”).
20. Alvin Plantinga, “Reason and Belief in God,” in Plantinga and Wolterstorff,
Faith and Rationality, 75.
21. Ibid., 73
22. Ibid., 78
23. Ibid., 80.
24. Ibid., 82.
25. Ibid., 19.
26. Ibid., 20.
27. Wittgenstein, On Certainty, 359.
28. My view of doxastic agency, the subject of Chapter 6, actually shares many
of the features of Wolterstorff’s Reidian account.
29. Nicholas Wolterstorff, “Can Belief in God Be Rational If It Has No Founda-
tions?” in Plantinga and Wolterstorff, Faith and Rationality, 176.
30. Although Anscombe never actually uses this terminology, the distinction is
often attributed to her discussion in her Intention.
31. Mark Platts, Ways of Meaning, 257.
32. For an illuminating discussion on what kind of commitment is involved in
beliefs, see John Gibbons, The Norm of Belief, 224–230.
33. Jonathan Cohen, An Essay on Belief and Acceptance, 11.
34. It can be argued that some of the examples I have given of beliefs that persist
without evidence would be better classified as what Tamar Gendler has
termed “aliefs.” See her “Alief and Belief,” 634–663 and “Alief in Action
(and Reaction),” 552–585. Her way of explaining the fear-behavior people
manifiest when they walk on a glass skywalk overlooking the Grand Canyon
when they claim to believe it is safe is as follows: “Although the venture-
some souls wholeheartedly believe that the walkway is completely safe, they
also alieve something very different” (635); the subject’s hesitation to walk
out over a skywalk can be explained by a “belief-discordant alief” (641).
It is interesting that one of her motivations in developing a conception of a
new kind of mental state is that, without it, it seems some of the ways phi-
losophers tend to think about belief is put into question. She says “a notion
of alief is crucial if we wish to hold on to a notion like belief that relates to
action in anything like the way philosophers have traditionally assumed.”
(“Alief and Belief,” 647). These “traditional assumptions” about belief are
all evidentialist assumptions. They include that they are “reality sensitive”
or “sensitive to evidence,” that “actions generated by beliefs are generated
by a mental state that is proportioned to all-thing-considered evidence and
subject to rational and normative revision” (“Alief in Action”, 570), that
“Belief aims to ‘track truth’ in the sense that belief is subject to immediate
revision in the face of changes in our all-things-considered evidence” (“Alief
in Action,” 565). If we accept these assumptions, then we need a way of
explaining people’s behavior in examples like the skywalk case. As Gendler
puts it: “Any theory that helps itself to notions like belief, desire and pre-
tense needs to include a notion like alief in order to make proper sense of a
 Unity of Norms: A Defense of Pragmatism 73
wide range of otherwise perplexing phenomena [. . . in short, if] you want to
save belief, then you need to make conceptual room for the notion of alief”
(“Alief and Belief,” 641–642). Another way one can “save belief” is to give
up on the view that belief and evidence are always so tightly connected.
35. Eric Schwitzgebel, “A Phenomenal, Dispositional Account of Belief,”
249–275.
36. McKay and Dennett, 493.
37. Shah, “A New Argument for Evidentialism,” 498.
38. Chrisman, “Ought to Believe,” 346–370.
39. Ibid., 369.

REFERENCES

Allan, G. Lorraine, Shepard Siegel, and Samuel Hannah. “The Sad Truth about
Depressive Realism.” The Quarterly Journal of Experimental Psychology 60,
no. 3 (2007): 482–495.
Alloy, L. B., and L. Y. Abramson. “Judgment of Contingency in Depressed and
Nondepressed Students: Sadder but Wiser?” Journal of Experimental Psychology:
General 108 (1979): 441–485.
Anscombe, G.E.M. Intention. Oxford: Basil Blackwell, 1957.
Avnur, Yuval. “In Defense of Secular Belief.” Oxford Studies in Philosophy of Religion
Vol. IV, edited by Jonathan Kvanvig, 2–20. Oxford: Clarendon Press, 2011.
Bering, Jesse. “The Existential Theory of Mind.” Review of General Psychology 6,
no. 1 (2002): 3–24.
Bishop, John. Believing by Faith. Oxford: Oxford University Press, 2007.
———. “How a Modest Fideism May Constrain Theistic Commitments: Exploring
an Alternative to Classical Theism.” Philosophia 35 (2007): 387–402.
BonJour, Laurence. The Structure of Empirical Knowledge. Cambridge: Harvard
University Press, 1985.
Chrisman, Matthew. “Ought to Believe.” The Journal of Philosophy 105, no. 7
(2008): 346–370.
Clifford, William K. “The Ethics of Belief.” In The Ethics of Belief Debate, edited by
Gerald D. McCarthy, 19–36. Atlanta: Scholars Press, 1987.
Cohen, Jonathan L. An Essay on Belief and Acceptance. Oxford: Oxford University
Press, 1992.
Gendler, Tamar. “Alief and Belief.” The Journal of Philosophy 105, no. 10 (2008):
634–663.
———. “Alief in Action (and Reaction).” Mind and Language 23 (2008): 552–585.
Gibbons, John. The Norm of Belief. Oxford: Oxford University Press, 2013.
Hume, David. An Enquiry Concerning Human Understanding. Edited by Tom
L. Beauchamp. Oxford: Oxford University Press, 1999.
———. A Treatise of Human Nature, ed. L. A. Selby-Bigge, 2nd edition, revised by
P. H. Nidditch Oxford: Clarendon Press, 1978)
James, William. “The Will to Believe.” In The Ethics of Belief Debate, edited by
Gerald D. McCarthy, 55–71. Atlanta: Scholars Press, 1987.
Kvanvig, Jonathan L. ed. “In Defense of Secular Belief.” In Oxford Studies in
Philosophy of Religion 4. Oxford: Clarendon Press, 2011.
Marusic, Berislav. “Belief and Difficult Action.” Philosopher’s Imprint 12, no. 18
(2012): 1–30.
———. “Promising against the Evidence.” Ethics 123, no. 2 (2013): 292–317.
McGrath, Matthew. “Dogmatism, Underminers, and Skepticism” (forthcoming). In
Philosophical and Phenomenological Research.
74 Doxastic Norms
McKay, Ryan T., and Daneil C. Dennet. “The Evolution of Misbelief.” Behavioral
and Brain Sciences 32 (2009): 493–561.
Moore, G. E. “Proof of the External World.” In Philosophical Papers. New York:
Collier Books, 1939.
Nozick, Robert. Philosophical Explorations. Harvard: Belknap Press, 1981.
Alvin Plantinga, “Reason and Belief in God,” in Faith and Rationality, edited by
Plantinga and Wolterstorff, 75. Notre Dame: University of Notre Dame Press,
1983.
Plantinga, Alvin, and Nicholas Wolterstorff. Faith and Rationality: Reason and
Belief in God. Notre Dame: University of Notre Dame Press, 1983.
Platts, Mark. Ways of Meaning. London: Routledge and Kegan Paul, 1979.
Pryor, James. “The Skeptic and the Dogmatist.” Nous 34, no. 4 (2000): 517–549.
Schwitzgebel, Eric. “A Phenomenal, Dispositional Account of Belief.” Nous 36
(2002): 249–275.
Sellars, Wilfred. “Empiricism and the Philosophy of Mind.” In Minnesota Studies
in the Philosophy of Science, edited by H. Feigl and M. Scriven, 253–329.
Minneapolis: University of Minnesota Press (1956).
Shah, Nishi. “A New Argument for Evidentialism.” The Philosophical Quarterly
56, no. 225 (2006): 481–498.
Stroud, Sarah. “Epistemic Partiality in Friendship.” Ethics 116, no. 3 (2006):
498–524.
White, Roger. “Problems for Dogmatism.” Philosophical Studies 131 (2006):
525–557.
Whiting, Daniel. “Reasons for Belief, Reasons for Action, the Aim of Belief, and
the Aim of Action.” In Epistemic Norms, edited by Clayton Littlejohn and John
Turri. Oxford: Oxford University Press, 2014.
Wittgenstein, Ludwig. On Certainty. Oxford: Blackwell, 1669.
Wolterstorff, Nicholas. “Can Belief in God Be Rational If It Has No Foundations.”
In Faith and Rationality: Reason and Belief in God edited by Alvin Plantinga and
Nicholas Wolterstorff, 135–186. Notre Dame: University of Notre Dame Press,
1983.
Part II

Doxastic Responsibility
This page intentionally left blank
4 The Puzzle of Doxastic
Responsibility

1. THE PUZZLE

We have been concerned with the question of how ought we to believe and
I have been urging that one way of answering that question is misguided.
I have argued, contrary to those who I have termed “evidentialists,” that
it can, at times, be possible and permissible to believe without evidence.
It seems that the practice of making normative assessments of belief,
and that we can even argue about which beliefs are permissible, reveals
that we hold people responsible for the beliefs they hold. This notion of
responsibility is not simply one pointing out the causal genesis of the belief.
Holding someone responsible for his beliefs is not like holding the wind
responsible for knocking over the tent. That we praise and blame each other
for the beliefs we hold indicates, rather, that, at least sometimes, we view the
beliefs that one forms and maintains as resulting from one’s agency.1
This whole discussion of an “ethics of belief” may be misguided, however,
if beliefs are largely involuntary. Attributions of responsibility and other
deontological judgments in the doxastic realm are puzzling. For, much of
what we believe is beyond our control; it seems we cannot decide to believe
the way we can decide to act. But such lack of control should exempt us
from responsibility and judgment. How can we sensibly talk of how we
ought to believe if we cannot exercise control over what we believe? We can
formulate our puzzle by considering the following argument.

The Voluntarism Argument

1. If attributions of responsibility for beliefs are appropriate, then people
have voluntary control over their beliefs.
2. People do not have voluntary control over their beliefs.
Conclusion: Attributions of responsibility about beliefs are not
appropriate.2
Yet, our practices seem to assume such attributions are appropriate.
We would disapprove of someone who believes that whales are fishes, or
that her neighbor littered the sidewalk when she did not. We think a typically
78 Doxastic Responsibilities
well-informed American ought to believe the Earth revolves around the sun
and would be critical of someone who believes the sun revolves around the
Earth.3 And yet it seems quite clear that one with such a belief could not just
decide to change it in the direct way he could simply decide to change his
shirt. Neither can one decide to acquire the belief the way one could decide
to acquire a new pair of shoes.
Three responses to this puzzle are possible. The first response denies
the second premise, arguing that, at times, we can effectively decide to
believe; this view has come to be called “doxastic voluntarism.”4 A second
response to the puzzle accepts the argument as sound and so denies that
we are responsible for beliefs, and argues that our common practices of
attributions of responsibility are misguided; we are mistaken if we think,
for example, someone should be praised or blamed for a belief he holds.
Just as I cannot help feeling wet when rain falls on me or hot when the
sun is strong, my believing that it is raining or the sun is strong is not
something that is “up to me;” it is a state I find myself in when the world
impinges on me in certain ways. On such a view, belief is a passive phe-
nomenon and must be to do its job. Such a revisionist theory that claims
common practices are erroneous should only be accepted if there is no
way of resolving the tension without plunging us into such widespread
error.
The third, and currently most common, response to the puzzle argues that
although we lack voluntary control over our beliefs, we can nonetheless be
held responsible for them, thus denying the first premise of the Voluntarism
Argument. In Chapter 5, I will consider a number of ways one might
account for our normative judgments about beliefs while maintaining we
lack voluntary doxastic control. In this chapter. I will discuss arguments for
doxastic voluntarism and arguments for the view that we actually are not
responsible for beliefs.

2. DOXASTIC VOLUNTARISM

In his fourth meditation, Descartes articulates a robust form of voluntarism.

He says that our will is completely free to affirm or deny what is presented
to the intellect: “The will consists solely in the fact that when something is
proposed to us by our intellect either to affirm or deny, to pursue or to shun,
we are moved in such a way that we sense that we are determined to it by
no external force.”5 For Descartes, how I act or how I believe is completely
up to me. I have freedom to choose both my beliefs and my actions; the per-
son who assents to a proposition has made a choice. This freedom explains
both doxastic and moral error; it makes as much sense to blame me for my
beliefs as it does to blame my actions, which come from the same faculty.
So, I can consider the proposition, “The sun revolves around the Earth,”
and then decide whether to assent to it or not. According to Descartes,
I ought to restrain my will and only assent to those propositions that are
 The Puzzle of Doxastic Responsibility 79
“clearly and distinctly represented to it [the will] by the understanding,”
I am, thus, clearly responsible for not restraining my will when I let myself
believe those propositions that I “conceive more confusedly and obscurely.”
If all beliefs were directly under the control of the will, then it would seem
that my decision to believe something (whatever the reason for this decision)
would result in my believing it. I can raise my arm whenever I decide to
(absent external force keeping it down), whatever my reason for willing my
arm raised; in such a situation, I effectively decide to raise my arm because
I succeed in executing my decision. Sometimes, we decide to do things that
we do not end up doing; something might get in the way of my decision being
effective. Most opponents of doxastic voluntarism argue that it is impossible
to decide to believe. What they really mean is that it is impossible for a deci-
sion to believe to be effective. The question here is whether effective decisions
are possible in the doxastic realm. I cannot effectively decide to believe that
I am, for example, six feet tall, or that Caesar died in his bed, or that there
is an odd number of stars. Any contemporary voluntarist will accept that we
cannot control all our beliefs, in the same way that we can raise our hands or
imagine the Eiffel tower. But the reasonable voluntarist will ask, is it the case
that deciding to believe is sometimes possible? Voluntarists will say it is and
argue that there are times when belief is voluntary. Carl Ginet argues that
this can happen when the evidence it not conclusive. The position he argues
for is “that coming to believe something just by deciding to is possible, that it
sometimes seems to us that we do this, and that our doing so need not offend
against epistemic reason.”6 Here is another example of a case he considers
to be a good candidate for being described as someone’s deciding to believe
something. In such cases, he argues, the subjects believe a proposition and the
belief “came into existence directly by the subject’s deciding that it would.”7

Before Sam left for the office this morning, Sue asked him to bring from
his office, when he comes back, a particular book she needs to use in
preparing for her lecture the next day. Later, Sue wonders whether Sam
will remember to bring the book. She recalls that he has sometimes, though
not often, forgotten such things, but, given the inconvenience of getting in
touch with him and interrupting his work and the thought that continu-
ing to wonder whether he’ll remember it will make her anxious all day,
she decides to stop fretting and believe that he will remember to bring it.8

Ginet argues that in cases like this we decide not to consider the possibility
of not-p; we count on p’s being true. Ginet is careful to distinguish count-
ing on p from staking something on p. “To stake something on p” means
to hope that it turns out to be the case, perhaps even to think that it is more
likely to be the case than not. If Sue has merely staked her hopes on Sam
remembering her book, she has not closed off the possibility that he will
forget it. Unless she does close off this possibility, she will be anxious and
fretful. So, in deciding to cease the fretting, according to Ginet, she must
decide to do more than stake something on Sam’s remembering the book;
80 Doxastic Responsibilities
she must count on it, and this kind of counting, Ginet argues, is believing.
Thus, if we can decide to count on something being the case, we can decide
to believe it. In each of the cases Ginet considers, the decision to believe is
coupled with another decision. As Ginet puts it: “In these examples, the
subject S decided to believe a certain proposition p. S did this in deciding to
act, or not act, in a certain way.”9 In the case cited above, Sue decides not
to remind Sam to bring the book, and in so deciding, she decided to believe
that he would remember to bring it. In the case of a poker game in deciding
to bet, the bettor decides to believe her opponent is bluffing. In a courtroom
case, in deciding to vote for acquittal, Sam decided to believe the statement
of the witness.10 Ginet says that in each of these examples the agents could
have acted as they did without also deciding to believe as they did. In such
cases, they would all be more prepared to deal with the possibility of not-p.
For example, Sue could have decided to not remind Sam but continue to fret
and to think about what to do were he to forget it.
Adler argues that it is a mistake to see the ultimate epistemic position
regarding the inconclusively supported proposition as “full belief.” He
admits that there may be times when I do not yet have a belief and then I
“make up my mind” and judge something true (thus, issuing in full belief)
even when no new evidence has been added. What we decide to do is to end
inquiry and, according to Adler, belief occurs when inquiry ends; so what
I am doing, he says, is “not deciding to believe; rather, I am deciding to
place myself in a position where I will come to believe.”11 Belief, therefore,
remains passive and nonvoluntary.
I agree with Adler that viewing belief as a direct result of a decision seems
misguided. It is significant that in all of Ginet’s cases there are doubts that
require active suppressing. In each case, the subject’s state of mind becomes
more tranquil once the belief is acquired. But when we assess each case, we
end up wondering whether the agent has closed off the possibility of not-p
to the extent required for Ginet’s account. When I decide to continue down
the road and, in deciding to continue, decide to believe that I locked the door
(that is to count on it being the case), can I really forget my earlier anxiety
so that the possibility is gone from my mind? Contrast the situation with
an unmotivated, ordinary perceptual belief. I believe there is a computer in
front of me and my having this belief requires that I refuse to take seriously
the possibility that I am a brain in a vat or am being deceived by an evil
genius, but this refusal requires no active suppression or self-deception.
An uncontroversial example of deciding to believe should not involve self-
deception, suppressing of memories, wishful thinking, etc. If I act in a way
that exhibits these rational defects, we would wonder whether it resulted
from a decision at all. My decisions to act often come about as a result of
deliberation, in which I reflect on what is best for me to do. If, swept away
by momentary passion, I refuse to even consider the consequences of infidel-
ity, concentrating my attention only on present pleasure, it is questionable
whether my action results from a decision.
 The Puzzle of Doxastic Responsibility 81
Perhaps my construal of what is required for an act to result from a deci-
sion is overly narrow. While saying someone decided to do something often
suggests that they were deliberating between more than one option, it
could be that such deliberation is not required. We do, for example refer to
“split-second decision.” It could be that, at times, we contrast actions that
result from decisions to those that are nonvoluntary, like reflexes or behavior
that has its cause wholly outside the agent. While there might be a sense
in which decision could have this wider scope, the more narrow meaning
seems to be the more common and central one. “Decision theory” is taken
to be synonymous with “rational choice” theory. Even if actual agents devi-
ate from what decision theory says is the rational choice, there is still the
idea that when making decisions, people are choosing among options, even if
their deliberation can be faulty. In the context of medical ethics, in trying to
evaluate where a patient is capable of consent, one of the questions asked is
whether the patient has “decisional capacity.” Whether one has this capacity
depends, in part, on whether one has the capacity for reasoning: “Without the
mental ability to engage in reasoning and manipulate information rationally,
it is impossible for understanding and appreciation to issue in a decision.”12
I have been reflecting on what it means to make a decision to help assess
whether Ginet’s examples are examples of decisions to believe. Because the
idea of rational deliberation seems to be so closely connected to forming a
decision, these cases that seem to require a kind of sabotage or ignoring of
the deliberative process do not seem like decisions, at least as we usually
think of them. And it seems Ginet views decisions as coming at the end of
process of deliberation. In each case, the subject is considering evidence that
both supports and undermines a particular proposition and then acts in
such a way that, according to Ginet, entails counting on (and, so, believing)
the proposition. What would the subjects in these cases state as their reasons
for believing? It seems that there are two answers that Ginet can give to this
question and each will be problematic for his thesis that “coming to believe
something just by deciding to is possible . . . and that our doing so need not
offend against epistemic reason.”13 The first answer points to some kind of
non-epistemic reason, such as the need for peace of mind. This need figures
prominently in two of the cases—the one concerning Sue and the book, as
well as the one about believing I locked the door when on a road trip. I have
some evidence that I did (an unclear, not-fully confident memory sensation)
and I have some evidence that I did not (that such unclear, unconfident
memory impressions have, at times, been mistaken). But, “given the great
inconvenience of turning back and the undesirability of worrying about it
while continuing on, I decide to continue on and believe that I did lock it.”14
Did I decide to believe it because doing so was desirable? Shah’s discussion
of the difference between first-person and third-person attribution of reasons
(and causes) for belief is important here. From a third-person perspective,
one could say that the desirability of believing, that it would be inconvenient
to not believe it, was part of what caused me to have the belief. But Shah
82 Doxastic Responsibilities
seems right that such a reason could not figure in a straightforward way as
a premise in doxastic deliberation. It seems that I would not say to myself:
Not believing I locked the door would be very inconvenient, so therefore it
is the case that I locked the door. Even though I think Shah is wrong that
the concept of belief makes such an inference impossible, it would clearly
be deeply irrational.
Given the way Ginet describes the case, the practical reasons serve as
my reasons for acting in a certain way, but in so acting I come to view
the evidence as supporting the belief. It seems likely that if asked what my
reason was for believing that I locked the door, I would say that I remember
doing so; that is, I would cite the evidence in support of the proposition. But
if this is so, then it seems Adler’s characterization of these cases is correct.
My deliberation results in a decision to act and then, once I so act, the
evidence dictates what I believe. The doxastic control is, thus, of a more
indirect kind that Ginet claims it to be.15
I do think these cases point to some kind of doxastic control, but Ginet
has tied the kind of control too closely to a particular kind of standard
voluntary action. Once Adler dismisses the possibility of deciding to believe,
he also dismisses the possibility of doxastic control. He says that only those
actions that result from decisions are voluntary, and then equates an act
being voluntary in this sense with having control over it. It seems there are
ways I can exhibit control over my actions even if they do not issue directly
from my decisions. If I decide to take a walk, it seems that I am in control
of each step, although I do not decide to take each step before taking it.
Perhaps, beliefs are under my control even if they do not result directly from
my decisions, in the sense argued for by some doxastic voluntarists. I will
argue that it is possible for us to take a much more active role in the forma-
tion of beliefs than those opposed to doxastic voluntarism have allowed.
I will turn to my account after considering the other two ways of responding
to the puzzle of doxastic responsibility.

3. NO DOXASTIC RESPONSIBILITY

The second response to our puzzle accepts that, given the involuntary
nature of belief, we are, in fact, not responsible for them and deontologicial
judgments are not appropriate in the doxatic realm. David Hume is often
invoked as the exemplar of such a view. For example, David Owens char-
acterizes Hume’s position in the following way: “In denying the existence
of epistemic agency, doxastic responsibility and intellectual freedom, Hume
means to reject the idea that belief is subject to reason. He allows that beliefs
are governed by the sort of biological norms that apply to the process of
breathing, or the workings of the human heart but no one thinks us respon-
sible for non-compliance with such norms.”16 While I ultimately think this
is a mistaken characterization of Hume’s view, some of what he says about
 The Puzzle of Doxastic Responsibility 83
belief supports this passive reading. Hume’s explanation of the difference
between a proposition believed and one not believed is entirely mechanistic.
For Hume, the difference between incredulity and belief lies in the manner
of conception of an idea. The ideas I believe feel different from those I do
not believe; in Hume’s terms, they have more “force and vivacity.” When
I experience a type of impression that has constantly been conjoined with
another, a habit is formed in my mind such that when I have an impression
of one, its vivacity is transmitted to the associated idea. For example, I hear
a friend’s voice in the next room. This present impression (the sound of the
voice) has always been conjoined with this person, and so my idea of that
person is so enlivened by the impression of this effect that it approaches
the liveliness of that impression. And it is this manner of conceiving of that
person in the next room that we call believing that the person is in the next
room. A belief feels very much like an impression, impressions being “our
more lively perceptions, when we hear, or see, or feel, or love, or hate, or
desire, or will.”17
We can see why such an account of belief seems to leave no room for
the notion of responsibility in belief formation. When I am faced with
certain impressions, I cannot help but to believe certain related ideas; it
seems I have no choice in what I believe. This inevitability seems to exempt
me from blame; I cannot be criticized for having the beliefs I do. Given
Hume’s mechanistic story about the enlivening (and diminution) of beliefs,
the answer to why I believe x must be: I believe x because of an inexplicable
habit of the mind to associate present impressions with their related ideas.
Questions about whether I ought to believe x seem out of place.
Hume, however, is clearly committed to the idea that an ethics of beliefs
is possible. Many passages in Hume’s works reveal that it is evident that he
does blame people for having or failing to have certain beliefs. For example,
he says that beliefs that are a result of prejudice, namely beliefs formed on
the basis of “general rules contrary to present observation and experience”
are “errors” that I ought to take care to avoid.18 I can similarly be criticized
if I believe x rather than y simply because x occurred more recently and is
thus conceived by my mind in a more lively manner. The man who trembles
when looking at the precipice below him, despite the complete security
afforded by the iron cage he is in, ought not to believe he is in danger.
Hume says we “ought to regulate our judgment concerning causes and
effect” with rules that are formed in the understanding. These rules teach
us to “distinguish accidental circumstances from the effacacious causes.”19
And of course, Hume famously says in the first Enquiry, “a wise man . . .
proportions his belief to the evidence.”20 Hume scholars have extensively
discussed how such normative assessments are consistent with Hume’s view
of belief as resulting from certain natural mechanisms of the mind.21
When one looks more deeply into Hume’s views about beliefs, it becomes
clear that Hume is not committed to the view that we are utterly passive in
our belief formation. Even if an essential component of belief, for Hume,
84 Doxastic Responsibilities
involves a nonrational feeling, one can make choices regarding the data
one attends to, and the processing of it one does, before this belief-feeling
occurs. Hume has the resources to show how we can form higher-order
beliefs about our beliefs and that we can revise our beliefs according to
these higher-order beliefs. Further, we can be blamed for failing to revise
them accordingly. To see how belief revision is possible for Hume, consider
his discussion of the effect of general rules on what we believe, found in
Treatise 1.3.13. Hume wonders how it can be that even when faced with the
lively impression of a witty Irishman, one can still maintain the belief that an
“Irishman cannot have wit.” His answer is that the same principle that gives
rise to our conclusions of causal reasoning produces our prejudiced beliefs,
namely custom and habit. When faced with many instances of Irishmen
conjoined with lack of wit, we conclude that there is a necessary connection
between the two. But when we form beliefs of this kind, even against the evi-
dence, we are mistaking “superfluous circumstances” with “the essential”22
But we can use the more “extensive and constant” general rules that guide
our causal judgments to “correct” the “propensity” to make such mistakes.
And if we do not do so, it seems we can be blamed for this failure. This
discussion shows that Owens was mistaken in characterizing Hume as deny-
ing epistemic agency and doxastic responsibility. Though he disagrees with
Descartes that beliefs are the direct result of the will, many practices related
to forming and maintaining our beliefs are under our control and are the
appropriate target of reactive attitudes.
My reason for taking a little time to discuss Hume’s views on belief is just
to show that accepting the second premise of The Voluntarism Argument—
that people do not have voluntary control over their beliefs—does not lead
directly to the conclusion that there is no agency and responsibility in the
doxastic realm. For Hume, it seems the kind of indirect control that we have
over beliefs is enough for him to allow such responsibility. Like Hume, many
theorists deny that we have any kind of meaningful control over belief, but
most do not want this lack of control to commit us to giving up on the
notion of doxastic responsibility and so they will try to explain how our
attributions of praise or blame are compatible with our lack of voluntary
control. For the implications that follow, if we do lack any responsibility
for beliefs is problematic. For example, Nell Levy, a recent proponent of
this “no responsibility” view,23 considers what we would want to say about
Dr. Fritz Klein, a concentration camp doctor who justified his practices by
saying, “I am a doctor and I want to preserve life. And out of respect for
human life, I would remove a gangrenous appendix from a diseased body.
The Jew is the gangrenous appendix in the body of mankind.” It seems
apart from the horrific actions he undertook, we also would blame him for
the beliefs he held.
Levy canvasses a number of arguments aiming to show that doxastic
responsibility is possible despite our inability to acquire or eliminate them at
 The Puzzle of Doxastic Responsibility 85
will. These include accounts that argue that doxastic responsibility does not
require control, as well as those that claim we do have the right kind of control
over belief to ground attributions of responsibility. He finds flaws in all the
accounts and concludes that “our lack of control over belief typically excuses
responsibility for them.”24 The only exception he makes is in cases where one
fails to gather more evidence when one ought to do so. But, he asks: when is
one in such a position? Levy adopts the position (argued for extensively by
Adler and discussed in Chapter One) that if I take myself to believe something,
then I must take myself to have sufficient evidence for it. But if I take a propo-
sition to be true, then why would I investigate it any further?

Sincerely and wholeheartedly to assert that p is to commit myself to

saying that the evidence I possess, including my higher-order evidence
(i.e., my evidence which bears upon the availability and the evidential
valence of evidence I do not currently possess), supports the conclusion
that p. Since this is the case, I take myself to be under no obligation,
moral or epistemic, to gather further evidence. Why gather evidence for
a proposition I know to be true?25

The only time such an obligation exists, according to Levy, is when one is
not fully confident in the belief, when doubts about its truth remain. And even
in such a case, Levy says two other conditions must hold for me to be indi-
rectly responsible for a belief: 1) I must believe that more evidence gathering
will put me in a better position to properly judge that what I believe is true,
and 2) I must believe that the costs of this evidence gathering are outweighed
by the potential value of being more confident in my belief. This last condi-
tion is more likely to hold when the content of the belief in question concerns
something of considerable importance.
If someone meets all these conditions and fails to gather more evidence, that
person could be blamed for holding the belief. It seems Clifford’s shipowner
may satisfy these conditions. He was not fully confident, he knew further
evidence would likely be helpful, and it was a matter of great importance.
But Levy thinks the conditions are rarely satisfied, because they are so strin-
gent, and “the comparative rarity of doxastic responsibility gives us little
reason to expect that we shall often be able to find a way to hold agents
responsible for their actions which derives from their responsibility for their
beliefs.”26 He thinks is very unlikely that they will hold in Klein’s case, for
Klein was probably entirely convinced of the truth, thus not thinking that
further evidence gathering would be needed. He says that people who hold
repellent beliefs like those of Klein “are more often morally insane than mor-
ally responsible.”
It seems we would often like to say of people when they hold beliefs with
complete certainty that they ought not to do so. If they have come to faulty
conclusions based on scanty evidence, and so have ended inquiry to their
86 Doxastic Responsibilities
satisfaction, they are blameworthy for getting themselves into that state.
Levy insists that holding one responsible in such a case is not fair because,
when one is in a state of full belief, one is compelled by the evidence; one
cannot do anything but believe when one views the evidence in a certain
light. And it is this passive, involuntary nature of belief that undermines the
legitimacy of attributions of responsibility. But, thinking again of Hume’s
view can help show why this ultimate passivity of belief does not excuse
us from responsibility. Belief is the output of certain input (memories, per-
ception, reasoning) conjoined with certain mechanisms of the mind. Now,
even if we cannot control the mechanism, that is we cannot control that
given certain input belief will come, we do have a lot of control over the
input. We can vary the context in which beliefs arise and, at times, we can
be held responsible for failing to vary the context. One could, for example,
when questioned about a belief that one holds with full confidence, decide
to attend to the question, thus adding input. And depending on the circum-
stances, one can be blamed for failing to add and attend to it.
What Levy (and others) may say about such cases is that we are actually
responsible and blameworthy for actions that led to the belief but not for
the belief itself. When I hold you responsible for robbing the bank, I am
holding you responsible for that action, not the earlier actions that led to
the robbing. I may well hold you responsible for those as well, but if you
didn’t have control over robbing the bank (say you chickened out at the end
and your co-conspirator held a gun to your head, or took over your brain
somehow), then I would not hold you responsible for robbing the bank,
even if you were responsible for the earlier actions that led up to bank rob-
bing. And so, one may argue, if the ultimate state of believing is not “up to
me,” I cannot be blamed for this state even if I can be blamed for actions
that led me to that state.
Arguments against doxastic control and responsibility often turn on point-
ing to asymmetries between actions and beliefs.27 It is true that beliefs are not
simply actions, but there are many ways in which they are like actions, and
at times is can be useful to attend to their action-like aspect. Even if beliefs
are not under our direct voluntary control, it seems that some of our actions
are like this as well; they are not a direct result of our decisions but still under
our control. We are told that one cannot believe “just like that” but that
we can act “just like that.”28 And if we ask, “Just like what?” a particular
action will be given as an example, such as raising one’s hand. It seems that
there is a kind of directness in acting that is not found in believing. As soon
as I intend to act, I do so, whereas a belief cannot be the direct result of an
intention. This is true, but it seems this kind of argument always privileges
a certain kind of action. There are many kinds of actions for which it seems
we are responsible that are not the direct result of intention. In arguing for
doxastic freedom, Matthias Steup considers a number of such actions that
are considered free but are not done because one has formed an intention
to do so. He says these include those performed because of habit, and those
 The Puzzle of Doxastic Responsibility 87
performed as automatic responses. He offers a list of examples belonging to
the latter category. “Intending to drive from my house to campus:

• I insert the ignition key,

• engage the clutch,
• shift into reverse, and
• step on the gas.”

Given most accounts of compatibilist freedom, Steup argues, these would

count as free.29 This discussion is part of Steup’s larger project to show that
no good argument can show that beliefs differ from actions in such a way
that renders us capable of being responsible for actions but not for beliefs.
Either an account is too stringent, saying that a criterion is needed that
many actions do not have (such as having intention as their direct cause), or
the account claims that beliefs fail to have a quality that Steup argues they
do have. For example, it is claimed that beliefs are not reasons-responsive,
but Steup says they clearly are and that when it is presumed they are not,
there is a “chauvinistic” assumption that freedom requires responsiveness
to practical reasons. Steup argues instead that “responsiveness to practical
reason and responsiveness to epistemic reasons equally ground freedom.”30
The accounts I discuss in Chapter 5 share this view. The view I provide of
doxastic control will build on some parts of Steup’s discussion, but Steup is
not centrally concerned in saying what it means to be responsible for beliefs;
he is more concerned with showing that the standard arguments showing
we are not responsible fail.
Ginet exaggerates the similarities between beliefs and voluntary actions
in his account of doxastic voluntarism. Levy exaggerates their differences.
In the next two chapters, instead of focusing on how alike or unalike beliefs
and actions are, I will turn, instead to trying to articulate what it takes to be
responsible for a state and then see how beliefs can be the kind of state for
which one can be responsible.

NOTES

1. I do not take it that one is only responsible if blame is appropriate. I follow

Fischer and Ravizza (1998)(and they follow Peter Strawson, 1974) as seeing
an agent as responsible if he is an apt candidate for a reactive attitude. Reactive
attitudes include gratitude, indignation, resentment, love, and respect. They
need not always coincide with judgments of blame-worthiness, but they often
will. In Conor McHugh’s discussion of doxastic agency, “Exercising Doxastic
Freedom,” he uses “reproach” rather than “blame” to describe our negative
attitude toward beliefs considered impermissible, because he considers “blame”
to have particularly moral connotation. I will use both fairly interchangeably.
2. Here, I follow the presentation of the argument by Richard Feldman, “The Eth-
ics of Belief,” except that I replace his “deontological judgments about beliefs
are true” with “attributions of responsibility for beliefs are appropriate.”
88 Doxastic Responsibilities
Feldman solves the puzzle by denying the first premise. This is Feldman’s char-
acterization of William Alston’s argument in “Epistemic Justification.” Many
discussions about issues of doxastic control and responsibility take Alston’s
argument as their starting point and then either defend Alston’s position or
argue against it. Alston is not, however, directly concerned with responsibil-
ity. Rather, he is primarily interested in whether a belief’s justification should
be thought of in terms that are analogous to how actions are assessed as
permissible or impermissible in ethics. I will not be discussing how best to
understand justification. But Altson’s argument seems to cast doubt on all
our talk about belief that assumes responsibility. Feldman and Levy (Doxas-
tic Responsibility) understand his argument to have this wide-reaching scope.
3. These are examples of blameworthy beliefs that Adler (Belief’s Own Ethics)
discusses (64).
4. My discussion mainly focuses on the positive ability to decide to believe. But
the strongest form of doxastic voluntarism would also allow one to withhold
belief.
5. Descartes, Meditations on First Philosophy, 38–39.
6. Ginet, “Deciding to Believe,” 63.
7. Ibid.
8. Ibid., 64.
9. Ibid., 65.
10. This last case was quoted in full in Chapter 1.
11. Adler, Belief’s Own Ethics, 61.
12. Stanford Encyclopedia of Philosophy, “Decision Making Capacity.”
13. Ginet, “Deciding to Believe,” 63.
14. Ibid., 64.
15. Bratman, “Practical Reasoning and Acceptance in a Context,” 1–13. It can
be argued that the cognitive attitudes held by the subjects in Ginet’s examples
might be better characterized as what Michael Bratman calls “acceptances”
rather than beliefs. In Bratman’s view, if one’s cognitive attitude shifts
depending on certain “practical but context-relative pressures,” then it is
better classified as acceptance rather than belief. If I were closer to home
when I wondered whether I locked the door, I could go back and check and
so would not accept that I locked it, but, given the context, I accepted it
in my deliberations and planning. One of Bratman’s primary motivations
for arguing that we should think of there being a distinct cognitive attitude
of “acceptance” is that he thinks it is essential to beliefs that they are con-
text independent and not under voluntary control. Ginet’s argument and his
reflection on the cases given are supposed to put into question those assump-
tions about belief. Both Bratman and Ginet use the language of betting; Brat-
man says that some propositions he can accept or take for granted in the
background of practical reasoning are not ones that he would bet on being
true which he would if he believed them. But Ginet’s contention is that once
the subjects in his examples count on the propositions being true, they would
be willing to accept the bet that they are.
16. Owens, Reason Without Freedom, 2.
17. Hume, Enquiry Concerning Human Understanding, 12. There are a number
of ways that an idea can acquire sufficient force to become a belief, the most
common being when the idea and impression have been constantly conjoined
in past experience. The most extensive discussion of the different types of
belief-forming mechanisms is found in Lorne Falkenstein, “Naturalism,
Normativity and Scepticism in Hume’s Account of Belief,” 29–72. Although
he only identifies three main types—reason, association, and educa-
tion—he provides a detailed analysis of the different ways that ideas can
 The Puzzle of Doxastic Responsibility 89
be enlivened by impressions (34–40). For a discussion of Hume on belief
that emphasizes the its phenomenological features, see Jennifer Smalligan
Marusic, “Does Hume Hold a Dispositional Account of Belief,” 155–184.
18. Hume, A Treatise of Human Nature, 147.
19. Ibid., 149.
20. Hume, Enquiries Concerning Human, 110.
21. See introduction, footnote 9 for references.
22. Hume, A Treatise of Human Nature, 148.
23. Levy, “Doxastic Responsibility,” 127–155. I will later address whether his
arguments apply to the account I defend.
24. Levy’s use of the language of excuses here is odd. Usually, we are excused
when we have done something that is normally wrong but where the specific
circumstances exempt us from blame. Yet, Levy argues that the whole notion
of blaming people for beliefs is mistaken; there is no wrong to be excused.
25. Levy, “Doxastic Responsibility,” 145.
26. Ibid., 147.
27. In arguing against doxastic voluntarism, Robert Audi enumerates several rea-
sons why beliefs are not actions. Actions are events, he says, “in the ordinary
sense in which the occurrence of an event entails that of a change. Beliefs
are not events . . . To believe is not to do something or change anything . . .
Beliefs then are not actions” (Robert Audi, “Doxastic Voluntarism and the
Ethics of Belief,” 105). Audi goes on to say that belief formation is an event
but that it is not an action because it cannot be done intentionally or for
practical reasons. In his discussion of doxastic agency, Matthew Boyle sug-
gests that besides exercising our agency in “actively changing thing,” we can
also exercise it “in actively being a certain way” and “for a rational creature,
belief itself is an active condition” (19). To make sense of this kind of activity,
Boyle uses Aristotle’s kinesis/energia distinction. An energia consists in actu-
alizing a capacity, but not for any further end, a kind of “actively maintained
condition” (20). Our capacity to regulate our beliefs and respond to rea-
sons, Boyle argues, reveals that “a person’s believing something on a certain
basis is in a perfectly good sense, an energia of her capacity for doxastic self-
determination” (21).
Conor McHugh’s recent discussion of doxastic agency also questions the
simplicity of the belief/action divide. While he concludes that believing is not
an action, he does argue that beliefs are under agential control “because nor-
mally we can directly determine what beliefs we have by acting.” Again our
capacity to reason and react to reason in belief regulation reveals this activ-
ity. See McHugh’s “Epistemic Responsibility and Doxastic Agency,” 132–57,
and his, “Exercising Doxastic Freedom,” 1–37. Boyle and McHugh’s accounts
have a lot in common with the reasons-responsive views I will discuss in the
next chapter.
28. See for example, Jonathan Bennett “Why Belief is Involuntary,” and Audi,
“Doxastic Voluntarism.”
29. Steup, “Doxastic Freedom,” 375–392.
30. Ibid., 388.

REFERENCES

Adler, Jonathan. Belief’s Own Ethics. Cambridge: MIT Press, 2002.

Alston, William P. “The Deontological Conception of Epistemic Justification”
Philosophical Perspectives 2 (1988): 257–299.
90 Doxastic Responsibilities
Audi, Robert. “Doxastic Voluntarism and the Ethics of Belief.” In Knowledge,
Truth, and Duty, edited by Matthias Steup, 93–108. Oxford: Oxford University
Press, 2001.
Bennett, Jonathan. “Why Belief Is Involuntary.” Analysis 50 (1990): 87–107.
Boyle, Matthew. “ ‘Making Up Your Mind’ and the Activity of Reason.” Philoso-
phers’ Imprint 11, no. 16 (2011): 1–24.
Bratman, Michael. “Practical Reasoning and Acceptance in a Context.” Mind 101,
no. 401 (1992): 1–13.
Descartes, Renee. Meditations on First Philosophy. Translated by David Cress, 3rd
ed. Indianapolis/Cambridge: Hackett Publishing Company, 1993.
Falkenstein, Lorne. “Naturalism, Normativity, and Skepticism in Hume’s Account
of Belief.” Hume Studies 23 (1997): 29–72.
Feldman, Richard. “The Ethics of Belief.” Philosophy and Phenomenological
Research 60, no. 3 (2000): 667–695.
Fischer, John, and Mark Ravizza. Responsibility and Control: A Theory of Moral
Responsibility. Cambridge: Cambridge University Press, 1998.
Ginet, Carl. “Deciding to Believe.” In Knowledge, Truth, and Duty: Essays on
Epistemic Justification, Responsibility, and Virtue, edited by Matthias Steup,
63–76. New York: Oxford University Press, 2001.
Hume, David. Enquiries Concerning Human Understanding and Concerning the
Principles of Morals, edited by Selby-Bigge and Nidditch. Oxford: Clarendon
Press, 1975
———. A Treatise of Human Nature, edited by L. A. Selby-Bigge and P. H. Nidditch,
2nd ed. Oxford: Clarendon Press, 1978.
Levy, Neil. “Doxastic Responsibility.” Synthese 155 (2007): 127–155.
Marusic, Jennifer Smalligan. “Does Hume Hold a Dispositional Account of Belief.”
Canadian Journal of Philosophy 40, no. 2 (2010): 155–184.
McHugh, Conor. “Epistemic Responsibility and Doxastic Agency.” Philosophical
Issues 23, (2013): 137–157.
———. “Exercising Doxastic Freedom.” Philosophy and Phenomenological
Research 88, (2014): 1–37.
Owens, David. Reason without Freedom: The Problem of Epistemic Normativity.
London: Routledge, 2000.
Stanford Encyclopedia of Philosophy, “Decision Making Capacity.” Last revised
June 20, 2011. http://plato.stanford.edu/entries/decision-capacity/
Steup, Matthias. “Doxastic Freedom.” Synthese 161 (2008): 375–392.
Strawson, Peter. “Freedom and Resentment.” In Freedom and Resentment and
Other Essays. London: Methuen.1974.
5 Responsibility without
Voluntary Control

Given that beliefs cannot directly result from decisions or intention, what
does this tell us about our being responsible for beliefs? As we have seen,
one possibility is that it is a mistake to praise or blame people for the beliefs
they hold, or to see them as responsible for those beliefs. If this were the case,
then our attributions of responsibility in the doxastic realm, which are com-
mon, would all be mistaken. Most theorists, however, do not want to accept
the view that there is such widespread error when it comes to common views
about beliefs. If one takes it that responsibility entails voluntary control and
one agrees that we lack this control, then one would have to deny that we
are responsible for our beliefs. If one thinks we can (and ought) to be respon-
sible for our beliefs, and agrees that we lack voluntary control over beliefs,
one must conclude that responsibility does not entail voluntary control.
Here are three prominent characterizations of doxastic responsibility
without voluntary control:1

Responsibility is what is registered when blame is applied to persons as

such . . . we are blamed for our vices, for those traits of character, mani-
fested in thought, feeling, and action, which makes us bad people . . . the
key concept for any theory of responsibility should be responsiveness to
reasons, not agency or control. As well as actions, I am accountable for
those states of mine (beliefs, desires, and emotions) that are governed
by reason, at least where I am capable of responding to reasons. Neither
the scope of the will nor the power of reflection determines the bound-
aries of responsibility. Virtue and vice are matters of my responsiveness
to certain kinds of reason—ethical virtues concern ethical reasons, epis-
temic virtues concern epistemic reasons—and I am praised or blamed
accordingly.2
By settling certain questions for oneself, by having a take on what
is true, what is important, what is to be done, one thereby constitutes
those bits of one’s mind relevant to the quality of one’s relations with
others—and so establishes what we might call one’s moral personality,
or . . . the quality of one’s will. But that bit of one’s mind—one’s moral
personality or one’s will (broadly construed)—just is the object of moral
assessment and reaction . . . The scope of the voluntary should start to
92 Doxastic Responsibilities
seem considerably narrower than the scope of the active or the range of
things for which we are responsible.3
In order for a creature to be responsible for an attitude, on the rational
relations view, it must be the kind of state that is open, in principle, to revi-
sion and modification through the creature’s own processes of rational
reflection . . . it is those mental states which we regard as normatively
connected to a person’s evaluative judgments that we take to be attribut-
able to her for the purposes of moral appraisal.4

While accounts of responsibility for nonvoluntary states, of which there

has been a recent proliferation, differ in their details, most fall into two catego-
ries, and there is overlap between the two. The first can be grouped under the
heading of “character-based accounts” of doxastic responsibility. They argue
that the proper target of assessment when considering attributions of respon-
sibility is not the particular state being judged (for example, a particular
belief), but rather something deeper about the person. The second can be
called “reasons-responsive accounts” of doxastic responsibility. For these
accounts, it is sufficient that a state be “responsive to reasons” for one to
be held responsible for it, and beliefs are open to rational assessment in this
way. Owens clearly falls into the first category, using the familiar language
of character and virtue that one finds in virtue epistemology. While some of
what Smith and Hieronymi say also suggests that the proper target of assess-
ment is not the actual state but something deeper about the person—their
“moral personality” or “moral personhood”—their accounts do not appeal
as straightforwardly to character or virtue and so they are not vulnerable to
some of the problems that Owens, and other virtue accounts, faces. In my
critical discussion of these accounts and in arguing for my view of doxastic
agency, I am assuming it is a strike against an account if it cannot accommo-
date some of the central features of our ordinary practices. This is not to say
that everything we actually do is what we should do but, all else being equal,
a view that captures more aspects of our ordinary responsibility practices is
preferable to a view that captures fewer aspects.5

1. SOME PROBLEMS WITH CHARACTER-BASED ACCOUNTS

According to the character-based view, if I hold a belief and am not responsive

to the reasons that reveal the belief’s irrationality, I can be held respon-
sible (and blamed) for holding this belief. This is the case even if I could
not have exercised better control, or even if I cannot alter my belief now.
Rather, holding irrational beliefs is an epistemic vice, which, in turn, reveals
a defect in my character; we are responsible for what determines our virtu-
ous or vicious character, even if we cannot control these determinants. For
example, Owens considers gullibility, dogmatism, and “weak-mindedness”
all to be distinctively epistemic vices. Even if the beliefs that result from these
 Responsibility without Voluntary Control 93
vices do not result in any morally problematic actions, I am blameworthy
for holding them, because they reveal a defect in my character: “Even before
anything bad happens, my gullibility means that I cannot be an esteemed
human being because I cannot be trusted to think and feel as I ought.”6
Locating all responsibility for belief at the level of character is problematic.
First, our tendency to blame people with whom we have little experience
casts doubt on the character-centered view. It seems like we are able to make
justified general claims about what adult believers should or should not
believe, and we blame them for particular beliefs, regardless of what their
general doxastic tendencies are. Consider someone who does, for example,
possess the epistemic virtue of wisdom, who as Owens puts it, “knows to
whom credit is due, at what point to form a view, when to open his mind
and when to close it,”7 and most of the time believes in such a way that
manifests this virtue. If, one time, perhaps when overcome with jealousy,
he forms the false belief that his wife is unfaithful on insufficient evidence,
would it really make sense for us to say, “Shame on you, you are lacking
merit as a person as you clearly have not cultivated the virtue of wisdom”?
Can we really even assess whether someone possesses a particular virtue
based on one instance? It seems not, but blame still seems appropriate in
this one instance. He is blameworthy because he has a belief he ought not to
have, and, I will argue,8 at least a part of our blame does indicate we think
he has failed to exercise a kind of control.
As we have seen, these character-based views think that the particular
belief is not the proper target of assessment, and so the practice of blaming
people for holding particular beliefs is inappropriate. Adler’s view of doxastic
responsibility is also in the family of those now under discussion, except he
is clear that the belief itself is the target of our reactive attitudes. He says,
“Although the blame is attached only when these beliefs are visible or acted
on, the blame is for the belief itself. We think it is blameworthy for people
to hold such beliefs, given the available evidence.”9 Adler thinks that it is
“deeply misleading” to apply a certain kind of deontological language to
beliefs. When he says “one ought to believe that p only if one has adequate
reasons that p,” this “ought” is not pointing to a duty or a direction. Because
Adler thinks it is conceptually impossible to believe without taking yourself
to have adequate reasons for your belief, the “ought” is taken as more of
a “must,” and thus, “when I recognize that the evidence establishes (fails
to establish) that p, it makes no strict sense to say I ought (or that it is not
the case that I ought) to believe p.”10 He does not, however, want to deny
responsibility, because he thinks it is possible for one to be responsible for
nonvoluntary states of mind. He says, “Even if one cannot choose to believe
otherwise, beliefs are capable of responsivity to reasons, and that is crucial
for responsibility.”11
Adler does not offer a detailed account of responsibility; his discussion
of this issue comprises less than three pages and relies heavily on references
to Adams and Fischer. Perhaps if he were to provide a fuller view, he would
94 Doxastic Responsibilities
come to a similar conclusion concerning the actual target of assessment that
these character-based accounts do. But it is interesting that the more intuitive
and straight-forward answer to the question of what is being blamed when
one believes, for example, that the earth is flat, is the holding of the par-
ticular belief. Though I may later learn some deeper facts about the believer
which could either mitigate or exacerbate my reaction, my initial reaction
seems not to be targeting anything beyond the specific state of holding this
particular belief.
If this seems right—that we normally take ourselves to be assessing or
reacting to a particular state rather than ones’ personality or character—then
these character-based views which argue that we are responsible despite our
lack of any kind of control, seem to be closer to those which deny the legiti-
macy of our attributions of doxastic responsibility than it initially appears.
For ultimately, it seems, according to these views, that the kind of respon-
sibility we have for beliefs is derivative. We are responsible for our beliefs
only insofar as we are responsible for cultivating virtues. But, as we have
seen, those who argue we lack doxastic responsibility also allow for a kind
of derivative responsibility. Both accounts claim that we lack control over
what we believe; they differ on whether this lack of control exempts us from
responsibility. As I have stated before, that a theory fails to capture central
features of our ordinary practices does not mean it is incorrect. But if it is
possible to make sense of the idea of a non-derivative doxastic responsibil-
ity, that preserves the idea that we are responsible for specific beliefs, then
such a view is preferable.
Another problem with these character-based accounts is that if we really
blame believers for their character traits, then we would blame someone less
for the occasional lapse than someone who often fails to respond to reason.
I think the opposite tends to be true. If one has grown up in such a way as to
be able to exercise so-called epistemic virtues, but fails to keep one’s beliefs
in line with one’s more reflective judgment, then one can be blamed for this
lack of control. If someone has never, for example, learned to proportion
his belief to the evidence, this person seems less blameworthy than one who
has the capacity but fails to do so. For example, if someone grew up in
a family and community with a racist ideology, where access to evidence is
controlled and limited, his responsibility for the racist belief that whites are
essentially superior to blacks will likely be somewhat mitigated. We would
be less likely to mitigate responsibility for someone who has had more open
access to the evidence and, in general, is seen as a wise and reflective person.
The fact that we can even make the distinction between the lapses of the
reflective person and the unreflective one seems problematic for this character-
based view. For, the wise person’s beliefs will be more responsive to reasons
than those of the unwise, and when the wise person’s beliefs fail to be so,
we blame him, at least partly, for lacking self-control. This is not to say that
he is generally a person of incontinent character, but that he is lacking self-
control in this instance. We are blamed, it seems, when we do not exercise
 Responsibility without Voluntary Control 95
our reflective competence that helps us believe the way we ought to believe.
Interestingly, W. C. Clifford points this out, saying: “Every time we let our-
selves believe for unworthy reasons, we weaken our powers of self-control,
of doubting, of judicially and fairly weighing evidence.”12 Clifford thinks
that believing for unworthy reasons can lead to immoral actions and can
exhibit a number of vices, but it also reveals a lack of control.

2. REASONS-RESPONSIVE VIEWS

I have been discussing a family of views and considering problems that they
all face. I will turn now to some of the details of these accounts. We will see
that they may not all be equally vulnerable to the objections I have raised.
While reasons-responsiveness is central to them all, some of the concerns
I have raised apply more to accounts that explicitly invoke character and
virtue. But there are further problems that apply to the more specific ver-
sions of these characterizations of responsibility without voluntary control.

(i) Passivity vs. Activity: Is There Any Kind of Doxastic Control?

Because beliefs cannot simply and directly result from decisions the way
many actions can, some have tried to separate responsibility entirely from
freedom and control. I have been urging, however, that beliefs are products
of our agency, something we have an active role in shaping and maintain-
ing. Even if we cannot believe “at will,” neither are we passive in the beliefs
we form and maintain. Of the three representative views I have been con-
sidering, the one who most clearly delinks responsibility from any kind of
control or agency is Owens’s view. To repeat, he says, “The key concept for
any theory of responsibility should be responsiveness to reasons, not agency
or control.”13 Owens argues that not only can we not believe at will, but
neither do we have reflective control over our beliefs; a belief’s rationality
is not constrained by one’s judgment of it the way an action is. A state is
under reflective control, for Owens, “when it can be motivated by higher
order judgments about the probative force of the reasons for it.”14 He cites
McDowell, Locke, and Descartes as theorists who think we do have such
control over our beliefs, that our higher order judgments about what we
ought to believe help to dictate what we do believe. Owens denies that we
have this control; it is not by reflecting on reasons for believing, he says, that
one comes to believe. I may reflect on evidence for or against p in forming
a belief of p, but I will not reflect on whether believing p is rational, justi-
fied, desirable, or useful. We cannot, he argues, decide to favor one goal of
believing over another and thereby believe it is raining or that Jones is guilty
as a result of this decision. “Purposiveness implies that a subject form beliefs
in pursuance of a certain goal, that they control their beliefs by aiming at
that goal.”15 But believers lack this kind of control; our beliefs do not seem
96 Doxastic Responsibilities
goal-oriented the way that our actions are. So, Owens argues, our blaming
people for beliefs they hold is not a result of our taking them to have failed
to exercise control. It is not the case, he says, that they have failed to keep
their beliefs in line with their more rational goals.
Smith’s rational-relations account of responsibility may share some
features of the McDowlenian view that Owens criticizes, and she seems to be
trying to make some room for doxastic agency. Smith begins her discussion
by pointing out an apparent tension in our ordinary thinking about the
conditions for moral responsibility. It seems, on the one hand, that we think
of voluntary control or choice as a precondition of legitimate moral assess-
ment and yet “we regularly hold ourselves and others responsible for things
that do not appear to reflect a conscious choice or decision.”16 These include
involuntary attitudes, reactions, and patterns of awareness. We fault people,
for example, for their failures to notice, for forgetting, for having certain
worrisome thoughts occur, for having certain emotional reactions. Beliefs
are included in Smith’s list of states for which we lack voluntary control
but are nonetheless responsible, though none of the specific examples she
discusses focus on beliefs. Smith’s way of resolving the tension, a tension
very similar to the one revealed by the puzzle of doxastic responsibility that
I outlined, is to argue against the volitionalist criterion of responsibility and
argue instead for a “rationalist” one. The activity implied in our moral prac-
tices, she says, is not the activity of choice, but rather is the activity of evalu-
ative judgment: “what makes an attitude ‘ours’ in the sense of relevant to
questions of responsibility and moral assessment is not that we have volun-
tarily chosen it or that we have voluntary control over it, but that it reflects
our own evaluative judgments or appraisals.”17
Unlike Owens, Smith does not want to deny that responsibility is tied
to agency, rather she thinks that the voluntary/ nonvoluntary division does
not map on to the passive/active division. Instead, she says, “The dividing
line between activity and passivity goes through judgment: I am active, and
responsible, for anything that falls within the scope of evaluative judgment
(i.e., anything that is, or should be sensitive to my evaluative judgments
and commitments).”18 For example, if I express fear, then it is reasonable
to suppose that this fear is rationally connected to an evaluative judgment,
a judgment that the object feared is dangerous. If my good friend forgets
to meet me when I have said I really need to talk, according to Smith, the
reason I will feel angry or resentful is because I will see this attitude as
expressing a judgment, one that places less value on my well-being than
I can reasonably expect of a friend. And these evaluative judgments and
commitments are of the kind for which we take a person “to be directly
morally answerable.”
Given that, according to Smith, these involuntary states are a “direct
reflection of what we judge to be of value, importance, or significance”; they
“are the kind of states for which reasons or justifications can appropriately
be requested.”19 Such is not the case with nonintentional mental states—such
 Responsibility without Voluntary Control 97
as physical pains, sensations—or physiological conditions—such as hunger
or thirst. We do not expect these states to be “rationally sensitive to our eval-
uative commitments . . . we are essentially ‘passive’ in respect to these states,
then because they are not the kinds of states that either directly reflect or
are supposed to governed by our underlying evaluative judgment.”20 One
of Smith’s central aims is to show that we express our agency “not only in
our explicit choices and decisions, but also in what we unreflectively think,
feel, desire and notice.”21 And again, this is the case because these states are
thought to express what one values and cares about, and such evaluative
judgments express what kind of person one is.
We have seen that Owens says “no agency” is required for responsibility,
whereas Smith thinks that any attitude that can be taken to reasonably
imply an evaluative judgment is an expression of rational agency. But does
her view connect responsibility to control at all, even if it is not the control
of choice or decision? Are these states, for which we are responsible, under
our reflective control? This is a difficult question but, in the end, I think
the answer is “no.” Smith is clear that the connection of the unreflective
states to evaluative judgments need not be, and most of time is not, one
of which the agent is aware; it is usually not a deliberate expression of
an evaluative judgment. These judgments that comprise the things we care
about or regard as important or significant “do not always arise from con-
scious choices or decisions, and they need not be consciously recognized by
the person who hold them.” She says that one may only discover that one
has certain evaluative commitments through one’s own responses in certain
situations. For example, I could discover that I care more about being liked
by others than I do about standing up for my moral principles. It is at this
point of awareness where it seems some control enters: “Once one realizes
that one holds a certain evaluative judgment, it is open to one to determine
whether one has adequate justification for that judgment or to modify it
or give it up if that justification cannot be provided.”22 If I react with fear
to a spider but am convinced that there is no justification for the judgment
that spiders are dangerous, I should no longer be fearful of them. And the
“should” here in question is “the should of rationality and, therefore, marks
a normative ideal.”
It seems, then, that there is an expectation that rational agents’ attitudes
will, for the most part, be consistent with the evaluative judgments to which
they seem to be linked. It may seem that the locus of responsibility, and so
where blame may be appropriate, is when the awareness, and capacity to
control, revise, and modify enters in. But Smith thinks reactive attitudes,
and attributions of responsibility, are appropriate for one’s involuntary state
(such as an emotional responses or a belief), even if one is not conscious
of the evaluative judgments to which it is taken to be rationally related.
This is because the norm of rationality allows us to form expectations that
certain attitudes and evaluative judgments are linked together. If these eval-
uative judgments are morally or epistemically problematic, then negative
98 Doxastic Responsibilities
reactive attitudes are appropriate, because this shows that there is something
problematic or defective about the person’s values and, ultimately, with the
person.
We see, then, that, despite Smith’s talk of activity and agency, her view dif-
fers little from Owens’s account. Remember, for Owens, “I am accountable
for those states of mine (beliefs, desires and emotions) that are governed by
reason, at least where I am capable of responding to reasons.”23 But what
kind of capacity is required here? Owens considers the case of anger and
“the associated vice of intemperance”24 as an example of state for which
I am blameworthy despite lacking the capacity to alter it, or having volun-
tarily acquired it. If my reactions and behavior “expresses a psychological
state itself subject to rational assessment (like anger) I am culpable. I am
responsible for the harmful effects of my anger even though my intemper-
ance nor its manifestation are under my direct control.”25
Owens imagines that the theorist who wants to maintain that responsibility
entails control, who he refers to as the “juridical” theorist, will say there
are various strategies I can take to cultivate temperance and self-restraint
and that what I am being blamed for is my failure to take the steps needed
for such cultivation. To counter the idea that this is what is going on when
we hold people responsible for, for example, their road rage, Owens consid-
ers two cases where each person puts the same amount of time and effort
into self-improvement. Each person assiduously attends anger management
classes and does their relaxation exercises in the car simulator with the same
level of commitment, but the therapy works for one and not for the other.
Though knowledge of these efforts may temper one’s reproaches, one remains
blameworthy for these reactions and “continued guilt . . . would not be inap-
propriate.” Because, according to Owens, what these reactions are targeting
is one’s bad character, blame is appropriate no matter its cause or its possibil-
ity of changing. I can be admired for courage of perseverance in my efforts,
but if my anger persists, “that does nothing to diminish my responsibly for
my feelings.” This is because, according to Owens, the possession of a virtue
(and the absence of a vice) has an “ethical significance which is quite indepen-
dent of the (ethical) feasibility required to cultivate it.” Excuse can only arise
when one’s states are not at all responsive to reasons. We see now that being
responsive to reasons does not mean that one can actually respond to reasons
showing a state to be unjustified. It means only that one has a capacity to
rationally assess the state, to see it as being connected to a virtue or a vice
that either increases or decreases one’s personal merit: “one is held to account
for one’s merit as a person; one gets blamed for those things (both moral and
non-moral) which are thought to make one a bad person.”26 Again, as we
saw with Smith, we have a normative ideal and an expectation that rational
agents will live up to it; if they fall short, they are blamed.
Hieronymi’s account has a lot in common with Smith’s view. As we have
seen, she, like Smith, does not think the boundaries of agency should be
 Responsibility without Voluntary Control 99
marked by what is voluntary: She also thinks that when we hold people
responsible for nonvoluntary states, we do so because we view these states
as expressing a person’s “take on, opinion about, or orientation to, some
object”27; they express what one values and finds important. These states
include beliefs. She further says that for such states to retain their signifi-
cance, they cannot be voluntary; we cannot simply decide to choose what
kind of people we are: “beliefs, then, are a central example of the sort of
thing for which we are most fundamentally responsible, in part because they
are the sort of thing which could not be voluntary.”28 She calls beliefs, along
with emotions, reactions, and intentions, “commitment-constituted atti-
tudes.”29 Our beliefs and feelings, “reveal one’s answers to certain questions,
and so can reveal one’s mind, one’s (moral or epistemic or rational) self,
which is the object of both assessment and reaction.”30
So, like both Owens and Smith, Hieronymi thinks we do not have volun-
tary control over beliefs. Much like Owens, and others, she says one cannot
believe something because one thinks it would be good to believe that thing:

One cannot believe by forming and executing an intention to believe.

This is what makes beliefs different from voluntary actions. We have
seen, one might well take the fact that believing p would make a good
joke, or relieve one’s boredom, or spare one anguish, to count sufficiently
in favor of believing p. But one cannot, for those reasons, settle the ques-
tion of whether to believe p, therein form an intention to believe p, and
execute that intention by believing p. This is because, when one believes,
one is understood to have settled the question whether p, but one can-
not be understood to have settled that question for reasons which one
only takes to show believing good to do.31

This is contrasted with the case of a regular voluntary action:

One intends to raise one’s right hand in settling the question of

whether to raise it, where that question could be settled by any (set of)
consideration(s) that one takes to count sufficiently in favor of raising
one’s right hand.32

But where does Hieronymi stand on the question of whether we exercise

any kind of control over belief, and whether our holding people responsible
entails that they have a kind of control? She says that to say they are not
voluntary, “is not to say that they are involuntary, or something in relation to
which we are passive . . . The fact that we do not exercise voluntary control
over these attitudes need not entail that we do not exercise any control over
them. One must simply allow that there are other forms of control,”33 and
“because these attitudes embody our answer to some question(s), we exercise
a distinctive form of control over them.”34 So what kind of control is this
100 Doxastic Responsibilities
“distinctive,” “non-voluntary” control that we have over beliefs and other
commitment-constituted attitudes? Here is a little more elaboration on its
nature:

Because these attitudes embody our take on the world, on what is or is

not true or important or worthwhile in it, we control them by thinking
about the world, about what is or is not true or important or worthwhile
in it . . . Because our minds change as our take on the world changes—
because our minds change as we change our minds—we can be said to
be “in control” of our commitment-constituted attitudes. This is an odd
kind of control, to be sure: the object controlled and the subject who
controls are not distinct. There is no reflective distance in the offing. We
change our minds, and so control our attitudes, not by reflecting on or
thinking about our mind, but rather by thinking about the object of our
thoughts. The controlling happens “behind the lens,” so to speak. The
thinking subject controls its thoughts in thinking them.35

It is not obvious why what is here described counts as a kind of control.

At times, it seems Hieronymi is simply equating being in control with being
“active” or not passive. In laying out her description of this kind of control,
she says that we do not control them the way we do our actions but they
are not “for that states to which we are passive, or simply things which we
can only affect and manage through our actions, like our furniture or our
allergies.”36 How do we affect and manage them? She says it is by thinking
about the objects of our thoughts. We don’t control them by thinking about
the thoughts themselves, which is the kind of control Smith suggests we
have. According to Smith, I could reflect on what a certain reaction seems to
indicate about me and then alter my thinking if I do not find the evaluative
judgments defensible. But here, Hieronymi says that “there is no reflective
distance” between what is being controlled and what is controlling. How
much control do I have over the thoughts I think? I can turn my attention
to certain features of the world, be open and sensitive to different kinds of
information. I can ask myself questions. If this constitutes any kind of con-
trol, it is of a very indirect kind. What seems most crucial to her account,
however, as is the case with Smith, is that these attitudes are the product of
reflective activity and, it is because they are expressive of such activity, that
we can ascribe responsibility to them.
In a more recent paper,37 Hieronymi elaborates on this “odd kind of con-
trol.” There, she terms the kind of control we have over nonvoluntary states
“evaluative control.” She says,

that “if an attitude embodies our answer to a question or set of ques-

tions, then it seems we will form or revise such an attitude in forming or
revising our answers to the relevant question(s) . . . We might say that we
control these aspects of our minds because, as we change our mind, our
 Responsibility without Voluntary Control 101
mind changes—as we form or revise our take on things, we form or revise
our attitudes. I call this exercising evaluative control over the attitude.38

She notes that this kind of control does not have some of the paradig-
matic features of agency (of voluntariness and reflective distance), but that
it should be considered a kind of agency nonetheless. Her main argument
in support of this view is that our intentions do not exhibit these familiar
features and yet no one would want to deny that we exercise agency over
our intentions. While this discussion elaborates some on the different kinds
of control, it says little about how we exercise this control over belief and,
more significantly for the purpose of the present discussion, nothing about
how or whether this kind of control grounds attributions of responsibility.
Does evaluative control come in degrees? Can we lose the capacity to con-
trol our attitudes in this way, and, if so, are we no longer responsible for
them? Hieronymi shares my basic contention about beliefs, namely that it is
a mistake to see ourselves as passive in respect to them. Our ways of trying
to articulate the way in which we are active in our doxastic lives differ in
the details though they share a lot of features in common. Which is prefer-
able will depend on their explanatory power. Given that the aims to which
we put our respective accounts differ, this may be hard to discern.

(ii) Responsibility for Our Values and Commitments

All accounts that claim that one is responsible for one’s character, or for
the kind of person one is, need to respond to the worry that so many of the
sources of who one is lie outside of oneself. Owens clearly states that we are
responsible for our merits as persons even if we cannot control these deter-
minants or alter them. Again, in his discussion of temperance, Owens says
that when assessing someone’s culpability for vice, we may, at times, take
into account unusual features of his personal history. He says, “The rage of
someone terribly abused as a child is less resented than that of a person with
a normal upbringing, and temperance in such a person is more admired.”
But, he claims this difference in assessments does not turn on any question
of having more or less control over the state being assessed. Instead, we
are inclined to mitigate blame, because we have “doubts about whether
this person’s emotions are responsive to reasons at all, given his unusual
upbringing.” But remember what a minimal criterion it is for a state to be
responsive to reasons; one need only be aware that the state can be ratio-
nally assessed, that it can be justified or unjustified, and that this awareness
has some motivational force. It seems that what it would take for a state
to lose this kind of responsiveness is severe pathology. It would mean that
one has lost the capacity of any rational assessment in a particular domain.
This could happen, and I will consider some such cases later, but it seems
such severity is not needed to mitigate the blame in the case of someone with
a history of abuse. It seems, instead, that part of what matters is that they do
102 Doxastic Responsibilities
not seem to have constituted their own minds the way that others have. It is
not clear that their commitment constituting attitudes are as clearly theirs.
Even if we accept Owens’s diagnosis of this kind of case, what ends up
explaining our mitigation of responsibility is recognition of a diminished
capacity. And again, in his summary of what is required for responsibility, he
says I am accountable for those states where “I am capable of responding” to
reasons. But talk of capacity seems to bring back in questions of freedom and
control. If I am furious at you for missing our date, and I find out that you
were incapable of coming because you got into an accident on the way, then
I will no longer blame you. But this is the picture of responsibility Owens is
arguing against. He says that whether someone has control over a state does
not matter for responsibility. In the road-rage example, I am responsible even
if there is no way I can get rid of it. So the capacity to rid oneself of the state
or the vice does not matter; the capacity to assess, judge, and be motivated
matters. But, it seems, such a capacity does entail a certain kind of control.
One who can recognize that an emotional reaction is problematic or a belief
is unjustified and be motivated to alter it seems to be exercising more agency
over such states than someone who lacks such recognition and motivation.
Smith would say the first person is active while the other is passive.
Smith explicitly addresses the concerns about past histories mattering for
assessments of responsibility. For much of her discussion, she is concerned
to show that history does not matter; how one came to have the evaluative
judgments one does is irrelevant for assessments or responsibility for
attitudes. All that matters is that

the attitude is, or should be, sensitive to her evaluative judgments and
that she can properly be asked to defend or justify it . . . The fact that
a person’s evaluative judgments are usually shaped in various ways by
her early attachments and environment does nothing to undermine the
claim that they are still genuinely her judgments.39

But, she says, the conditions under which a person’s evaluative judgments
were formed can be relevant for some kinds of assessments. She considers
two people, Abigail and Bert, whose attitudes reflect evaluative judgments
that are deeply racist and religiously intolerant. Abigail was raised in a
racist community and developed “evaluative tendencies and corresponding
attitudes in line with those she sees operative in her family or surrounding
community. As a adult, her attitudes may continue to reflect the vicious eval-
uative judgments thus formed in her childhood.” Bert, on the other hand, is
someone who was raised in “a loving and tolerant home and community, but
who later in life reflectively comes to adopt racist and intolerant values.”40
Smith says that if the question is about being responsible for becoming
a racist or intolerant person, then Bert is responsible and Abigail is not
(or she is less so) but, she says, “this question of responsibility (namely,
the responsibility one has for becoming a certain kind of person) must be
 Responsibility without Voluntary Control 103
distinguished from the question of one’s responsibility for attitudes one in
fact holds.”41 All that matters for responsibility for an attitude is that it is
reasonable to ask one to defend the judgments the attitude reflects. The ori-
gin of such attitudes does not affect the question of their justification. Like
Owens, Smith says we may be less critical of Abigail, realizing it would be
harder to modify these judgments of such deeply entrenched patterns, but
she cannot escape responsibility. Again, as in Owens’s account, the only
way one ceases to be responsible is if a person “is literally incapable of
appreciating and responding to rational criticism directed at her evaluative
judgments.”42 To deny someone such a capacity, Smith says, “can be deeply
patronizing and disrespectful, and we should not be too eager to resort to
them, either in our case or in our treatment of others.”43
I find Smith’s discussion of these cases puzzling in a couple of ways. First,
we once again find an appeal to capacities or lack of capacities as crucial
to assessments of moral responsibility. This may not be as problematic for
Smith as for Owens, since she is mainly concerned with showing that a cer-
tain kind of control, namely the kind of control presupposed by the voli-
tionalist account of responsibility, is not actually required for appropriate
attributions of responsibility. This control centers on choice and decision,
and people are not excused from responsibility if they cannot choose or
decide otherwise. In Smith’s account, one is excused from responsibility if
one does not have the capacity to alter one’s evaluative judgment. But to say
one is incapable of altering one’s evaluative judgment seems very similar to
saying one cannot choose or decide to alter it. We blame Bert more, it seems,
because he went through a deliberative process, one over which he had con-
trol, and so he could exercise the same control to undo these judgments. It
seems it is not only that he is more responsible for being the kind of person
he is but also is more responsible for having the particular belief because he
has more control over its existence than does Abigail.
Smith has said that being responsible for being a certain kind of person
must be distinguished from being responsible for the attitudes one has. But
Smith’s way of explaining our responsibility for attitudes is by connecting
them to the values and commitments we have which reveals one’s moral per-
sonhood. If one really were not responsible for being the kind of person one
is, how does it make sense to say that one’s evaluative judgments are really
“her judgments”? It seems whether one’s judgments are really one’s own
depends on facts about how they came about. I think part of our tendency
to mitigate our blame in a case like Abigail’s has to do with our questioning
whether her mind is as self-constituted as is Bert’s.

(iii) The Possibility of Generalizing and of Mitigation

The last concern I brought up for character-based accounts is that in some
ways the assessments of responsibility can only be very specific. As I argued,
it seems I would need to know what kind of person you are before I can
104 Doxastic Responsibilities
assess whether you are defective or vicious in certain ways. If the target
of assessment is not actually the particular state, for example a particular
belief, but something deeper about you, then I would need a lot of data
before I could attribute a certain vice to you and so be able to react appro-
priately. And again, if you do not have the character trait generally associ-
ated with the attitude to which I am reacting, this doesn’t seem to excuse
you; as I noted in the case of wisdom, it could make matters even worse.
And yet it seems we have the capacity to make general statements about
problematic beliefs and that they are problematic whether or not they reflect
a stable character trait or vice. I can fault you for being dogmatic in one
instance even if, in general, you are extremely open-minded.
Do the reasons-responsive accounts which do not refer to character
and virtue do better in their capacity to make generalizations about when
people are responsible or blameworthy, on the one hand, and when such
responsibility is mitigated on the other? It seems they can do better on the
generalizations but not on mitigations. They appeal to what can be expected,
what can be “properly asked” of one, to what connections between atti-
tudes or judgments can be “reasonably” thought to hold. We can say that
people who hold certain beliefs tend to have takes on the world that are a
misguided or worthy of criticism. Even if it turns out they do not, in fact,
have traits which we usually associate with these attitudes, if it is reason-
able to assume they do, then, it is claimed attributions of responsibility are
appropriate. Here is Hieronymi again on the target of assessment for moral
responsibility:

To morally assess someone on account of some action or attitude—to

say that she was, in that action or attitude, respectful or disrespectful of
others, or to say, in light of that action or attitude, that she was being
kind, generous, magnanimous, spiteful, petty, or cruel—is not just to
make a statement about the action or attitude; it is to make a statement
about the quality of the will or mind or moral personality of which the
action or attitude is a product or part. It is to make assumptions about
the mind from which the action emanated, or the mind into which the
attitude fits, and to react to it—to that mind or moral personality.44

Hieronymi is not saying here that our assumptions are always correct. Her
point is that we will only hold someone responsible for an action or attitude
if we take it to be an expression of her moral personality. This shows how
we can assess someone for holding a particular belief even if we do not know
much about her mind or moral personality. But it becomes unclear what
someone could say to disown a particular nonvoluntary attitude. In the case
of an action, if it is not voluntary, then this shows it is not an expression of
the quality of my will. When it comes to nonvoluntary states, though, such as
beliefs, that they are a part of who I am, of what I value, seems inescapable.
But do we really want to say that we are equally responsible for all our beliefs
 Responsibility without Voluntary Control 105
and emotions and that all agents are equally responsible for such states,
regardless of their circumstances? It seems we need some way of showing
when these generalizations do or do not hold in specific cases, and a way to
have a better understanding of when responsibility is mitigated. I will argue
for a view of doxastic responsibility and agency that allows us to make these
distinctions. My view requires making sense of a kind of doxastic control and
tying responsibility to this control. I will argue that we take responsibility for
our beliefs, and that taking responsibility includes taking control of them.

NOTES

1. The main accounts I will discuss are from Owens, “Reason without Free-
dom”; Angela Smith, “Responsibility for Attitudes”; and Pamela Hieronymi,
“Responsibility for Believing,” and “Believing at Will.” Recently, Pascal Engel
has argued that responsibility for belief requires neither control nor agency.
But the kind of responsibility he discusses is minimal; it simply requires that
there is “a minimal capacity on the part of the subject to recognize his men-
tal state as a belief and be sensitive to the norm for belief.” As long as he is
not a “dumb brute” who observes the “passing show,” then there is a sense,
Engel says, in which the subject is responsible. Whatever kind of responsi-
bility this is, it is not one for which the reactive attitudes are appropriate
(Pascal Engel, “Epistemic Responsibility”). Other recent defenses of doxastic
agency are found in Steup, “Doxastic Freedom”; Maurisic, “Belief and Dif-
ficult Action”; Boyle, “ ‘Making Up Your Mind’ ”; and McHugh’s “Epistemic
Responsibility Doxastic Agency” and “Exercising Doxastic Freedom.” All
of these are variants of the reasons-responsive view for which I use Hiero-
nymi and Smith as examples. Of all of them, McHugh’s is the closest to my
view and he also seems concerned with understanding doxastic freedom and
agency in a way that connects it to agency and freedom, in general. While he
gets closer than others at offering a unified conception of agency, this unity is
threatened by his maintaining a strict division between the kinds of reasons
and norms applicable for belief and action. I will argue that one advantage
of my view is that it does a better job in explaining why the control we have
over beliefs does not differ substantially from the control we have over many
actions, 217.
2. Owens, Reason without Freedom, 123–126.
3. Hieronymi, “Responsibility for Believing,” 361–362, 369.
4. Smith, “Responsibility for Attitudes,” 256.
5. I discuss this view of how to evaluate a philosophical theory in the introduction.
6. Owens, Reason without Freedom, 124.
7. Ibid., 125.
8. In Chapter 6.
9. Adler, Belief’s Own Ethics, 64.
10. Ibid., 51.
11. Ibid., 66.
12. Clifford, “The Ethics of Belief,” 23.
13. Owens, Reason without Freedom, 126.
14. Ibid., 2.
15. Owens, “Epistemic Akrasia,” 381–397.
16. Smith, “Responsibility for Attitudes,” 236.
17. Ibid., 237.
106 Doxastic Responsibilities
18. Ibid.
19. Ibid., 251.
20. Ibid., 257.
21. Ibid., 263.
22. Ibid., 252–253.
23. Owens, Reason without Freedom, 126.
24. Ibid., 117.
25. Ibid., 119.
26. Ibid., 118–119, 121.
27. Hieronymi, “Responsibility for Believing,” 367.
28. Ibid., 372–373.
29. Ibid., 368.
30. Ibid., 363.
31. Ibid., 367.
32. Ibid.
33. Ibid., 369.
34. Ibid., 370.
35. Ibid., 370–371.
36. Ibid., 370.
37. Hieronymi, “Two Kinds of Agency,” 138–162.
38. Ibid., 3.
39. Smith, “Responsibility for Attitudes,” 120–126, 267.
40. Ibid., 267.
41. Ibid.
42. Ibid., 269.
43. In addressing problems with what he calls “the answerability view,” a view
very similar to Smith’s, McHugh (“Epistemic Responsibility”) points out that
such a view would make those who suffer from paranoid delusions respon-
sible for them for most such people are capable or reflective self-governance.
While I understand the worry about the patronizing potential of denying
responsibility, it seems some way of explaining how those how have not
totally lost this capacity can fail to be responsible is needed.
44. Hieronymi, “Responsibility for Believing,” 362.

REFERENCES

Adler, Jonathan. Belief’s Own Ethics. Cambridge: MIT Press, 2002.

Boyle, Matthew. “ ‘Making Up Your Mind’ and the Activity of Reason.” Philoso-
phers’ Imprint 11, no. 16 (2011): 1–24.
Clifford, William K. “The Ethics of Belief.” In The Ethics of Belief Debate, edited by
Gerald D. McCarthy, 19–36. Atlanta: Scholars Press, 1987.
Engel, Pascal. “Epistemic Responsibility Without Epistemic Agency.” Philosophical
Explorations 12, no. 2 (2009): 205–219.
Hieronymi, Pamela. “Believing at Will.” In Belief and Agency, edited by David
Hunter. The Canadian Journal of Philosophy Supplementary Volume 35, (2009):
149–187.
———. “Responsibility for Believing.” Synthese 161, no. 3 (2008): 357–373.
———. “Two Kinds of Agency.” In Mental Actions, edited by Lucy O’Brien and
Matthew Soteriou, 1–30. Oxford: Oxford University Press, 2009.
Marusic, Berislav. “Belief and Difficult Action.” Philosopher’s Imprint 12, no. 18
(2012): 1–30.
 Responsibility without Voluntary Control 107
McDowell, John. Mind and World. Cambridge: Harvard University Press, 1994.
McHugh, Conor. “Epistemic Responsibility and Doxastic Agency.” Philosophical
Issues 23, (2013): 137–157.
———. “Exercising Doxastic Freedom.” Philosophy and Phenomenological
Research 88, (2014): 1–37.
Owens, David. “Epistemic Akrasia.” The Monist 85 (2002): 381–397.
———. Reason without Freedom: The Problem of Epistemic Normativity. London:
Routledge, 2000.
Smith, Angela M. “Responsibility for Attitudes: Activity and Passivity in Mental
Life.” Ethics 115, no. 2 (2005): 236–271.
Steup, Matthias. “Doxastic Freedom.” Synthese 161 (2008)” 375–392.
6 The Possibility of
Doxastic Agency

1. GUIDANCE CONTROL: TAKING RESPONSIBILITY

AND OWNERSHIP

I have now examined and critiqued three ways of addressing the argument
that leads to the conclusion that our attributions of responsibility about
belief are not appropriate because people lack voluntary control over their
beliefs. Here is the argument again:

I. If attributions of responsibility for beliefs are appropriate, then peo-

ple have voluntary control over their beliefs.
2. People do not have voluntary control over their beliefs.
Conclusion: Attributions of responsibility about beliefs are not
appropriate.

One way of responding is to accept the argument as sound and so deny

that we are responsible for beliefs. Doxastic voluntarists, as discussed in
Chapter 4, deny the second premise, namely that people do not have vol-
untary control over their beliefs. The views discussed in Chapter 5 deny
the first premise, namely that if attributions of responsibility for beliefs are
appropriate, then people have voluntary control over their beliefs. They
claim that we can be responsible for more than what is under our voluntary
control. I turn now to my response to the puzzle of doxastic responsibility.
I think it is wrong to divorce responsibility from control, and I accept
that if we do not have control over our beliefs, then our attributions of
responsibility are misguided, and so it seems that I accept the first premise.
But as we have seen some of the accounts that claim we have responsibility
for nonvoluntary attitudes, including beliefs, think we have some kind of
control over these attitudes; Hieronymi, for example, calls it “evaluative
control.”1
Even some of those who deny that we are responsible for beliefs will
allow that we have a kind of indirect control over beliefs. The character-
based accounts of responsibility will similarly allow for this kind of indirect
control, but they do not think our attributions of responsibility are tied in
 The Possibility of Doxastic Agency 109
any way to this kind of control. The kind of control for which I will argue
is not the indirect kind to which these theorists concede; it is not simply
derived from other states over which we do have control. What I want to
defend is (a) that attributions of responsibility and other deontological judg-
ments about beliefs are appropriate and (b) these attributions and judg-
ments presuppose that we have control in the doxastic realm.2
Is the kind of control I claim we have over beliefs that I think is required
for doxastic responsibility a kind of voluntary control? In my discussion of
doxastic voluntarism, I have denied that our beliefs can be the direct result
of decisions, the way some of our actions can, and so if having “voluntary
control” over x entails that we can directly decide to x, then I am accepting
the second premise and denying the first. What is required for voluntary
control is a controversial and complex issue and I won’t get into details here
of various characterizations. Paradigmatic voluntary actions are those that
can be directly executed once one intends the action, like the raising of one’s
hand. It is sometimes said that actions of this kind can be done “at will”
and since we cannot believe “at will,” we do not have voluntary control
over beliefs. If the kind of control I argue we have over our beliefs can be
understood as believing at will, then I am denying the second premise. While
the main purpose of this chapter is to elaborate on the nature of control that
I think is required for doxastic responsibility, I will begin by making some
general claims about it that distinguish it from the views of doxastic agency
discussed in the previous chapter, and will help show why my view could be
counted as a kind of doxastic voluntarism.
I discussed some accounts that agreed the boundaries of agency extend
beyond what can be the direct result of decision, choice or intention.
As Hieronymi rightly points out, such a narrow construal of agency will
leave out many actions as being under our control. While I agree that one
cannot directly decide to believe, I do not think this denial is equivalent to
denying the possibility of believing at will. What exactly is meant by “the
will,” and what does it mean to be able to act or believe “at will”? Again, just
as characterizations of the voluntariness vary, so do views of what constitutes
a person’s will. Historically, it was viewed as a mental faculty with a particu-
lar function, namely that of choice. Possessing such a faculty was thought to
be what allowed a person to act in a way that accords with his own determi-
nations and reasons (barring physical limitations).
While we don’t now generally think of “the will” as a kind of mental
faculty with a particular function, we still employ the language of “will”
quite often, both in everyday and in philosophical discourse. Thinking about
some of these common expressions can help us begin to develop a conception
of what it might mean for a state to be subject to the will. We call some peo-
ple strong-willed and others weak-willed. Someone who has trouble keeping
her actions in line with what she thinks, all things considered, she ought to
do, is weak willed.3 One who can resist temptation and act as he thinks best,
even when it is difficult, has a strong will. Sometimes, we may say (usually of
110 Doxastic Responsibilities
a child) that she is “strong willed,” to mean something like stubborn or head-
strong. But even in this case, the child knows what she wants, what she thinks
is best and does not want her actions to deviate from these determinations.
Frankfurt identifies the will with the desire that is effective, that leads
all the way to action. In such a view, an action done “at will” is one that
accords with what one most (perhaps all things considers) wants to do.4 On
Hieronymi’s view I can do something “at will” if I can do it intentionally,
that is do it for any reason I take to bear sufficiently on it.
This collection of thoughts suggests that one way to think of the will is as
intimately connected with practical reason. An action done “at will” is one
done for reasons, intentionally, decisively, or in accordance with one’s best
judgment.5 One pertinent question then in trying to determine if one has
voluntary control over a mental state or attitude is to ask whether it subject
to the deliverances of practical reason; can reasoning about whether it is
good or right have any effect on whether it manifests or endures? As we
have seen, most theorists (even those arguing for some kind of doxastic
agency) do not think one can ever believe for practical reasons. For example,
Steup argues for a kind of doxastic freedom, arguing that any compatibil-
ist account on conditions for responsible action can be applied to beliefs as
well. But Steup aligns himself squarely with the evidentialists. In considering
Feldman’s view against the idea that we have control over beliefs, he says,
“Feldman is an evidentialist. So am I. I don’t find myself frequently in dis-
agreement with him.”6
Part of Steup’s argument against opponents of doxastic freedom is
to accuse them of “practical reasons chauvinism.”7 Many who consider
reasons-responsiveness to be a necessary condition of freedom (and, thus,
responsibility) say beliefs are not appropriately reasons–responsive because
they are not responsive to practical reasons. Steup agrees that they are not
but thinks that no good argument is forthcoming for why responsiveness
to epistemic reasons cannot “equally ground freedom.” He says the view
that “voluntary control consists (exclusively) of responsiveness to practical
reasons” is “unmotivated and ad hoc.”8 Just as whether I take a walk or not
is responsive to practical reasons, “whether or not I believe cats are mammals
is a response to my epistemic reasons.”
Hieronymi also is clear that one can only believe for evidential or
epistemic reasons. Although she thinks there is a sense in which we can
exercise control over our beliefs, part of her explanation for what makes
beliefs different from voluntary actions appeals to something very much like
transparency, arguing that the only reasons that bear on whether or not to
believe something bear on the truth of what is believed:

In believing, you are answerable to reasons that you take to show the
belief is true . . . If we were to try to make belief into an action, one
would have to be, in believing answerable for reasons that one takes
to show something good about believing . . . Because believing brings
 The Possibility of Doxastic Agency 111
about its own distinctive form of answerability, believing p cannot be
understood as an action its own right.9

I have argued that deliberation about what to believe can include practical
reasons. I have further claimed that the kind of reasoning that leads to some
of our most important beliefs is not just centered on questions related to the
truth of the proposition, but also includes questions related to what would
be best to believe. While usually it is best to believe what is true, under cer-
tain conditions it can be best (or at least as good) to believe even when the
evidence does not support the truth of the proposition believed. The view of
doxastic agency and control, then, that I will be articulating is much more
continuous with how we think of agency and control in general. We have
seen that many accounts of doxastic agency, responsibility, or control think
of it as “odd” or “distinctive” because of what is thought to be an essential
difference between beliefs and actions. Given my view that the norms for
actions and beliefs are unified, I prefer, if possible, to develop a view of
responsibility and control that can be applied to both actions and beliefs.10
One of the reasons it may seem that we lack control over our beliefs is
that so many of them are unavoidable and irresistible. If it is impossible
for me to avoid holding a particular belief, how can I be said to have any
control over my holding it? For doesn’t being in control entail that I could
have done otherwise or could have chosen differently? Many theorists have
been concerned with formulating a concept of control (and responsibility)
that does not entail that one could have done otherwise. For if causal deter-
minism is true (or if God has created one and only one perfect world plan),
then there may be a sense in which we can never act other than we do, but
we would not want our notions of responsibility along with all the practices
that go along with them to be rendered meaningless if it turned out that we
did live in a deterministic universe. John Fischer calls the kind of control
that does not entail alternative possibilities “guidance control.” To illustrate
the kind of thing he has in mind, imagine that you are driving a car that is
a “driver instruction” automobile with dual controls. As long as you are
driving in a relatively safe manner, the instructor lets you control the car and
so when you, at the correct time, turn to the right it is you who is guiding the
car to the right. But if you had shown signs of confusion and were about to
mistakenly turn to the left, the instructor would have stepped in and steered
it to the right. Thus, you could have gone in no other direction but to the
right. So, although you have guidance control over the car, you lack what
Fischer calls “regulative control”; the instructor has that. Regulative control
entails a kind of dual power: the power to have control over some act A, and
the power to freely to do something else instead. The instructor can steer the
car any way he likes, while you can only steer it one way.
In their book, Fischer and Mark Ravizza11 provide detailed elaboration
of this concept of guidance control and argue that it is sufficient for moral
responsibility. I am not concerned with the details of their analysis or with
112 Doxastic Responsibilities
whether the kind of control we have over beliefs directly maps onto what
they have articulated, but much of their way of thinking about control seems
to be getting at what we can do in the doxastic realm. There has been exten-
sive criticism of this notion of guidance control, and I will not be concerned
with addressing all of these concerns. What I want to argue is that thinking
of doxastic control as a kind of guidance control is helpful, clarifying, and
preferable to other accounts given.12 It is being the author or guide of one’s
actions that Fischer and Ravizza argue is the relevant sense of control in
assessments of responsibility, not the principle of alternative possibilities.
There are two main components to this notion of guidance control: reasons-
responsiveness and ownership. We have seen that those who want to extend
responsibility beyond the voluntary tie responsibility to reasons-responsive-
ness and, thus, argue that responsibility does not require voluntary con-
trol. Fischer and Ravizza are adamant, however, that responsibility requires
control, just not the kind of “regulative control” that requires alternative
possibilities.13
An agent exhibits guidance control of an action “insofar as the mech-
anism which actually issues in his action is his own, reasons-responsive
mechanism.” The idea of ownership is left out of most of the character-
based and reasons-responsive accounts of doxastic responsibility but it is
essential in trying to understand how one can have guidance control over
beliefs. If all that were required for responsibility is that the mechanism
issuing in the action (or belief) is reasons-responsive, then even if you were
directly manipulated (say, had scientists kidnapped you and implanted such
a reasons-responsive mechanism), you would still be responsible.14 As we
have seen, one of the problems with these accounts is they do not allow for
mitigation or exceptions; the scope of responsibility it too wide. It seems
that those with what I have called “compelled beliefs” would be as respon-
sible a normal believers.15 On this view, one only ceases to be responsible if
one has lost the capacity to appreciate and respond to reasons. This would
mean that someone being treated for anorexia nervosa who can appreciate
the reasons to alter her belief about her weight but cannot yet respond to
those reasons is not excused from responsibility, that is unless her case is
completely hopeless.
For the mechanism that actually issues in certain behavior to be one’s
own, one must take responsibility for it. Taking responsibility is under-
stood historically. As one comes to view oneself as an agent—as having an
effect on the world as a consequence of one’s intentions, decisions, etc.—
one comes to view oneself as a fair target for the reactive attitudes, such
as being worthy of blame or praise. By viewing oneself as an appropriate
target for the consequence of a particular mechanism (say, ordinary prac-
tical reasoning), one thereby takes responsibility for it and the behavior
resulting from it. Once one takes responsibility for a particular mechanism,
then this ownership extends to future operations of the mechanism. It is a
process that occurs over time where we develop a concept of ourselves as
engaged in a kind of conversation.16 When we are addressed and treated as
 The Possibility of Doxastic Agency 113
responsible agents through such attitudes as praising and blaming, we begin
to form an internal view of ourselves as responsible, and develop our own
way of assessing and reacting to others. Fischer and Ravizza describe the
process as such: “The goal of achieving a correlation between external and
internal attitudes supports the practices that we use to train individuals who
are not yet full members of the moral community and to encourage them
to develop the internal view that we are extending to them.”17 Thus, taking
responsibility need not be any conscious act; rather, the way we react to oth-
ers and feel about ourselves reveal whether we have taken responsibility for
the mechanism in question.18
Can this notion of a mechanism be intelligibly applied to the doxastic
realm? Fischer and Ravizza switch from a focus on agents and their proper-
ties to a focus on mechanisms to help make clear how it is possible for an
agent to fail to be reasons-responsive while still being responsible. Part of
their motivation for developing this “mechanism-based” approach comes
from the challenge posed by “Frankfurt-type” cases.19 In such a case, an
agent carries out some behavior entirely according to his own deliberations
and reasons. For example, Sam carries out a plan to kill the mayor. But, if
somehow Sam had wavered, then Jack, who had installed a device that had
secretly been installed in his brain that monitors all of his brain activity,
would have activated the device to make Sam kill the mayor. In such a case,
the agent, Sam, is not reasons-responsive, but the actual-sequence mecha-
nism is. The alternative-sequence mechanism, the one that involves direct
stimulation of Sam’s brain, is not reasons-responsive and not one for which
Sam has taken responsibility.
Fischer and Ravizza are clear that all they mean by “mechanism” is the
process that leads to the relevant “upshot.” This upshot is some kind of
behavior, and by behavior they mean to include actions and omissions. They
also consider that the upshot may be a trait of character or, perhaps, an
emotional reaction. Given this analysis, it is plausible that one such upshot
can be a belief. Examples they give of mechanisms or processes are delib-
eration, practical reason, brain-stimulation, irresistible (physically based)
urges, hypnosis, addiction, and intentions. Given how broadly mechanisms
are construed, it seems perfectly legitimate to talk about the mechanisms
that result in beliefs. Some of the processes that result in beliefs are inquiry,
evidence gathering, attending, reasoning, memory, and perception. We can
now ask if any of these mechanisms are such that we can take responsibility
for them.
It is clear that the actual mechanism issuing in a belief is often reasons-
responsive. Say I believe, on inadequate evidence, that my brother stole
some money from me. I discover my error (perhaps by finding the money
or finding the real thief). I will then (under normal circumstances) revise my
belief about my brother. But do we take responsibility for these mechanisms
and for the resulting beliefs? Can we “own” them as a consequence of our
agency the way we can own the mechanisms that lead to actions? I think
we can see that this is possible if we think about our capacity to feel guilty
114 Doxastic Responsibilities
about simply having certain beliefs. So consider again the belief about my
brother. Even if I never acted on it in any way, I can still feel guilty for hav-
ing formed this belief at all. I am not suggesting that our feelings of guilt
are always appropriate or that feeling guilt is what marks off the domain
for what we are responsible. It is just one of many reactive attitudes, such
as blame and praise, that reveal that we find the initiator of the behavior
who provokes such reactions to be responsible; in the case of guilt or shame
the initiator is usually oneself. Just as is the case with action, these reactions
can sometimes be misplaced; but such attitudes are not always misplaced.
At times, guilt is an appropriate response to holding a belief. How can the
character-based view that denies that responsibility entails control account
for the appropriateness of feelings of guilt? According to this view, my guilt
does not result from a failure of control, but, instead, it results instead from
my feeling bad about not being a good person in general, given that good
people do not form beliefs too hastily.
But if I really viewed myself as lacking control over the belief, would I feel
the same kind of guilt for having it? Say I discovered that I had undergone
some kind of psychic manipulation such that the mechanism responsible for
issuing in beliefs about my brother was one which was controlled (via some
remote control) by some evil scientist. I might still feel some shame and
view myself as somehow defective, but part of the reason I feel guilty has
to do with this belief being a result of my agency. I have taken responsibil-
ity for the mechanisms that issue in evidentially based beliefs. When these
mechanisms are faulty, it is my fault, and I can be said to have lost some
control over these mechanisms given that I am failing to guide my beliefs
appropriately.20
That some notion of control is in play when assigning blame to beliefs
is reinforced if we consider when and why we mitigate such blame. If you
cannot make your higher order judgments effective about how you ought
to believe, there is a sense in which your belief is no longer your own;
you are divided and overpowered. I would blame you less if you really are
compelled to believe against your better judgment. You are not as open to
blame as someone who can believe the way he ought to but who fails to
put in the care and effort required to do so. Thinking about the account of
compelled belief discussed in Chapter 1, it seems we would be more apt to
excuse the person compelled to believe than we would the akratic believer.
Let us now look at a number of different cases of defective believing and
see where we are apt to assign responsibility and when we are not. It seems
that the more guidance control we have over the belief-issuing processes,
the more likely we are to attribute responsibility for it. This kind of control
comes in degrees and the more control we have, the less we are apt to miti-
gate responsibility.
Many of our beliefs result from perception. Perceptual beliefs also seem to be
the ones that are most obviously not under one’s control. But thinking again
about what it means to take responsibility and that this is properly seen as
 The Possibility of Doxastic Agency 115
a developmental process, we can see that even this most seemingly passive
mechanism is one over which we do have some degree of guidance control.
Imagine that someone is insistent that he sees a unicorn galloping toward
him and, based on the usual trustworthiness of his senses, believes there is
a unicorn galloping toward him. If this person later discovers that a power-
ful hallucinogenic drug had been slipped in his drink and he still believes
that a unicorn had galloped toward him, it is appropriate to criticize him.
For, normally, if I point out reasons for thinking that your normally well-
functioning mechanism has gone awry, you should revise your assessment
of the resulting upshot. Further, you see yourself as appropriately chastised
for being overly confident in, or hasty to assume, the proper functioning of
this mechanism.
Imagine a case of “misperception” that is not a result of an external agent
(like a hallucinogenic drug), where one is not hallucinating and has been
hypnotized. Bob believes his girlfriend was kissing Oscar in the car and he
believes this because he claims that he saw her doing so. It turns out she was,
in fact, not kissing Oscar and somehow Bob’s perception was faulty and
misled him into forming a faulty belief. If through pleading and insistence,
Bob’s girlfriend convinces him that she did not kiss Oscar, how should Bob
view the belief he formed? It is appropriate that he should feel some guilt
and, again, he should be criticized for his forming and maintaining it. How
much guilt or criticism is appropriate may depend on the details of the case.
If it would have looked liked kissing to all ordinary perceivers standing
where Bob was then, it would seem his initial formation is not blameworthy
but his maintaining it might be.
Of course, it is not possible for us to constantly monitor our perceptual
faculties to ensure that they are operating free of biases or neuroses that may
be leading us astray. But there is some presumption that we should be ready
to do some monitoring to ensure that this mechanism for which we have
taken responsibility is operating correctly. If I find this mechanism is regu-
larly leading me astray, something is wrong with me; it is not appropriate
for me to insist “but these beliefs result from perception over which I lack
control and so it is not my fault that I keep forming false beliefs.” If your
perceptual faculties really are “taken over” by some severe psychosis, a point
does come where we would excuse you from responsibility. But this again
underscores the difference from our assessment in the non-pathological case
where we think assessments of responsibility are appropriate and so some
degree of control is possible.
Which mechanism precisely am I claiming can both 1) be one for which
I can take responsibility and 2) is reasons-responsive? There is a wholly
passive aspect of perception that does not even satisfy the second crite-
rion. That I am experiencing the appearance of a unicorn while hallucinat-
ing and knowing I am hallucinating does not seems reasons-responsive;
no reasoning will lead me to revise the appearance. Perhaps a more vivid
example: the feeling that I am being stabbed as I am being stabbed is not
116 Doxastic Responsibilities
reasons-responsive. It is even hard to make sense of what it would mean
to be so. But the process whose upshot is simply an appearance and the
process whose upshot is a belief are different. Exactly what needs to be
added to perception such that what is perceived is also believed is a com-
plicated matter. For our purposes, we can just call the process perception
plus whatever x needs to be added (or Px). Pure perception and Px are
easily conflated because most of the time as soon as we perceive something
to be the case, we, at the same time, believe it to be so. This connection,
however, is severable and, further there are times when it ought to be.
When one fails to sever them when one ought one’s belief is appropriately
subject to criticism. And, again, I hold that part of what one is failing to
do in forming a belief one ought not is properly characterized as a failure
of control.
Another common belief-issuing mechanism is memory and, again, like
perception, this seems to be a process over which I exercise relatively lit-
tle control. If I have a vivid memory of an occurrence, it is almost impos-
sible for me to fail to believe that the thing occurred. But, as in the case of
perception, thinking about when and how this mechanism goes awry can
help show that the appropriateness of attributions of responsibility are
tied to our having some guidance control over this process. Imagine you
are conversing with a friend and you start talking about a conference you
were both at the previous summer. You begin recalling together who else
was at the conference. You say, “Oh, and John was there—I remember
liking his talk.” Your friend insists that John was not there and you are
emphatic that you remember him being there. If the next day your friend
shows you the program and convinces you that you misremembered, it is
appropriate for you will feel somewhat sheepish about your firm belief
that John was there and a certain degree of reproach is appropriate for
you having this belief.
One may wonder if you feel sheepish about your firm belief or about
your behavior. Perhaps the belief came about because you combined bits
of memories of one conference with bits of memories of another one into
a hybrid “memory” of a conference. And this “memory” may be very vivid.
But what can be wrong with firmly believing on the basis of a strong mem-
ory impression? If we recall Ginet’s example of believing I locked the door,
the problem was that the memory impression was unclear and not fully
confident. But given that your friend is insisting that John was not there,
based on her memory, it seems you should question the accuracy of your
vivacious memory. You may end up feeling sheepish about your own insis-
tent behavior, but you are also responsible and are appropriately criticized
for the belief. This would be the case even if you had not expressed your
view at all, if you had silently thought your friend is wrong because you
remember things correctly.
Contrast this case again with a pathological one. If someone has Alzheimer’s
disease and so has a severely defective memory mechanism, there comes a
 The Possibility of Doxastic Agency 117
point at which we excuse him from responsibility in his memory-induced
beliefs, and it seems to be at the point at which we acknowledge that he has
lost ownership of the process, where he is incapable of correction and of keep-
ing his beliefs in line with his and others’ judgments. Because such diseases
are often gradual, we can find that, at first, we do continue to react in ways
that reveal we hold the agent responsible for his beliefs. We will say, with
frustration, “Don’t you remember? You left the keys on your desk.” But as
the disease progresses, such admonishments seem less and less appropriate.
Just as lack of control over one’s faculties excuses one from being admonished
for how one acts, similarly it excuses one from being admonished for how one
believes. But, again, it is not a question of whether you could have believed
otherwise about John’s presence at the conference. Rather, this is a belief that
you have ownership over in that you have the capacity to keep it in line with
how you think you ought to believe.
One may think that once one discovers that his memory is impaired that
he has a special reason to be vigilant in its deliverances, and so is more
blamable than I am in my misremembering, given that I was reasonably
confident in my memory’s proper functioning. It may well be the case that
if one receives an early diagnosis when one has the capacity to exercise this
extra diligence, we may indeed hold him to an even higher standard than
those without impairment. When I said there comes a point when we would
no longer hold one responsible, I meant at the point when one has lost the
ability to take responsibility for one’s memory.
Attributions of responsibility are the most obviously warranted in
cases where beliefs result from deliberation or inquiry, and these are pro-
cesses over which we clearly have guidance control. These mechanisms are
responsible for issuing in actions as well as beliefs. One of the examples that
Fischer and Ravizza consider is taking responsibility for “acting from the
mechanism of practical reason.” They refer to these actions as “reflective
actions.” Beliefs that come about as a result of reflection are clearly ones for
which we have taken responsibility in Fischer and Ravizza’s sense. If I am
a juror and through a process of deliberation come to believe the witness
and acquit, if I am later convinced that there were good reasons to doubt
the witness, I will accept criticism and it seems criticism is appropriate.
Even if we maintain a distinction between practical reason that results in
action and theoretical reason that results in belief, a distinction that I have
been urging is not as clear-cut as often assumed, in both cases the process
involved is one of deliberation and inquiry, processes for which we take
responsibility.21
How, then, does this account of doxastic guidance control address the ini-
tial puzzle of doxastic responsibility and why is it preferable to the character-
based and reasons-responsive views discussed in the Chapter 5? Consider
again the example of the person who thinks the sun revolves around the
Earth. The character-based view argues that our criticism of one holding this
belief is in no one way tied to whether having the belief is in one’s control.
118 Doxastic Responsibilities
But we see that in cases where we lack guidance control over the belief-issuing
mechanism, attributions of responsibility and the reactive attitudes that come
with them cease to be appropriate. We expect a well-informed adult to be
sensitive to the amount of evidence that supports the heliocentric view and
expect that his belief will conform to this evidence. This expectation comes,
at least in part, from the fact that he has seen himself as a fair target for
being chastised in this way, and that he takes responsibility and ownership
for what he believes. These are all the necessary components of guidance
control. If one does not exhibit this kind of control, it would be misguided
to criticize him or otherwise hold him responsible. Thus, my account does a
better job on both the capacity to generalize and to allow for mitigation of
responsibility.
But isn’t saying that I am responsible for what flows from those aspects
of myself that I own very similar to saying I am responsible for what flows
from my character? Consider the example, which I argued is problematic
for the character-based account: the husband who has the misguided belief
that his wife is having an affair. It seems we could blame him for having this
belief, even if, in general, his character is such that he forms beliefs wisely.
I suggested that we would tend to blame him more if he were generally
wise, partly because we expect better of him; we expect him to be in better
control of his cognitive and emotional life. In the account I am proposing,
we are not blaming him for possessing a vice that reveals his defective
character. Rather, his history has displayed that activities like weighing evi-
dence and attending to arguments are processes for which he has taken
responsibility; he has accepted that it is appropriate to chastise him when
these processes send him off track. The appropriateness of our blame is tied
to the appropriateness of our expectations. In general, he has the capac-
ity to respond to reactive attitudes in the doxastic realm. If he lost this
general capacity, we would begin to question whether these processes and
the upshots of them really were his. The more questionable this connection
becomes, the more questionable become our attributions of responsibility.
The ownership account can do a better job explaining mitigation and
excusing bad beliefs. Remembering those cases where we are excused makes
it clear that responsibility is tied to control of the kind I have described.
The less our beliefs flow from ourselves, the less responsible we are. Once
we start to participate in the world as agents, we not only expect each
other to take responsibility for actions, but the responsibility extends to the
doxastic realm as well. How would the character-based view explain, for
example, the gradual diminishing of responsibility in the case of develop-
ing Alzheimer’s? Owens would say that responsibility disappears when the
person’s beliefs are no longer responsive to reasons. But what accounts for
this failure? For Owens, there is no meaningful question about whether you
ought to be responsive to reasons or not.
Owens has little to say about what excuses responsibility in cases like
this. In a discussion of memory and when it can serve as a reliable source of
 The Possibility of Doxastic Agency 119
evidence for rationally preserving belief, he says the following: “A person
with a mangled memory has an impaired capacity for reasoning; he has
no sensitivity to the past reasoning which could justify his current beliefs
and so can’t be held responsible for their lack of justification.”22 Here,
“mangled memory” refers to any time one’s memory “garbles the content”
of a reasonable belief. Owens’s point is that unless one has reason to doubt
one’s memory and to check its reliability, one is not responsible and so not
reproachable for continuing to hold the belief. So on Owens’s view, when
I mistakenly believed that John was at the conference last year I should
admit that my belief is groundless, but I should not be blamed for having
held it, as I did not reveal any vice in forming my belief. But, again, my
holding it does indicate some defect; I have done something wrong and, if
it is not the possessing of a vice then what is it? I submit that it is the lack-
ing a kind of self-control, a kind of carelessness. I am reproached, or feel
guilt, for the same reason I would if I acted in a way that manifests a lack
of self-control.
As we have seen, because Owens wants to preserve the notion that we are
responsible for beliefs and he is convinced that we have no reflective control
over beliefs, he argues that we should eschew what he calls “the juridical
theory of responsibility” that links responsibility to control. Given that his
concern is to explain how one can be responsible for involuntary states, he
never applies his theory to voluntary actions. In the realm of action, respon-
sibility is often diminished, as is legal responsibility, when the action is com-
pelled. For Owens, compulsion would be incidental; what must matter is
whether one is responsive to “ethical reasons” in carrying out the action. I
suppose a gun at your head will cause you to be less reasons-responsive, but
it seems more crucial that, in such a case, your agency has been overtaken.
Again, the reason that you cease to become responsive to ethical reasons
ought to matter.
The reasons-responsive accounts that eschew talk of character and virtue
do a better job with some of these worries, but because they are concerned
with divorcing responsibility from voluntary control, it is much harder for
them to explain how one can be more or less blameworthy for beliefs, or
how responsibility for beliefs can be diminished or augmented. Besides
being able to make better sense of the examples where we limit responsibil-
ity, the guidance control account allows for much more symmetry between
our responsibility for beliefs and our responsibility for actions than do those
accounts.
Finally, as discussed earlier, one important worry about having responsibility
based on character, moral personality, or one’s evaluative commitments, is
that these all seem contingent on factors wholly beyond one’s control, like
upbringing and environment, and at a certain point, are extremely difficult
to alter. Consider again Owens’s example of two agents who both become
possessed with unjustified anger. Given that for Owens such a state is blame-
worthy because it detracts from one’s goodness as a person, the genesis of the
120 Doxastic Responsibilities
state, as well as one’s own reaction to the state is irrelevant to our assessment
of it. We saw that Smith made a similar claim; history may matter for the
assessment of the kind of person you are but does nothing to mitigate respon-
sibility for the belief you now hold.
For the ownership account of responsibility, one’s history matters in terms
of assessing responsibility. Fischer and Ravizza suggest that the difference
between the cases in which we are inclined to hold an individual responsible
for an emotional reaction and those in which we are not so inclined may be
explained in terms of guidance control. If one’s upbringing was filled with
neglect and abuse to the point that one’s capacity for normal moral develop-
ment was undermined, this would make a difference in our assessment of moral
responsibility. And, if one tries very hard to gain control of one’s emotional life
and fails, blame would be severely mitigated. Given that taking responsibility
is understood as a developmental process, a point need not arrive when one
is beyond hope; one can always gain guidance control over one’s emotion-
issuing mechanisms and so become a fuller member of the moral community.
It is, thus, not only when one has completely lost the capacity to respond to
reasons that one ceases to be responsible, as was claimed by both Owens and
Smith.

2. RESPONSES TO OBJECTIONS

We have seen in Chapter 4 that Neil Levy is a recent advocate of the view
that we are not responsible for beliefs, and his way of arguing for this posi-
tion is to show that all accounts given of doxastic responsibility are flawed.
He uses the term “compatibilist” for accounts that articulate a concept of
doxastic responsibility that does not require having regulative control over
belief. This includes accounts that deny that responsibility entails any kind of
control and those that say responsibility entails guidance control. He begins
his section objecting to compatibilist accounts of doxastic responsibility by
mentioning Fischer and Ravizza, but neither of the two compatibilist views
he addresses refers directly to the idea of ownership and taking responsibil-
ity. The first one he considers, however, is closest to the account of control
I have given, and so I will assess whether his objection applies to my view. He
claims that this view will never allow one to have a blameworthy belief. The
account he considers, one argued for by Mark Heller,23 and which has a lot
in common with Smith’s view discussed in Chapter 5, says we are respon-
sible for beliefs if “we form and assess our beliefs in accordance with our
self-endorsed belief-forming dispositions.”24 I can thus be blamed whenever
I form beliefs in accordance with a bad reasoning process that I endorse.
Levy asks if such an account allows for the typical excusing conditions of
compulsion and ignorance. Compulsion does not seem to be a problem, for
it would seem that if my belief is compelled, then it is not formed in accor-
dance with my endorsed epistemic practices. If ignorance is to be an excuse,
 The Possibility of Doxastic Agency 121
however, it seems that every badly formed belief would be excused. For the
only way one would endorse a fallacious reasoning process would be out of
ignorance or irrationality. So it seems the only way one can be responsible is
if one believes well. Beliefs are either praiseworthy or the believer is excused.
There are two ways in which Levy’s criticisms do not apply to the
account given here. First, his description of belief-issuing mechanisms is
much narrower than what I have here described. We take responsibility
for more than just reasoning processes. If I am right that we can own, for
example, our perceptual faculties and the beliefs that result from them, then
our reproachable beliefs extend beyond those formed out of ignorance or
irrationality. Further, Levy’s talk of endorsement is a much more active and
conscious process than is the historic conception of ownership that Fischer
and Ravizza describe and which I have been employing. Even if we never
consciously endorse a mechanism, we can still have ownership of it. I can be
blamed for a badly formed belief if it results from a process for which I have
taken responsibility, and I have taken responsibility if my practices reveal
that I have accepted the expectation that I keep beliefs of this kind in line
with my higher order judgments of how I ought to believe. And, again we
normally accept being criticized for bad believing once we come to under-
stand how and why we were led astray.
Another worry with the ownership account is that on the one hand, it
can be too easy to duck responsibility by refusing to take responsibility
and, on the other hand, one can be held responsible when one ought not
to be because one has mistakenly taken responsibility. Should responsibility
really depend on attitudes of the fallible agent? Fischer and Ravizza discuss
this worry at length and I am satisfied with their response. First, we must
remember the dialogical and historical aspects of their account. As I empha-
sized in response to Levy’s objection, taking responsibility is not a single
act that one chooses to do or fails to choose to do. The price of failing to
take responsibility is high and not one that many people would be willing
to incur. In viewing oneself as an agent and as an “appropriate participant
in the family of reactive attitudes,” one thereby takes responsibility. If one
does not see oneself in such a way, one would be cut off from most mean-
ingful human relationships; it requires one “to relinquish autonomy and to
remain a fragmented self that is constantly in danger of ‘slipping away.’ ”25
There is, indeed, a “subjectivist” component to the ownership account in
that an agent has to have a certain view of himself to be responsible. But
one’s seeing oneself in such a way is not all that matters; one can mistakenly
see oneself as in control when one is not. If one were being directly manipu-
lated—for example by electronic stimulation—then one’s feeling like one
was an appropriate target for reactive attitudes would not thereby make one
responsible.26
Another question that arises with my account is whether the kind of
control I have described is any more direct than the kind of uncontroversial
indirect control that most everyone acknowledges. This is a difficult question
122 Doxastic Responsibilities
because the many people who address this issue each have a different view
about what kind of indirect control is uncontroversial. The important point
is that however they characterize the uncontroversial indirect control, they
claim that our attributions of doxastic responsibility do not depend on this
kind of control. For example, Richard Feldman says that we can have “non-
basic voluntary control” over belief because we can act in ways that will
cause beliefs. But the range of cases for which we attribute responsibility
is much larger than this small subset. Feldman thus says, “our ability to
control our beliefs in the way described here is epistemically insignificant,”27
and he thinks our deontological judgments about belief do not depend on
having this or any other kind of control.
We have seen that indirect control is also discussed in the context of
distinguishing between bringing about a belief and believing; I can act on
myself in various ways so that I will likely be in a state where a belief will
arise in me.28 This idea goes back to Pascal’s wager, where he suggests that
if one acts like a Christian, it is likely one will come to believe like a Chris-
tian. One has direct control over actions such as going to Mass and indirect
control over the beliefs that come about as a result of these actions. As we
have seen, it is a widely held view that we can act for practical reasons, but
we can only believe for truth-related reasons and so we could have reason
to engage in certain kinds of self-manipulation so that we are more likely to
view a certain proposition as true. In these cases of self-manipulation, if we
do attribute responsibility for the developed beliefs, it is because we attri-
bute responsibility for the actions that led to those beliefs; doxastic respon-
sibility is thus derivative. With the ownership view one’s responsibility for
beliefs is not derivative in this way. Your responsibility for your beliefs is
as direct as responsibility for many of your actions; in each case you are
responsible if the action or belief resulted from a mechanism for which you
have taken responsibility, in the sense I have described.

3. UNITY OF NORMS REVISITED

We have seen how allowing for the possibility of non-derivative doxastic

responsibility allows for symmetry between our responsibility for beliefs
and our responsibility for actions. Tying responsibility directly to control in
the doxastic realm reveals the possibility of doxastic agency. An important
implication of this view is that the norms governing belief are not isolated
from the norms governing other aspects of our agency. The central contention
argued for in Part I is that the tendency to isolate epistemic or doxastic norms
from practical or agential ones is misguided and that such a division is either
assumed or ill defended. We saw that one of the reasons for this tendency is
that it seems that the “ought” guiding belief and the “ought” guiding action
must be substantially different, because we can exercise control over actions
in a way that is substantially different from the control we exercise over belief.
 The Possibility of Doxastic Agency 123
Remember, in arguing that doxastic oughts should be seen as more like rules
of criticism and from rules of action, Chrisman says people “do not exercise
agency in believing what they believe.”29 Chrisman comes to this conclusion,
largely, because we do not seem to exercise the same kind of voluntary control
over beliefs as we do over many actions. If my account of guidance control is
correct, however, then we see that beliefs are products of our agency. We take
responsibility for our beliefs, and taking responsibility includes taking control
of them. We are blamed when we lose this grasp, when we do not exercise
our reflective competence that helps us believe the way we ought to believe.
Doxastic, moral, and prudential faults are all faults of agency.
Another concern that one may have with the account I have given is that
many examples of when we blame believers involve beliefs with moral con-
tent. When we blame someone for having a racist belief, for example, are we
really blaming him for any kind of having the belief, or is it because we closely
tie racist beliefs to immoral actions for which we blame him? When the con-
tent of a belief is rather insignificant, do we blame someone who formed it
hastily or on insufficient evidence? Included in Jonathan Adler’s examples of
blameworthy beliefs are “that whales are fishes” and “that some UFOs are
alien spaceships paying regular visits to Earth.”30 What do we think about
these beliefs? Would pity be more appropriate than blame for ignorance? Do
we react in any way that reveals that we hold them responsible? I think we
do; shaking your head in frustration reveals that you are passing judgment.31
Remember that when Clifford says, “It is wrong always, everywhere, and
for any one, to believe anything on insufficient evidence,” he means it is mor-
ally wrong.32 He thinks that the inculcated habits of forming beliefs on insuf-
ficient evidence could lead to dissolution of society and a violation of a duty
to humanity. As we have seen, most contemporary theorists resist this kind of
conflation, thinking of epistemic responsibilities as distinct from moral ones.
I have argued against this this separation of norms and so think that
Clifford is pointing to something important. It is not that epistemic faults
are moral faults. Rather, both moral faults and epistemic faults are pointing
to a fault in agency, revealing that you are becoming passive and unreflec-
tive where you should take active control. If this is true, then assessments
of epistemic responsibility are not only analogous to the moral, but, rather
these assessments are unified. That we correctly hold each other responsible
for the views we have about what is true, worthy, and good, reveals that we
can exercise control over these central aspects of who we are.

NOTES

1. Hieronymi, “Two Kinds of Agency,” 3.

2. To state (a) a little more carefully: Attributions of responsibility for belief are as
appropriate as are attributions of responsibility for actions. I will not provide
an argument that will convince someone who denies we have any freedom or
responsibility at all. My main contention is that, if we examine our practices,
124 Doxastic Responsibilities
the best way to make sense of them is to develop a plausible conception of
a kind of doxastic control. I am not trying to establish the metaphysical thesis
that we are free (either to act or to believe). It is possible that our attributions
of responsibility are radically mistaken or incoherent, but I intend to show
that we can articulate a kind of control that is possible in the doxastic realm
that would provide coherence to these attributions. The kind of argument
I will present in this chapter has much in common with what, following Kant,
have come to be called transcendental arguments. Beginning with the fact that
our attributions of doxastic responsibility are appropriate, I ask what are the
conditions of possibility such that this is the case. I argue that the coherence
of this central and prevalent practice depends on the possibility of doxastic
agency. I then seek to give a plausible account of doxastic agency that would
legitimate this common practice. Commonly, arguments of this form, take the
given to be something uncontroversial from which they deduce something less
than obvious. For example, Kant argues that sense-experience is only possible
if we possess pure intuitions of space and time. If his argument works, then
he has proven the existence of such pure intuitions because nobody would
question that we actually have sense-experience. But, as we have seen, one
can deny the appropriateness of our attributions of responsibility for belief.
The argument given for such skepticism, however, is that no plausible account
of the kind of control needed can be given. It seems preferable, if possible, to
articulate a way that can allow for a central part of our mental and social lives
to not be radically mistaken.
3. Not everyone agrees that is possible to act against one’s all-things-considered
best judgment because an intentional action is often taken to result from such
judgment. Thus, skepticism about weak willed (akratic) action is not uncom-
mon. For some discussion on if or how akrasia is possible, see Davidson,
“How is Weakness of Will Possible?” 21–42; Mele, Irrationality; Bratman,
“Practical Reasoning and Weakness of the Will,” 153–171; Watson, “Skep-
ticism about Weakness of Will,” 316–39; and Buss, “Weakness of Will,”
13–44.
4. Frankfurt, “Freedom of the Will,” 5–20. For Frankfurt, “the notion of the will
is not coextensive with the notion of what an agent intends to do. For even
though someone may have a settled intention to do X, he may nevertheless do
something else instead of doing X because, despite his intention, his desire to
do X proves to be weaker or less effective than a conflicting desire” (8). Those
who identify will and intention more closely may question how settled his
intention was given his failure to act.
5. McHugh, “Exercising Doxastic Freedom,” has a useful discussion of vari-
ous ways of understanding “what is involved in acting in accordance with,
or guided by the will” (4). He concludes that an action is under control of
the will, or, equivalently, under voluntary control, if it “meets the conditions
of reasons-reactivity and of intentions-reactivity. Voluntariness is a matter of
exercising a capacity to select and execute what you do on the basis of what
you see or judge yourself to have most reason to do” (9). Because beliefs
are not reactive to our intentions or reasons the way actions are, McHugh
concludes, “we do not have voluntary control over our doxastic states, Our
doxastic states are not under the command of the will in the way that our
actions are” (10). He goes on to offer a way of understanding doxastic free-
dom built on the notion of reasons-receptivity which he adopts from Fischer
and Ravizza’s account of moral responsibility.
6. Steup, “Doxastic Freedom,” 382.
7. Ibid., 387.
8. Ibid., 388.
 The Possibility of Doxastic Agency 125
9. Hieronymi, “Responsibility for Believing,” 365–366. As we have seen, there
are some exceptions to this view. In discussing the possibility of rationally
believing against the evidence, Marusic says, “when is it is up do us to
do something, we can rationally believe that we will do it even if we have
evidence that there is a significant chance we won’t do it, provided that it is
practically rational for us to do it; we can rationally believe against the evi-
dence” (“Promising Against the Evidence,” 11). While Marusic might resist
saying that in such cases I have a practical reason to believe, for example,
that I will keep my promise, if it is rational to keep my promise then, he
contends, the rationality of the action also makes the belief rational. For
other defenses of the possibility of believing for practical reasons, see Jordan,
“Pragmatic Arguments and Belief,” 409–420; and Foley, Working Without a
Net. An interesting recent discussion of when it is permissible to believe for
certain class of non-evidential reasons is found in Talbot’s “Truth Promoting
Non-evidential Reasons for Belief,” 599–618.
10. McHugh, “Exercising Doxastic Freedom,” points to a similar problem with
some defenses of doxastic freedom. He says “they propose criteria for doxastic
freedom quite different to the criteria for freedom of action, without showing
that what they are giving an account of is really a kind of freedom” (11). While
he thinks doxastic freedom is exercised differently from freedom of action, he
says that these species of freedom are of the same genus; “The more general
condition on freedom, that covers the various species of freedom I have dis-
cussed, in a condition of reasons-responsiveness” (31). But again, beliefs are
only responsive to reasons that are truth-related. I think that as long as one
maintains such a strict division between the nature and norms of belief and
action that the kind of freedom and agency that applies to belief bears little
resemblance to the freedom and agency exercised in action. To say reasons-
responsiveness unifies them is misleading if the ways in which they are reasons-
responsive is so much more constricted in one realm than in the other. I am
in danger of being subject to McHugh’s other criticism of existing defenses
of doxastic freedom, namely of claiming “that doxastic states are free in just
the same way that actions are free, and thus [failing] to do justice to the fact
that we don’t control our doxastic states as we do our actions” (11). I will say
a little more about this concern in the Conclusion of this book. What I have
been urging, and will continue to argue in this chapter, is that there is not as
much asymmetry between the kind of control we have over our beliefs and our
actions as many theorists contend even if we cannot control most of our beliefs
the way we can control some of our actions.
11. Fischer and Ravizza, Responsibility and Control.
12. Though, as we will see below, while Fischer and Ravizza briefly consider
how their account might be extended to emotions, they do not consider how
it might be extended to beliefs. When they discuss beliefs in passing, they
talk about them as being non-voluntary and not “up to an agent” (217) and
suggest that we find it less problematic if our beliefs are causally determined
by the external world than if our desires, choices and intentions are (236).
13. Fischer and Ravizza, Responsibility and Control, 39.
14. Steup, “Doxastic Freedom,” makes a similar point when considering
“reasons-responsiveness compatibilism.” If reasons-responsiveness could
result from systematic conditioning or manipulation (or even brainwashing),
reasons-responsiveness is not sufficient for freedom. Although Steup cites
Fischer and Ravizza as supporting this view, it is crucial that reasons-
responsiveness is a necessary but not sufficient condition, 39.
15. McHugh, “Doxastic Agency.” McHugh’s account of doxastic agency also
draws on Fischer and Ravizza’s account of guidance control and shows how
126 Doxastic Responsibilities
their idea of ownership can make better sense of why one is not responsible
for paranoid delusions or compulsive and repressed beliefs than does a view
like Smith’s (see especially pages 141–144).
16. It is important in what follows to keep in mind this special sense of “tak-
ing responsibility.” It may seem odd that one necessary condition of being
responsible is that one has taken responsibility. Why should my attitudes or
reactions to what I have done have any bearing on whether I am actually
responsible? But taking responsibility and thus having ownership for what
one has taken responsibility is something revealed in one’s practices over
time. I will return to this concern when discussing objections to my view.
17. Fischer and Ravizza, Responsibility and Control, 212.
18. Philip Petit and Michael Smith make a similar point about this in their
“Freedom in Belief and Desire,” where they discuss the “conversational
stance.” We are only able to engage in intellectual conversation if we view
each other as satisfying these three conditions: “first there are norms rele-
vant to the issue of what [we] ought to believe; second, [we are] capable of
recognizing this to be so and third, [we are] capable of responding appropri-
ately to the norms: that is, capable of believing the way [we] should” (Wat-
son, Free Will, 392).
19. Harry Frankfurt famously introduced examples which seem to show that we
can be responsible even when there is no alternative open to us, thus undermin-
ing the principle of alternative possibilities (P.A.P.). See Frankfurt’s “Alternative
Possibilities and Moral Responsibility,” 829–39; and “Freedom of the Will,”
5–20.
20. Which reasons-responsive mechanism precisely have I taken responsibility
for in the case here described? I have not taken responsibility for a faulty
mechanism. I identified a number of processes above that have beliefs as
their outputs and it seems my belief about my brother would result from
a combination of those and, in what follows, I will show that these are all
processes over which it makes sense to say one has ownership.
21. In a sense, one of the central contentions of this book is that a strict distinction
of practical and theoretical reasoning is untenable. I hope to have shown that
deliberating about what to believe and deliberating about what to do are not
radically different in kind. I will elaborate on this idea in the Conclusion of this
book.
22. Owens, Reason without Freedom, 159.
23. In Heller’s “Hobartian Voluntarism,” 130–141.
24. Levy, “Doxastic Responsibility,” 140.
25. Fischer and Ravizza, Responsibility and Control, 220.
26. Ibid., 236.
27. Feldman, “The Ethics of Belief,” 272.
28. Others who allow for indirect control but who do not think it is significant for
doxastic responsibility are Alston, “The Deontological Conception of Epistemic
Justification”; Adler, “Belief’s Own Ethics”; and Levy, “Doxastic Responsibil-
ity.” See also, Robert Audi “Doxastic Voluntarism and the Ethics of Belief,”
93–111.
29. Chrisman, “Ought to Believe,” 369.
30. Adler, Belief’s Own Ethics, 64.
31. Does “shaking your head in frustration” constitute a “reactive attitude”?
It seems to reveal a kind of lack of respect, or even contempt, which does
seem to be such an attitude.
32. Clifford, “The Ethics of Belief,” 22. This comes from the passage discussed
in Chapter 3, where he talks of our “universal duty of questioning all that
we believe.”
 The Possibility of Doxastic Agency 127
REFERENCES

Adler, Jonathan. Belief’s Own Ethics. Cambridge: MIT Press, 2002.

Alston, William P. “The Deontological Conception of Epistemic Justification.”
Philosophical Perspectives 2 (1988): 257–299.
Audi, Robert. “Doxastic Voluntarism and the Ethics of Belief.” In Knowledge,
Truth, and Duty, edited by Matthias Steup, 94–109. Oxford: Oxford University
Press, 2001.
Bratman, Michael. “Practical Reasoning and Weakness of the Will.” Nous 13,
no. 2 (1979): 153–171.
Buss, Sarah. “Weakness of Will.” Pacific Philosophical Quarterly 78 (1997): 13–44.
Chrisman, Matthew. “Ought to Believe.” The Journal of Philosophy 105, no. 7
(2008): 346–370.
Clifford, William K. “The Ethics of Belief.” In The Ethics of Belief Debate, edited by
Gerald D. McCarthy, 19–36. Atlanta: Scholars Press, 1987.
Davidson, Donald. “How Is Weakness of Will Possible?” In Actions and Events,
21–42. Oxford: Clarendon Press, 1980.
Feldman, Richard. “The Ethics of Belief.” Philosophy and Phenomenological
Research 60, no. 3 (2000): 667–695.
Fischer, John, and Mark Ravizza. Responsibility and Control: A Theory of Moral
Responsibility. Cambridge: Cambridge University Press, 1998.
Foley, Richard. Working Without a Net: A Study of Egocentric Epistemology.
Oxford: Oxford University Press, 1993.
Frankfurt, Harry. “Alternative Possibilities and Moral Responsibility.” Journal of
Philosophy 66 (1969): 829–839.
———. “Freedom of the Will and the Concept of a Person.” The Journal of Philoso-
phy 68, no. 1 (1971): 5–20.
Heller, Mark. “Hobartian Voluntarism: Grounding a Deontological Conception
of Epistemological Justification.” Pacific Philosophical Quarterly 81 (2000):
130–141.
Hieronymi, Pamela. “Responsibility for Believing.” Synthese 161, no. 3 (2008):
357–373.
———. “Two Kinds of Agency.” In Mental Actions, edited by Lucy O’Brien and
Matthew Soteriou, 1–30. Oxford: Oxford University Press, 2009.
Jordan, Jeff. “Pragmatic Arguments and Belief.” American Philosophical Quarterly
33, no. 4 (1996): 409–420.
Levy, Neil. “Doxastic Responsibility.” Synthese 155 (2007): 127–155.
Marusic, Berislav. “Promising Against the Evidence.” Ethics 123, no. 2 (2013):
292–317.
McHugh, Conor. “Exercising Doxastic Freedom.” Philosophy and Phenomenologi-
cal Research 88 (2014): 1–37.
Mele, Alfred. Irrationality: An Essay on Akrasia, Self-Deception, and Self-Control.
New York: Oxford University Press, 1987.
Owens, David. Reason without Freedom: The Problem of Epistemic Normativity.
London: Routledge, 2000.
Petit, Philip, and Michael Smith. “Freedom in Belief and Desire.” Journal of Philoso-
phy 93 (1996): 429–449.
Steup, Matthias. “Doxastic Freedom.” Synthese 161 (2008): 375–392.
Talbot, Brian. “Truth Promoting Non-evidential Reasons for Belief.” Philosophical
Studies 168 (2014): 599–618.
Watson, Gary. “Skepticism About Weakness of Will.” Philosophical Review 86
(1977): 316–339.
Watson, Gary, ed. Free Will. Oxford: Oxford University Press, 2003.
This page intentionally left blank
Conclusion

In one way, the claims I have argued for are very modest. I think that in
most cases, we should not deviate from evidentialist principles. Most of the
time, if one believes against the evidence, one is doing something wrong.
Thus, my evaluations of particular beliefs rarely differ from the evidential-
ist’s evaluations. In this sense, my anti-evidentialism is modest. In another
way, the claims I have argued for are new and controversial. I think that
the reason one is blamed or criticized for holding a particular belief is no
different in kind from the reason one is blamed or criticized for performing
a particular action. The ethics of belief is inseparable from ethics simpliciter.
The same kinds of considerations that bear on investigating what we should
do will also help illuminate how we should believe.
Most of the time, the answer to such questions is obvious; believing in
accordance with the evidence will be the way to have the best beliefs one
can—the beliefs that are the most helpful to oneself and others, the beliefs
that reflect the kind of person one wants to be. But most of our actions also
require little assessment or deliberation; we often operate almost automati-
cally, and we often manage not to violate the rules of prudence or morality.
It is when the right course to take is not obvious that deliberation comes in.
Again, determining what to believe is just like determining how to act.
Cases where external factors do not determine what to believe often center
around issues that matter deeply to us. Say that I have no way of determin-
ing whether my ex knowingly deceived me, but I know that believing he did
so will feed my antagonism, worsening our current relationship and hurting
our kids. If I take these factors into account when trying to figure out what
to believe, I am thinking about what it would be best for me to believe.
Thus, I am considering factors that are not solely truth related.
That practical reasoning can have this effect on beliefs shows that we have
some control over what to believe. But the idea that anything like choice or
control applies in the doxastic realm is very contentious. The argument of
Part I of the book is strengthened if it is possible to make sense of doxastic
agency. Thus, in a sense, the two parts of the book reach the same conclu-
sion via different paths. The conclusion that the norms of agency unify the
doxastic and practical realms demands that we can make sense of doxastic
130 Conclusion
agency. The notion that we can exercise control in the doxastic realm sug-
gests that unified norms guide both belief and action.
Pragmatism is often associated with doxastic voluntarism. We saw, in
Chapter 1, that one common line of defense for evidentialism is that the
evidentialist thesis explains why we are unable to choose what to believe. We
cannot choose what is true, and if beliefs in some sense aim at or are governed
by or are conceptually tied to truth, then this shows why we cannot choose to
believe. One way to counter this line of defense is to deny that the phenom-
enon it invokes is genuine. So, doxastic voluntarism supports pragmatism.
Nevertheless, many philosophers who endorse doxastic agency or doxastic
freedom are evidentialists. In fact, many such philosophers do not think it
is even possible to believe for practical reasons, let alone rational to do so.1
People who argue that we are responsible for our beliefs and have some
kind of control over them emphasize the similarities between belief and
action. And yet if the norms that apply to beliefs were wholly separate from
those that apply to actions, then beliefs and actions would differ in a crucial
way: doxastic responsibility would be a different kind of responsibility than
the kind we attribute to actions. The kind of failure that leads to blame in
one realm would be wholly different than the kind of failure that leads to
blame for action. Believing badly would be more like skiing badly or being
a poor chess player, something that is assessed according to its own distinct
set of norms. As we have seen, viewing beliefs in such a way is deeply prob-
lematic. I have shown how we can avoid viewing beliefs in this way and
develop a unified conception of ourselves, as both thinkers and doers, by
accepting that the norms of belief and action are unified.
In Chapter 6, I discussed a criticism of my view: that I have failed to
pay attention to the distinction between practical and theoretical reasoning.
I accept that it is a consequence of my view that there is no real distinction
between practical and theoretical reasoning. But I think this is an advantage
of my view. Part of what has led philosophers to a misguided view of the
norms governing belief is the misguided assumption that practical and theo-
retical reasoning are radically different.
Why does it seem so obvious to philosophers that there is a distinction
between practical and theoretical reasoning? Well, it seems they are used for
different purposes, to answer different kinds of questions. Theoretical rea-
son is used to address questions concerning what is true, whereas practical
reason is used to address questions about what is good.
I accept that we reason about different kinds of issues. But I do not see
why we need to posit two different kinds of reasoning. Reasoning about
theoretical issues and practical issues are often overlapping and intertwined.
To figure out what is best to do, we often must have a correct grasp of the
facts. If I deliberate about whether to stay home and grade or go out to see
a friend’s band, it is helpful if I know, for example, if and when the band will
play again, and what will happen if I put off my grading. Even Hume, who
thought reason could only do the job of discovering facts and making infer-
ences, knew that such reasoning had a role to play in our practical judgments
 Conclusion 131
While the role of theoretical reasoning in deliberating about what to do
is generally acknowledged, the role of practical reasoning in deliberating
about what to believe is not. I think this is because it is generally accepted
that knowing the truth can help you determine what’s good, but knowing the
good cannot help you determine what’s true. Depending on how one unpacks
this slogan, both clauses of this compound sentence turn out false. Even a
minimal acceptance of the is/ought distinction will lead to the denial that the
facts (what is) can determine what is good (or what ought to be). So, it is
only once certain normative principles are accepted that the facts can help
in one’s practical determinations. And there will be times when knowing all
the facts will be no help at all in determining what to do. The most difficult
moral dilemmas arise at such times, times when it seems either course of
action will be wrong.2 The attraction of strict and simple act utilitarianism is
that it does, in principle, allow the facts to dictate what is right. If one could
determine that overall suffering would be reduced by torturing one innocent
person, then the torture is permissible; there is nothing else to discern and
no more questions to ask. Of course, one’s epistemic position is rarely strong
enough to determine what the consequences of any action will be; this is one
of the standard objections to act utilitarianism. But another common, and
perhaps deeper, dissatisfaction with this kind of view, is that there should
be more questions to ask even if all the facts could be determined and that,
at times, there might not even be a correct answer about what is best to do.
These considerations reveal that the role of facts, and the theoretical rea-
soning that helps in their discovery, in determining what is best to do or believe
is both complex and limited. The same holds for the other half of the slogan,
namely that knowing the good cannot help you determine what is true. There
are times when questions about what is good do bear on the questions of what
is true, especially when part of whether something turns out true depends on
one’s own actions. Whether it is true that I will keep my promise or follow
through on a commitment is affected by my viewing it as good to do so; here,
questions about what is good help me determine what to believe.3
Questions about what is good also help determine what to believe about
the significance or meaning of human existence and endeavors. If one looks
starkly at the facts of our world and the history of humanity, the horror
is overwhelming. I doubt that if one were to make a list of all the acts of
cruelty and all the acts of selfless kindness ever performed that the kind acts
would outnumber the cruel ones. But believing that humans have a greater
capacity for love and kindness than for evil provides some hope, some rea-
son for optimism. Thus, it’s a good thing to believe. And that it is a good
thing to believe has some part to play in why I believe it.
Suppose you accept my claim that agency is exercised in the doxastic realm.
You might still ask what’s involved in assessing someone’s agency as a whole,
instead of simply assessing each dimension of agency. Many philosophers
who agree with me that doxastic agency is possible do not share my belief
that we can meaningfully assess agency as a whole. Such philosophers think
that one can be an excellent doxastic agent but terrible at making practical
132 Conclusion
decisions or completely immoral. One of my central arguments against phi-
losophers who deny doxastic agency is that such a picture fails to capture the
interwoven and overlapping nature of our practical and doxastic lives. The
same argument applies to philosophers who think that we can be assessed as
moral agents or epistemic agents, but not simply as agents.
Here, it may be helpful to think about how athletic excellence is assessed.
If you ask what makes someone an excellent hockey player, we can point to
various skills in which he excels: speed, stick handling, agility, shooting, etc.
Someone cannot be an excellent player if he is excellent at one of these but
poor at another; one must be excellent at all of them. But when someone
is assessed as an excellent hockey player, this evaluation goes beyond how
good he is at each aspect of hockey playing. The people considered the great-
est players are not necessarily the ones who score the highest in each of these
specific dimensions. Similarly, if you are an excellent agent, you must do at
least okay on all of the specific dimensions of agency. But there is more to
being an excellent agent than being excellent at each dimension. Explaining
what more there is, is difficult, just as it is difficult to say what makes some-
one a hockey great over and above their speed, puck-handling, and the like.
I conceive of the norms of agency as broadly eudaimonistic. In my dis-
cussion of epistemic virtue, we saw that one way philosophers argue for the
non-instrumental value of truth and knowledge is by claiming that they are
constitutive of a good life. Some argue that we should understand virtues,
in general, as being those traits that invariably coincide with a flourishing
life.4 I think the hopes of trying to establish such an invariable connection
between virtue and flourishing are dim if one takes “flourishing” as apply-
ing only to an individual. But if flourishing is thought of as applied both
individually and collectively, then the view is more plausible. The norms
of agency, it seems are those which, if everyone followed them, everyone
would be better off. Believing according to rules that help avoid falsehoods
and acting according to rules that help reduce suffering will contribute to
both individual and collective flourishing. But in the case of both belief and
action, there will be times when the general rules do not apply, times when
one needs to exercise one’s agency in a nonstandard, creative way, recogniz-
ing that what is best to do or believe is not what these rules dictate. It may
then be, finally, that agential norms cannot be codified but that violating
what is usually thought of as an epistemic norm (like believe only what
is true) or a practical norm (like do not harm others) need not violate the
norms of agency. Maybe part of what it takes to be an excellent agent is
recognizing when the rules can be broken.

NOTES

1. Steup, “Doxastic Freedom”; Hieronymi, “Responsibility for Believing.”

2. Nagel, “War and Massacre,” 123–144, vividly conveys how, at times, what-
ever ones does, that action will be morally wrong.
 Conclusion 133
3. Marusic, “Belief and Difficult Action.”
4. In the first chapter to Gopal Sreenivasan’s Emotion and Virtue (Princeton,
forthcoming), he considers various ways that virtue theorists might try to
defend this claim and finds them all unconvincing.

REFERENCES

Hieronymi, Pamela. “Believing at Will.” In Belief and Agency, edited by David

Hunter. The Canadian Journal of Philosophy Supplementary Volume 35 (2009):
149–187.
Hieronymi, Pamela. “Responsibility for Believing.” Synthese 161, no. 3 (2008):
357–73.
Marusic, Berislav. “Belief and Difficult Action.” Philosopher’s Imprint 12, no. 18
(2012): 1–30.
Nagel, Thomas. “War and Massacre.” Philosophy & Public Affairs 1, no. 2 (1972):
123–144.
Sreenivasan, Gopal. Emotion and Virtue (forthcoming). Princeton: Princeton
University Press.
Steup, Matthias. “Doxastic Freedom.” Synthese 161 (2008): 375–392.
This page intentionally left blank
Bibliography

Adler, Jonathan. “Akratic Believing?” Philosophical Studies 110, no.1 (2002): 1–27.
———. Belief ’s Own Ethics. Cambridge: MIT Press, 2002.
Allan, G. Lorraine, Shepard Siegel, and Samuel Hannah. “The Sad Truth about
Depressive Realism.” The Quarterly Journal of Experimental Psychology 60, no. 3
(2007): 482–495.
Alloy, L. B. and L. Y. Abramson. “Depressive Realism: Four Theoretical Perspec-
tives.” In Cognitive Processes in Depression, edited by L. B. Alloy. New York:
Guilford Press, 1988.
———. “Judgment of Contingency in Depressed and Nondepressed Students: Sadder
but Wiser?” Journal of Experimental Psychology: General 108 (1979): 41–485.
Alston, William P. “The Deontological Conception of Epistemic Justification” Philo-
sophical Perspectives 2 (1988): 257–299.
Anscombe, G.E.M. Intention. Oxford: Basil Blackwell, 1957.
Aristotle. Nicomachean Ethics. Translated by Terence Irwin. Indianapolis: Hackett,
1999.
Audi, Robert. “Doxastic Voluntarism and the Ethics of Belief.” In Knowledge,
Truth, and Duty, edited by Matthias Steup, 93–108. Oxford: Oxford University
Press, 2001.
Avnur, Yuval. “In Defense of Secular Belief.” Oxford Studies in Philosophy of Reli-
gion Vol. IV, edited by Jonathan Kvanvig, 2–20. Oxford: Clarendon Press, 2011.
Bennett, Jonathan. “Why Belief Is Involuntary.” Analysis 50 (1990): 87–107.
Bering, Jesse. “The Existential Theory of Mind.” Review of General Psychology 6,
no. 1 (2002): 3–24.
Berker, Selim. “Epistemic Teleology and the Separateness of Propositions.” Philo-
sophical Review 122, no. 3 (2013): 337–393.
Bishop, John. Believing by Faith. Oxford: Oxford University Press, 2007.
———. “How a Modest Fideism May Constrain Theistic Commitments: Exploring
an Alternative to Classical Theism.” Philosophia 35 (2007): 387–402.
BonJour, Laurence. The Structure of Empirical Knowledge. Cambridge: Harvard
University Press, 1985.
Boyle, Matthew. “‘Making Up Your Mind’ and the Activity of Reason.” Philoso-
phers’ Imprint 11, no. 16 (2011): 1–24.
Bratman, Michael. “Practical Reasoning and Acceptance in a Context.” Mind 101,
no. 401 (1992): 1–13.
———. “Practical Reasoning and Weakness of the Will.” Nous 13, no. 2 (1979):
153–171.
Buss, Sarah. “Weakness of Will.” Pacific Philosophical Quarterly 78 (1997): 13–44.
Chrisman, Matthew. “Ought to Believe.” The Journal of Philosophy 105, no. 7
(2008): 346–370.
136 Bibliography
Clifford, William K. “The Ethics of Belief.” In The Ethics of Belief Debate, edited by
Gerald D. McCarthy, 19–36. Atlanta: Scholars Press, 1987.
Cohen, Jonathan L. An Essay on Belief and Acceptance. Oxford: Oxford University
Press, 1992.
Davidson, Donald. “How Is Weakness of Will Possible?” In Actions and Events,
21–42. Oxford: Clarendon Press, 1980.
Descartes, Renee. Meditations on First Philosophy. Translated by David Cress, 3rd
ed. Indianapolis/Cambridge: Hackett Publishing Company, 1993.
Engel, Pascal. “Epistemic Responsibility Without Epistemic Agency.” Philosophical
Explorations 12, no. 2 (2009): 205–219.
Falkenstein, Lorne. “Naturalism, Normativity, and Skepticism in Hume’s Account
of Belief.” Hume Studies 23 (1997): 29–72.
Feldman, Richard. “The Ethics of Belief.” Philosophy and Phenomenological
Research 60, no. 3 (2000): 667–695.
Fischer, John, and Mark Ravizza. Responsibility and Control: A Theory of Moral
Responsibility. Cambridge: Cambridge University Press, 1998.
Foley, Richard. Working Without a Net: A Study of Egocentric Epistemology.
Oxford: Oxford University Press, 1993.
Frankfurt, Harry. “Alternative Possibilities and Moral Responsibility.” Journal of
Philosophy 66 (1969): 829–839.
———. “Freedom of the Will and the Concept of a Person.” The Journal of Philoso-
phy 68, no. 1 (1971): 5–20.
Gendler, Tamar. “Alief and Belief.” The Journal of Philosophy 105, no. 10 (2008):
634–663.
———. “Alief in Action (and Reaction).” Mind and Language 23 (2008): 552–585.
Gibbons, John. The Norm of Belief. Oxford: Oxford University Press, 2013.
Ginet, Carl. “Deciding to Believe.” In Knowledge, Truth, and Duty: Essays on Epis-
temic Justification, Responsibility, and Virtue, edited by Matthias Steup, 63–76.
New York: Oxford University Press, 2001.
Greco, John. Achieving Knowledge: A Virtue-Theoretic Account of Epistemic Nor-
mativity. Cambridge: Cambridge University Press, 2010.
Grimm, Stephen. “Epistemic Normativity.” In Epistemic Value, edited by Adrian
Haddock, Alan Millar, and Duncan Pritchard, 243–264. Oxford: Oxford Uni-
versity Press, 2009.
Heil, John. “Doxastic Incontinence.” Mind 93 (1984): 56–70.
Heller, Mark. “Hobartian Voluntarism: Grounding a Deontological Conception
of Epistemological Justification.” Pacific Philosophical Quarterly 81 (2000):
130–141.
Hieronymi, Pamela. An Enquiry Concerning Human Understanding. Edited by Tom
L. Beauchamp. Oxford: Oxford University Press, 1999.
———. “Believing at Will.” In Belief and Agency, edited by David Hunter. The
Canadian Journal of Philosophy Supplementary Volume 35, (2009): 149–187.
———. Enquiries Concerning Human Understanding and Concerning the Principles
of Morals, edited by L. A. Selby-Bigge and P. H. Nidditch. Oxford: Clarendon
Press, 1975.
———. “Responsibility for Believing.” Synthese 161, no. 3 (2008): 357–373.
———. “Two Kinds of Agency.” In Mental Actions, edited by Lucy O’Brien and
Matthew Soteriou, 1–30. Oxford: Oxford University Press, 2009.
Hume, David. A Treatise of Human Nature, edited by L. A. Selby-Bigge and P. H.
Nidditch, 2nd ed. Oxford: Clarendon Press, 1978.
Hume, Enquiries Concerning Human Understanding and Concerning the Principles
of Morals. Edited by L.A. Selby-Bigge and P.H Nidditch. Oxford: Clarendon
Press, 1975.
 Bibliography 137
James, William. “The Will to Believe.” In The Ethics of Belief Debate, edited by
Gerald D. McCarthy, 55–71. Atlanta: Scholars Press, 1987.
Jordan, Jeff. “Pragmatic Arguments and Belief.” American Philosophical Quarterly
33, no. 4 (1996): 409–420.
Kassin, S. M., and Studebaker, C. A. “Instructions to Disregard and the Jury:
Curative and Paradoxical Effects.” In Intentional Forgetting: Interdisciplinary
Approaches, edited by J. M. Golding and C. M. MacLeod, 413–434. Hillsdale,
NJ: Erlbaum, 1998.
Kornblith, Hilary. “Epistemic Normativity,” Synthese 94, no. 3 (1993): 367.
———. Knowledge and Its Place in Nature. Oxford: Oxford University Press, 2002.
Levy, Neil. “Doxastic Responsibility.” Synthese 155 (2007): 127–155.
Lynch, Michael. True to Life: Why Truth Matters. Cambridge: MIT Press, 2005.
Marusic, Berislav. “Belief and Difficult Action.” Philosopher’s Imprint 12, no. 18
(2012): 1–30.
———. “Promising against the Evidence.” Ethics 123, no. 2 (2013): 292–317.
Marusic, Jennifer Smalligan. “Does Hume Hold a Dispositional Account of Belief.”
Canadian Journal of Philosophy 40, no. 2 (2010): 155–184.
McCormick, Miriam. “Compelled Belief.” American Philosophical Quarterly 42,
no. 3 (2005): 157–169.
———. “Why Should I Be Wise?” Hume Studies 31, no. 1 (2005): 3–19.
McDowell, John. Mind and World. Cambridge: Harvard University Press, 1994.
McGrath, Matthew. “Dogmatism, Underminers, and Skepticism” (forthcoming). In
Philosophical and Phenomenological Research.
McHugh, Conor. “Epistemic Responsibility and Doxastic Agency.” Philosophical
Issues 23, (2013): 137–157.
———. “Exercising Doxastic Freedom.” Philosophy and Phenomenological
Research 88, (2014): 1–37.
McKay, Ryan T., and Daneil C. Dennet. “The Evolution of Misbelief.” Behavioral
and Brain Sciences 32 (2009): 493–561.
Mele, Alfred. Irrationality: An Essay on Akrasia, Self-Deception, and Self-Control.
New York: Oxford University Press, 1987.
———. Springs of Action. Oxford: Oxford University Press, 1992.
Moore, G. E. “Proof of the External World.” In Philosophical Papers. New York:
Collier Books, 1939.
Moran, Richard. Authority and Estrangement. Princeton: Princeton University
Press, 2001.
Nagel, Thomas. “War and Massacre.” Philosophy & Public Affairs 1, no. 2 (1972):
123–144.
Nozick, Robert. Philosophical Explorations. Harvard: Belknap Press, 1981.
Owen, David. Hume’s Reason. Oxford: Oxford University Press, 2002.
Owens, David. “Does Belief Have an Aim?” Philosophical Studies 115 (2003):
283–305.
———. “Epistemic Akrasia.” The Monist 85 (2002): 381–397.
———. Reason without Freedom: The Problem of Epistemic Normativity. London:
Routledge, 2000.
Petit, Philip, and Michael Smith. “Freedom in Belief and Desire.” Journal of Philoso-
phy 93 (1996): 429–449.
Plantinga, Alvin, and Nicholas Wolterstorff. Faith and Rationality: Reason and
Belief in God. Notre Dame: University of Notre Dame Press, 1983.
Platts, Mark. Ways of Meaning. London: Routledge and Kegan Paul, 1979.
Pryor, James. “The Skeptic and the Dogmatist.” Nous 34, no. 4 (2000): 517–549.
Schoenfield, Miriam. “Permission to Believe: Why Permissivism Is True and What It
Tells Us about Irrelevant Influences on Belief.” Nous 48, no. 2 (2014): 193–218.
138 Bibliography
Schwitzgebel, Eric. “A Phenomenal, Dispositional Account of Belief.” Nous 36
(2002): 249–275.
Sellars, Wilfred. “Empiricism and the Philosophy of Mind.” In Minnesota Studies
in the Philosophy of Science, edited by H. Feigl and M. Scriven, 253–329. Min-
neapolis: University of Minnesota Press (1956).
Shah, Nishi. “A New Argument for Evidentialism.” The Philosophical Quarterly
56, no. 225 (2006): 481–498.
———. “How Truth Governs Belief.” The Philosophical Review 112, no. 4 (2003):
467.
———, and J. David Velleman. “Doxastic Deliberation.” The Philosophical Review
114, no. 4 (2005): 497–534.
Smith, Angela M. “Responsibility for Attitudes: Activity and Passivity in Mental
Life.” Ethics 115, no. 2 (2005): 236–271.
Sosa, Ernest. Knowing Full Well. Princeton: Princeton University Press, 2010.
Sreenivasan, Gopal. Emotion and Virtue (forthcoming). Princeton: Princeton Uni-
versity Press.
Steup, Matthias. “Doxastic Freedom.” Synthese 161 (2008): 375–392.
Stich, Stephen. The Fragmentation of Reason. Cambridge: MIT Press, 1990.
Strawson, Peter. “Freedom and Resentment.” In Freedom and Resentment and
Other Essays. London: Methuen.1974.
Stroud, Sarah. “Epistemic Partiality in Friendship.” Ethics 116, no. 3 (2006):
498–524.
Talbot, Brian. “Truth Promoting Non-evidential Reasons for Belief.” Philosophical
Studies 168 (2014): 599–618.
Vahid, Hamid. “Aiming at Truth: Doxastic vs Epistemic Goals.” Philosophical Stud-
ies 131, no. 2 (2006): 303–335.
Velleman, David. The Possibility of Practical Reason. Oxford: Oxford University
Press, 2000.
Watson, Gary, ed. Free Will. Oxford: Oxford University Press, 2003.
———. “Skepticism About Weakness of Will.” Philosophical Review 86 (1977):
316–339.
Wedgwood, Ralph. “The Aim of Belief.” Philosophical Perspectives 16 (2002):
267–296.
White, Roger. “Problems for Dogmatism.” Philosophical Studies 131 (2006):
525–557.
Whiting, Daniel. “Does Belief Aim (Only) at the Truth?” Pacific Philosophical
Quarterly 93, (2012): 279–300.
———. “Reasons for Belief, Reasons for Action, the Aim of Belief, and the Aim
of Action.” In Epistemic Norms, edited by Clayton Littlejohn and John Turri.
Oxford: Oxford University Press, 2014.
———. “The Good and the True (or the Bad and the False).” Philosophy 88, no. 2
(2013): 219–242.
Williams, Bernard. “Deciding to Believe.” In Problems of Self, 136–151. Cambridge:
Cambridge University Press, 1976.
Wittgenstein, Ludwig. On Certainty. Oxford: Blackwell, 1669.
Wolterstorff, Nicholas. “Can Belief in God Be Rational If It Has No Foundations.”
In Faith and Rationality: Reason and Belief in God, edited by Alvin Plantinga
and Nicholas Wolterstorff, 135–186. Notre Dame: University of Notre Dame
Press, 1983.
Zagsebski, Linda. On Epistemology. Belmont: Wadsworth, 2009.
Zalabardo, Jose. “Why Believe the Truth? Shah and Velleman on the Aim of Belief.”
Philosophical Explorations 13, no. 1 (2010): 1–31.
Index

abandoning belief, conditions for 18–21 Alston, William 88n2

Abramson, Lyn 70n4 Anscombe, Elizabeth 65
acceptance vs. belief 88n15 Aristotle 9n6, 41–2, 68
accountability: for responsibility, mit- assessment: of agency 68; of attitudes
igation of 112; for states gov- 110; of belief, relevant criteria
erned by reason 91, 98, 102 for 2; of blame for moral respon-
actions: beliefs and 3, 122; beliefs sibility 120; of blame for vice
likeness to 86; responsibility for 119; of culpability 101; of desire
84–7, 123n2 65–6; of moral responsibility
act-utilitarianism 47 104, 120; of responsibility 98,
Adams 93 101–5, 112; of voluntary control
Adler, Jonathan 5, 9n2, 17–18, 21, over mental state/attitude 110
33n21, 61, 66–7, 80, 81, 85, 93, attitudes: assessing control over 110;
123 commitment-constituted 99–100,
affective vs. credal feelings 66 102; evaluative control over
agency: beliefs and 77, 95, 123; 100–1, 102–3; nonvoluntary,
boundaries of 109; boundaries control over 108; reactive 87n1
of responsibility and 95–100; attractiveness of believing 26–7, 81–2
capacity for 102; competence of, attributions of rationality 21
assessing 68; dimensions of 39; attributions of responsibility for actions
evaluative control as 100–1, 102; 86–7, 118–19, 123n2
faults of 123; guidance control of attributions of responsibility for belief:
112–13; paradigmatic voluntary appropriateness of 8, 77–8, 97,
actions 109; reasoning ability 101, 103–4, 108–9, 116–20,
and 81; reasons-responsive 113; 123n2; legitimacy of 85–6, 94;
responsibility for 84–5, 87n1, target of assessment in 92
112–13; resulting from decision Audi, Robert 89n27
80–2; voluntarism argument authenticity 42
78–9 automatic actions 87
agency norms 1–3, 68
akratic action 19, 33n17, 124n3; basic beliefs 61–3
see also weak willed basic good 41
akratic belief 33n21, 114 belief: Adler’s theory of 16–17; aim of
“Akratic Believing?” (Adler) 33n21 32nn3–5; appreciation-
alethic criteria: in assessing belief 2; motivation mismatch 19–21;
in deliberative belief-formation causative factors of 22–3; defined
26–7, 28 6, 10n11, 21–2, 67–8; essential
aliefs 72n34 components 83–4; meaning of
Alloy, Lauren 70n4 65–8; norms governing 4–6,
140 Index
10n11; ought of 48; processes choice 25, 57–8, 109
resulting in 113; purpose of 48; Chrisman, Matthew 69, 123
rocks vs. 10n19; truth-aim of Clifford, W. K. 5, 38, 53, 55–6, 95, 123
2–3, 15–16, 18, 21–3, 27–30, 37 cognitive pluralism 10n11, 34n35
belief/action divide 86, 89n27 cognitive systems: in defense of evi-
belief/action responsibility 122 dentialism 54, 60; evolution to
belief-discordant alief 72n34 achieve truth 23; generating true
belief formation: conscious 27; as event beliefs 31, 40; for goal achieve-
vs. action 89n27; evidential ment 31
ambiguity in 56–7; intention in Cohen, Jonathan 66
89n27; norms governing 37, 48; commitment, evaluative 96–7
proper 37 commitment-constituted attitudes
belief-forming mechanisms: memory 99–100, 102
116–17; perception 114–16; rea- commitments, responsibility for 101–3
son 117–20; reasons-responsive compatibilist freedom 87, 110, 120
112–14 compelled belief 18–21, 112, 114,
belief governance 64 120–1
belief norms 3 consequentialism 47
belief not grounded in evidence, a belief control: failure of 116; indirect 109,
not grounded in evidence 16–21 121–2; over memory in belief
belief perseverance 20–1 formation 88n15, 116–17; over
belief-truth-evidence connection 16–21 nonvoluntary attitudes 108; over
believers: incontinent 20; role of 48 perception in belief formation
believing, successful 48 114–16; regulative 111; without
believing at will 15, 109 alternative possibilities 111
believing correctly 3, 41 conventional norms 48
believing well 4, 38 conversational stance 126n18
Bering, Jesse 63 correct beliefs 21–30, 41
Berker, Selim 50n25 correctness: of irrational beliefs 24;
Bishop, John 56, 57 truth as standard of 26
blame: for beliefs, appropriateness of credal vs. affective feelings 66
82–7, 91, 92–5, 118–19, 120–1,
123; excusing conditions 118, deciding to believe: deliberation in
120–1; mitigating 101–2, 28–9, 78–82, 108; inconclu-
113–14, 120; for unjustified sive evidence in 28–9; with
beliefs 119; without voluntary insufficient evidence 57–8; for
control 84, 92–6, 113–14 non-evidentiary reasons 29; will
blameworthy belief 77–8, 82–7, 87n1, relationship 15; see also doxastic
94–5, 118, 120, 123 voluntarism
Boyle, Matthew 89n27 deciding what to believe 68–9
Bratman, Michael 88n15 decision theory 81
defective believing: blame assigned for
capacity, diminished 102 93, 118; perception and 114–16;
caring, real 46 responsibility for 113–14
causal determinism 111 deliberative belief-formation 26–31
character: responsibility for 101; virtue deliberative constraint 30
of 9n6 Dennett, Daniel 55, 67–8
character-based accounts of responsi- depressive realism 70n4
bility: blame in 92–3, 118–20; Descartes 45, 78–9, 95
generalization in 103–5; indirect desirability: of believing 26–7, 81–2; as
control in 109; mitigation possi- an indicator of truth 29
bility 103–5; ownership in 111; desire: belief vs. 65–6, 68; facts in
values and commitments 101–3; assessments of 65–6; irresistible
without voluntary control 92–5, 33n17, 113
118–19 determinism 111
 Index 141
dispositional accounts of belief 67–8 epistemic vices 92–3
dogmatism 92 epistemic virtue 24
doxastic agency: excusing conditions error, avoiding vs. knowing truth
120–1; memory and 62, 116–17, 55–6
119; perceptual beliefs and ethics of belief 1–3, 9n2, 18, 30–1, 48,
114–16; reason and 117–20; 77, 83
responsibility in 105n1, 108–14 eudaimonia 41
doxastic control: active/passive division evaluative control 100–1, 108
95–101; deliberation in 78–82, evidence: appreciation-motivation mis-
108, 111; delinking responsibil- match in light of 19–21; ignor-
ity from 95–6; as guidance con- ing, danger of 60–1
trol 112; responsibility for 77–8; evidential ambiguity 56
voluntarism argument 77–8 evidentialism: belief within 6; concep-
doxastic deliberation 26–8, 30–1, tual necessity of 17–21; correct
78–82, 108, 111, 117–20 beliefs and 21–30; in defense of
doxastic freedom 87, 105n1, 110, 53–4; defined 1; deviation from,
125n10 benefits of 54–5, 60–1; epistemic
doxastic ought 68–70, 122–3 value of 38–9, 60; Hume on 4–6;
doxastic responsibility: control require- opposition to 5–6; transparency
ment in 105n1, 111–12; doxastic and 26–7, 30; value of truth and
guidance control and 111–12, knowledge 52
117–18; mechanisms-based evidential norms 17, 21–4, 37
approach 112–14; no doxastic evolutionary theory: beliefs track
responsibility response 77–8, truth within 22–3; meaning-
82–7; non-derivative 94; making beliefs about life events
rational-relations account of 53–4; permissibility, relevance
95–6; voluntarism argument of to 54
77–82, 84, 108–9 excused wrongs 89n24, 98
doxastic responsibility without voluntary external-world hypothesis 58–9, 63, 64
control: attributions of respon- extrinsic value 49n10
sibility in 84, 92–8, 113–14;
character-based accounts 92–5, false beliefs: abandonment’s relation to
101–5, 109, 111, 118–20; char- recognition of 18–21; no beliefs
acterizations, primary 91–2; vs. 55–6; possibility of 22–3;
deontological judgments, appro- rationality of 24; value and dis-
priateness of 82–7; passive/active value in 50n18, 53, 54–5, 65
division 95–101; reasons-respon- family see loved ones
sive accounts 95–105, 119 feelings 9n6, 66–8, 120; see also guilt
doxastic venture 56, 57 Feldman, Richard 38–9, 60, 87n2, 110,
doxastic voluntarism 27, 77–82, 84, 122
108–9, 130 final good 41
duties to believe 9n1 final value 49n10
Fischer, John 93, 111–12, 117, 120,
emotion-issuing mechanisms 120 121
Engel, Pascal 105n1 Fish, Stanley 46, 113
epistemic agency, existence of 82 flourishing life 41–8, 52–4
epistemically irrational beliefs 24–5 foundationalism, classic 61–2
epistemic excellence 39 Frankfurt, Harry 110, 113
epistemic goods 39–48 free actions 87
epistemic norms 48 freedom, necessary conditions of 110;
epistemic ought 39 see also compatibilist freedom;
epistemic performances, norms guiding doxastic freedom
24 friendship, practical permissible beliefs
epistemic teleology 50n25 in 60–1
epistemic value 37–8, 40, 47 full belief 80, 86
142 Index
Gendler, Tamar 72n34 James, William 5–6, 53, 55–7, 63
Ginet, Carl 27–8, 79–82, 87, 116 judgment: active/passive division 96–7;
goal achievement 31, 40, 44, 48 of blame-worthiness 82–7, 87n1;
goals, universality of 40 evaluative 95–8, 100, 102–3; of
God, belief in 5, 61–5 moral responsibility 95–6; sus-
goods: epistemic 39–48; of happiness pension of 52–8, 60, 61; see also
42; instrumental 44; value of blame
particular vs. other 40–1 juridical theory of responsibility 119
Great Pumpkin, belief in 62–3, 64 justifying conditions 62–3
Greco, John 40
Grimm, Stephen 50n26 kinesis/energia distinction 89n27
guidance control: concept illustrated knowledge 39–48, 52, 59
111; as form of doxastic control Kornblith, Hilary 7, 30–1, 40, 41,
112; over memory in belief for- 44–5, 46
mation 116–17; over perception
in belief formation 114–16; legitimate beliefs 5, 18–21,
ownership component 112, 117, 62–3, 66
118–22; reasons-responsiveness Levy, Nell 84, 87, 120–1
component 112–14 life events, meaning-making beliefs
guilt 98, 113–15, 119 about 53–4
gullibility 92–3 living well (life lived well) 41, 46; see
also flourishing life
happiness 41–3, 47 Locke, John 95
Heil, John 19, 20 loved ones, non-evidentially based
Heller, Mark 120 beliefs about 20–1, 60–1
Hieronymi, Pamela 92, 98–101, 104, Lynch, Michael 45, 46, 50n16
108, 109, 110
Hume, David 4–6, 37, 58, 59, the Matrix 45
63, 82–4 McDowell, John 95
McHugh, Conor 87n1, 89n27, 105n1,
ignorance as excusing condition 120–1 124n5, 125n10, 125n15
illegitimate beliefs 5, 18–21, 62–3, 66 McKay, Ryan 55, 67–8
impermissible beliefs 3, 31, 52–4, 58–9, meaning-making beliefs 63
87n1 mechanisms of belief: memory 116–17;
incoherence tests 17 perception 114–16; reason
incorrect beliefs, ought of holding 23–4 117–20; reasons-responsive
incorrigible beliefs 23, 61 112–14; responsibility for 113–14
incredulity vs. belief 83 Mele, Alfred 19, 33n17
indirect control over belief 84, 109, memory-beliefs 62, 116–17, 119
121–2 mental illness: capacity for abandoning
inescapable/indispensable beliefs 58–60 belief and 20; correctness of
inherent good 41 beliefs 22–3; depression 70n4;
inquiry and belief 80 responsibility for 106n43
instrumental goods 44 misbeliefs: adaptive 53–4, 71n17;
instrumental value 49n10 defined 68
intellectual integrity 42–3 Moore, G. E. 58, 70n13
intention, belief and 25, 66, 86–7, 99, Moore’s paradox 17, 18
101, 109–10, 111–12 moral constraints 56, 57, 61
intrinsic good 41 moral faults 123
intrinsic value 49n10 moral insanity 85
irrational beliefs 22–5, 92–3 morality 54–5
irresistible beliefs 32n13, 111 moral personality 91
irresistible desires 33n17, 113 moral personhood 99, 103
 Index 143
moral responsibility 95–6, 104, philosophical wisdom 42
111–12, 120 Plantinga, Alvin 61–4
Moran, Richard 34n36 practical deliberation 28–9
practical permissible beliefs see permis-
naturalism 10n19 sible practical beliefs
neutral evidence, belief with 53–8 practical reasons for belief 24–5,
neutrality rule 59 110–11, 117–20, 129
non-basic voluntary control 122 practical value of knowledge 40–2
non-evidentially based belief 5–6, 25, practical wisdom 42
29–31, 53–4; see also permissible pragmatism 1, 5–6, 24–7, 130
practical beliefs pragmatist, defined 30
non-instrumental value 41 pragmatist’s dilemma 30–1
non-instrumental value of truth and prejudiced beliefs 83–4
knowledge 42–5 prima facie good 43–4
nonvoluntary attitudes, control over principle of alternative possibilities
108 (P.A.P.) 126n19
normative concepts 23 proper basic beliefs 61–3
norms of agency 2 prudence 42
Nozick, Robert 59–60 Pryor, James 70n13
psychological mechanisms of belief
ought: of beliefs 93; of belief vs. action 22–3
122; of caring about knowledge purposiveness 95
48; the epistemic 39; of holding
rational beliefs 25; value associ- rational belief formation 34n31
ated 39 rational beliefs 24–5, 32n12, 34n31,
oughts 68–70 64–5
Owens, David 32n4, 82, 92–3, 95–8, rational choice theory 81
99, 101–2, 118–20 rationality norms 24, 25
ownership of beliefs 112–13, 116–22, Ravizza, Mark 111–12, 113, 117, 120,
126n16 121
reactive attitudes 87n1
paradigmatic voluntary actions 109 reason: belief as subject to 82; in deci-
Pascal, Blaise 5, 122 sion-making 81, 118–19; deliber-
passional life 56 ative constraint on 27–9; virtue/
passional nature 5, 53 vice and 91
perceptual beliefs 62, 114–16 reasonableness, virtue of 4
permissible action 47 reason-induced beliefs 117–20
permissible practical beliefs: about reasons-responsive accounts of doxastic
loved ones 60–1; constraint of responsibility without voluntary
worth 56; desires (hopes) vs. control 95–105, 119
68; determining 39; inescapable/ reasons-responsive belief-forming mech-
indispensable beliefs 58–60; anisms 112–14
moral constraints 56, 57, 61; reflective competence 94–5
with neutral evidence 53–8; per- Reformed epistemologists 61, 62, 63
nicious distinguished from 52; regulative control 111
religious belief 61–5; with silent religious belief 5, 10n15, 20, 61–5
or no evidence 53–8 religious hypothesis 63
permissivism 34n35, 51n28 representationalist accounts of belief
pernicious beliefs 5–6 67–8
pernicious practical beliefs 52, 55, responsibility: accountability for
60–1, 63 112; for actions 84–7, 123n2;
persistent beliefs 59 agency and boundaries of
philosophical investigation 6–7 95–100; assessment of 98,
144 Index
101–5, 112; capacity for 97–8, truth: attaining 52; desirability as an
101–2, 106n43; cessation of indicator of 29; doxastic
103, 106n43, 112, 118, 120; for deliberation and 30–1; evidence’s
commitments 101–3; gradual provision of 38; goodness of 44;
diminishing of 117, 118; link knowing vs. avoiding error
to control 112, 119; mitigating 55–6; role in a flourishing life
101–2, 112; requirements for 91, 41–8, 52; value of 42–5, 54–5
98, 102, 112; for values 101–3; truth-aim of belief 2–3, 15–16, 18,
will and boundaries of 91; see 21–30, 37
also character-based accounts of “Truth and Happiness” (Lynch) 46
responsibility; doxastic responsi- truth norm 24
bility; moral responsibility truth-value of beliefs 39–48, 54
right action 47
right belief 47 uncorrectable beliefs, possibility of
role ought 48 22–3, 26
rules of the wise 4 unworthy reasons, consequences of
believing for 94–5
Schwitzgebel, Eric 67, 68
self-control 95, 98, 106n43, 119; value: in false beliefs 50n18, 53, 54–5,
see also akratic action 65; instrumental 49n10; intrinsic
self-deception 80 vs. extrinsic 49n10; of truth and
self-determination, doxastic 89n27 knowledge 39–48; of virtuous
self-evident beliefs 61 belief 37
self-knowledge 50n16 values, responsibility for 101–3
self-manipulation resulting in belief 15 Velleman, David 21–6, 54, 66
Sellars, Wilfrid 61 vice 91, 101, 119
senses, nature of existence independent virtue 4, 41–2, 91, 98
of 58 virtue of character 9n6
sensitive belief 71n15, 72n34 virtues of thought 41–2
Shah, Nishi 25–31, 68–9 virtuous action 42
silent or no evidence, belief with 53–8 virtuous belief 37
skepticism 58 virtuous person, characteristics of 44
Smith, Angela M. 92, 96–100, 102–3, voluntariness 109
120 voluntarism, Descartes on 78–9
Sosa, Ernest 24, 40, 50n26 voluntarism argument in doxastic
Steup, Matthias 86–7, 110 responsibility 77–8, 108
Stich, Stephen 10n11, 34n35 voluntary actions 109–11
strong willed 109–10 voluntary control, assessing 110
Stroud, Sarah 60
suspension of belief 56, 61 weak willed 109–10, 124n3; see also
suspension of judgment 52–8, akratic action
60, 61 Wedgwood, Ralph 16, 23–6, 32n12
Whiting, Daniel 32n5, 35n45
teleologist’s dilemma 26 will 91, 109–10
temperance 98, 101 at will actions 110
transparency 26–7, 30, 110–11 at will belief 15, 109
trivial true beliefs 44, 51n26, 56 Williams, Bernard 15–16, 18, 30, 68–9
true beliefs: cognitive systems in gen- wisdom 42, 92
erating 31, 40; as correct beliefs the wise 4, 37, 94
21–3, 26; goals achievement and wishful thinking 26, 80
48; obligation to investigate 85; Wittgenstein, Ludwig 64
ought of holding 23–4; Wolterstorff, Nicholas 61, 64–5
pragmatic reasons for caring
about 31 Zagzebski, Linda 40