Peels, Rik - Responsible Belief - A Theory in Ethics and Epistemology (2017, Oxford University Press)

i
RESPONSIBLE BELIEF
ii
iii
RESPONSIBLE BELIEF
A Theory in Ethics and Epistemology
Rik Peels
1
iv
1
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.
Published in the United States of America by Oxford University Press

198 Madison Avenue, New York, NY 10016, United States of America.
© Oxford University Press 2017
All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.
You must not circulate this work in any other form

and you must impose this same condition on any acquirer.
CIP data is on file at the Library of Congress

ISBN 978–0 –19–0 60811–8
1 3 5 7 9 8 6 4 2
Printed by Sheridan Books, Inc., United States of America
v
To Herman Philipse
Ferrum ferro acuitur et homo exacuit
faciem amici sui (Proverbia 27:17)
vi
vii
CONTENTS
Acknowledgments xi
Introduction: Why Responsible Belief Matters 1

I.1 Responsible Belief and Blameworthy Belief 1
I.2 A im of the Book 3
I.3 W hy We Need an Account
of Responsible Belief 6
I.4 A
New Approach: Intellectual Obligations
and Doxastic Excuses 8
I.5 Outline 11
1. Doxastic Responsibility: What Is It? 14
1.1 Introduction 14
1.2 R esponsibility and Appraisal 15
1.3 Occurrent, Dormant, and Tacit Belief 28
1.4 R esponsible Belief Is Blameless Belief 43
1.5 Conclusion 51
vii
viii
C ontents
2. The Problem: Doxastic Control

and Doxastic Obligations 52
2.1 Introduction 52
2.2 Th
e Argument from Doxastic Involuntarism 53
2.3 Obligations and Control 56
2.4 First Response: We Have Direct Intentional
Doxastic Control 61
2.5 S econd Response: We Have Indirect
Intentional Doxastic Control 66
2.6 Th
ird Response: We Have Compatibilist
Doxastic Control 72
2.7 F
ourth Response: Doxastic Obligations
Without Any Kind of Control 81
2.8 Conclusion 87
3. Th
e Solution: Doxastic Influence
and Intellectual Obligations 89
3.1 Introduction 89
3.2 Belief-influencing Factors 90
3.3 W hat Are Intellectual Obligations? 96
3.4 W hat Intellectual Obligations Do We Have? 100
3.5 Six Objections and Replies 108
3.6 Intellectual Obligations and Responsible Belief 117
3.7 Doxastic Excuses: Force, Ignorance, and Luck 124
3.8 Conclusion 130
4. R
esponsible Belief Entails the Ability
to Believe Otherwise 132
4.1 Introduction 132
4.2 W hat Is Force? 134
4.3 F orce as a Doxastic Excuse 137
4.4 B lameworthy Force 143
viii
ix
C ontents
4.5 C ontrol over or Influence on Properties

of Beliefs 148
4.6 Doxastic Frankfurt-style Scenarios 150
4.7 Th
e Doxastic Asymmetry Objection 157
4.8 Conclusion 162
5. Responsible Belief Is Radically Subjective 164
5.2 W hat Is Ignorance? 165
5.3 Ignorance as a Doxastic Excuse 169
5.4 Ignorance as a Narrow and a Broad
Doxastic Excuse 182
5.5 B lameworthy Belief and the Regress Problem 185
5.6 Conclusion 198
6. R
esponsible Belief Is Compatible
with Doxastic Luck 199
6.2 W hat Is Luck? 200
6.3 F our Varieties of Doxastic Luck and Luck as
a Doxastic Excuse 207
6.4 C onsequential Luck and Accidentality 215
6.5 The Problem of Doxastic Luck 218
6.6 Conclusion 234
Appendix: Responsible Belief and Epistemically Justified Belief 237

Bibliography 251
Index 267
ix
x
xi
ACKNOWLEDGMENTS
My first encounter with philosophy took place when it was around five
o’clock in the morning. As a fifteen-year-old, I participated in a twenty-
four-hour class marathon to raise money for some charity organization.
Ieke Haarsma, who taught philosophy at my school, asked us whether
plants can sleep. I first considered this a silly question—of course plants
cannot sleep. In fact—just to assure the reader—I still think plants
cannot sleep. Upon further consideration, however, it seemed to me
not as obvious why precisely we should think that plants cannot sleep.
Ever since that little event early in the morning, philosophy has gained
increasing importance in my life. What started as an intellectual game
has become for me a highly influential method for approaching, under-
standing, and sometimes even changing reality. Indeed, it has become
an inalienable part of my own identity, bringing along both the advan-
tages and disadvantages of a philosophical life. I thank Ieke Haarsma for
introducing me to the wonderful world of philosophy.
I could not have written this book without the help, support, and
critical feedback from my colleagues and students at the Philosophy
Departments of Utrecht University and the Vrije Universiteit
Amsterdam. These departments, as well as the Philosophy Department
of Oxford University during Hilary and Trinity terms 2011, have
xi
xii
A cknowledgments
provided a stimulating research environment in which working on this

project was a true pleasure.
Much of the material of this book was presented in some earlier (and,
I am afraid, often premature) version at several conferences. I thank
the audiences for their searching questions and helpful suggestions at
the conference on Doxastic Agency and Epistemic Responsibility at the
University of Bochum, Germany; the Fifth Annual Dutch OZSW
Conference at the Erasmus University of Rotterdam, the Netherlands;
the Fourth Annual Dutch Conference on Practical Philosophy at the
Technological University Eindhoven, the Netherlands; the Epistemology
Meeting: Doxastic Attitudes at Ghent University, Belgium; the 85th
Joint Session of the Aristotelian Society and the Mind Association at
the University of Sussex, United Kingdom; the Second Annual Dutch
Conference on Practical Philosophy at the University of Groningen, the
Netherlands; the conference on Moral Responsibility: Neuroscience,
Organization, and Engineering at the Technological University of Delft,
the Netherlands; the Meta-ethics and Moral Psychology Study Group at
Utrecht University; the 19th Inter-u niversity Workshop on Philosophy
and Cognitive Science at the University of Zaragoza, Spain; the confer-
ence on Recent Trends in Philosophy of Language and General Analytic
Philosophy at Tilburg University, the Netherlands; The Young European
Epistemologists Workshop and the Epistemic Agency Conference at
the University of Geneva; the conference on Responsible Belief in the
Face of Disagreement, the First and Second Workshops in Analytic
Philosophy; and the conference on The Value of Knowledge at the Vrije
Universiteit Amsterdam.
For their careful reading and rigorous comments on the entire
book, I would like to thank Tony Booth, René van Woudenberg, and
Ralph Wedgwood. I am heavily indebted to them. I would also like
to thank Matthias Steup and Nicholas Wolterstorff who reviewed the
manuscript for Oxford University Press and went to great lengths to
provide helpful comments. These have led to the rewriting of many
confused and unclear passages. I would also like to thank Niko
Nottelmann; I have benefited immensely from many enjoyable dis-
cussions with him on a large number of the issues dealt with in this
book.
xii
xiii
A cknowledgments
For their astute comments on one or several parts (of earlier versions
or presentations) of the book, I would like to thank Joel Anderson, Max
Baker-Hytch, Arianna Betti, Martijn Blaauw, Martijn Boot, Gerhard
Bos, Niek Brunsveld, Adam Carter, Bob Coenraats, Lieven Decock,
Boudewijn de Bruin, Leon de Bruin, Henk de Regt, Jeroen de Ridder,
Catarina Dutilh Novaes, Richard Feldman, Mikkel Gerken, Jeffrey
Glick, Alvin Goldman, John Greco, Dirk-Martin Grube, Peter Hacker,
Paul Helm, Peter Kirschenmann, Hilary Kornblith, Andrea Kruse,
Pierre Le Morvan, Joop Leo, Clayton Littlejohn, Sebastian Lutz, Rosja
Mastop, Conor McHugh, Anne Meylan, Jesse Mulder, Fred Muller,
Thomas Müller, Bert Musschenga, Philip Nickel, Dawa Ometto, Marina
Oshana, Alvin Plantinga, Stefan Roski, Emanuel Rutten, Marcel
Sarot, Katrien Schaubroeck, Jeroen Smid, Michael Smith, Nicholas
Southwood, Matthias Steup, Derek Strijbos, Jesper Tijmstra, Dorette
van der Tholen, Janneke van Lith, Gert van den Brink, Jessica van der
Schalk, Vincent van Oostrom, Han van Wietmarschen, Albert Visser,
Jan Vorstenbosch, Heinrich Wansing, Brian Weatherson, Jan Willem
Wieland, Timothy Williamson, and Michael Zimmerman.
I thank the Templeton World Charity Foundation, through whose
support publication of this book was made possible. The opinions
expressed in this publication are mine and do not necessarily reflect the
views of Templeton World Charity Foundation.
A heartfelt thank you to my executive editor, Peter Ohlin, my pro-
duction editors Emily Sacharin and Radha Subburathnam, and my copy
editor Johanna Baboukis, for their continual encouragement and their
hard work to make the publication of this book possible.
Several of the ideas and arguments in this book are based on work
that I published previously. Much of it has undergone substantial
revision—hopefully for the better—but many of the core ideas survived
and made it into this manuscript. I thank the editors involved for their
kind permission to use this material.
The final section of chapter 1 is based on “Why Responsible Belief
Is Blameless Belief ” (co-authored with Anthony Booth), The Journal
of Philosophy 107.5 (2010), 257–265. The main body of chapter 2
is derived from “Against Doxastic Compatibilism,” Philosophy and
Phenomenological Research 89.3 (2014), 679–702. The paragraphs on
xiii
xiv
A cknowledgments
belief-policies in c hapter 3 summarize what I argue in “Belief-Policies

Cannot Ground Doxastic Responsibility,” Erkenntnis 78.3 (2013), 561–
569. The main body of c hapter 4 draws from material published in “Does
Doxastic Responsibility Entail the Ability to Believe Otherwise?,”
Synthese 190.17 (2013), 3651–3669.
The first section of c hapter 5, on the nature of ignorance, makes
use of material from “What Is Ignorance?,” Philosophia 38.1 (2010),
57–67, “Ignorance Is Lack of True Belief: A Rejoinder to Le Morvan,”
Philosophia 39.2 (2011), 345–355, and “The New View on Ignorance
Undefeated,” Philosophia 40.4 (2012), 741–750. The main argument of
the chapter can be found in an earlier version in “What Kind of Ignorance
Excuses? Two Neglected Issues,” Philosophical Quarterly 64.256 (2014),
478–496. The final section draws on material from “Tracing Culpable
Ignorance,” Logos and Episteme 2.4 (2011), 575–582.
The parts of chapter 6 on the nature of luck and the problem of
doxastic luck are based on “A Modal Solution to the Problem of Moral
Luck,” American Philosophical Quarterly 64.256 (2015), 478– 496.
The Appendix makes use of material from “Let’s Bite the Bullet on
Deontological Epistemic Justification: A Response to Robert Lockie,”
Social Epistemology Review and Reply Collective 4.12 (2015), 42–49, and
can be found in a more elaborate form in “Epistemic Justification and
Responsible Belief,” Synthese (2016), forthcoming.
Finally, I would like to thank Herman Philipse, who was also my
PhD supervisor, for his many invaluable, constructive comments.
I especially want to thank him for his hospitality and the fun we had dur-
ing our memorable fortnightly paper discussion sessions at his home.
Even, or maybe especially, where our views do not seem to converge,
such as our opinions in the philosophy of religion, his ideas and criti-
cisms have been greatly inspiring for me. It is to him that I dedicate this
book, with gratitude.
xiv
1
INTRODUCTION
Why Responsible Belief Matters
I.1 R ESPONSIBLE BELIEF

AND BL AMEWORTHY BELIEF
Toward the end of World War II, in April 1945, the Germans assem-
bled a fleet of three ships in the Bay of Lübeck: the Cap Arcona, the SS
Deutschland, and the Thielbek. They filled them with close to ten thou-
sand concentration camp survivors and war prisoners. On May 3, 1945,
Sir Arthur Coningham, commander in the British Tactical Air Force,
ordered the attack on all German ships in the Baltic, including these
three ships. However, he was ignorant of certain important facts. For
instance, he falsely believed that there were only SS men and other
German soldiers on board. All three ships were sunk. Most of the SS
guards survived, but an estimated 7,800 camp survivors and war
prisoners died.
Let us assume that Coningham’s other beliefs were formed and
maintained responsibly rather than culpably. Thus, he responsibly
believed that taking down ships with German soldiers was a good deci-
sion at this stage of the war, he responsibly believed that these ships were
playing an important role in the military defense of Germany, he respon-
sibly believed that they would escape and be harmful if he were to let
them go, and so forth. Then the answer to the question whether he acted
responsibly in taking down these ships seems to depend on whether or
not his belief that there were only German soldiers on board was held
1
2
R esponsible B elief
responsibly or culpably. It seems that if he was not culpable for hold-

ing this belief, it is unfair to blame him for his decision. However, if he
should have known better, then it seems that he is at least partly blame-
worthy for the tragedy.1 This example, then, shows that whether or not
we act responsibly sometimes crucially depends on whether we believe
responsibly and, thereby, shows the relevance of responsible belief.
It is not only important to hold responsible factive beliefs, but also
to hold responsible normative beliefs. Let me give an example to illus-
trate the point. On July 22, 2011, the Norwegian fundamentalist Anders
Breivik bombed government buildings in Oslo, killing eight people,
and then murdered another sixty-n ine people in a mass shooting on the
island of Utøya. He did not act from a sudden, vehement emotion, but
planned the massacre carefully. Most importantly for the issue at hand,
he acted from a firm conviction that his country and Europe as a whole
are under threat from Islam and that he needed to make a statement.
He believed he ought to kill a large number of people in order to put the
topic on the national and international political agenda. It seems that if
he had formed his beliefs responsibly, he would not have held this belief
and he would, therefore, not have committed such horrible crimes.
Mutatis mutandis, the same can be said about the terrorists attacks in
Paris in January and November 2015 that were carried out on the basis
of Islamic fundamentalist beliefs, and other instances of extremism.
Again, whether one acts culpably or responsibly sometimes crucially
depends on whether one believes culpably or responsibly.
Doxastic responsibility—responsibility for belief—does not only
figure prominently in situations where much is at stake. It seems that
we also hold each other responsible for our beliefs in more mundane
situations. The following sentences can plausibly be interpreted as evi-
dencing this: “Laurence had no right to think that I lied to him,” “Julia
shouldn’t have believed that demagogue so readily,” and “Henry ought
to believe that he will pass the exams.” We can also think of such expres-
sions as “The prime minister should have known about the fraud” and
“You shouldn’t have known about the surprise party,” for statements like
1. For a famous fictional example that is in some ways similar, see Clifford (1901, 163–165).
2
3
I ntroduction : W hy R esponsible B elief M atters
these can be understood as saying that someone should (not) have held
certain true beliefs about some matter. It seems that the phenomenon of
holding someone doxastically responsible is pervasive in our lives.2
I.2 A IM OF THE BOOK
The purpose of this book is to develop an account of responsible belief.

By an “account of responsible belief ” I mean an informative analysis
in terms of necessary and sufficient conditions of what it is to believe
responsibly. I consider such an account informative if it tells us in
terms on which we have a fairly firm grip what it is to believe respon-
sibly. Thus, upon being fully informed about someone’s situation, this
account should provide us with the basic means to distinguish whether
that person believes responsibly or not. Let me say explicitly that I do
not think that this requires some kind of reduction of normative proper-
ties to natural properties, if that is possible at all. For, as Peter Strawson
rightly points out, we can elucidate a thing or a concept by showing its
relation to other things or concepts. 3 For example, modal notions like
“necessity” and “counterfactual truth” can be explicated in terms of pos-
sible worlds, even if the notion of a possible world itself needs as much
clarification.
Let me also point out what I do not seek to do. In setting these limits,
I bring the aim of this book more sharply into focus.
First, the account is confined to doxastic responsibility of individu-
als rather than groups of people. It would be another and surely inter-
esting project, within social epistemology, to develop an account of
the ethics of group belief or maybe group knowledge.4 Clearly, though,
2. D uring the last few decades, several philosophers have pointed out that we are often
doxastically responsible and that we should believe responsibly rather than culpably,
e.g., BonJour (1985, 42), Chisholm (1977, 14), Kim (1994, 282–2 84), Smith (1983, 545),
Wolterstorff (2005, 326–338).
3. He calls this “connective analysis.” See Strawson (1959, 9–12; 1992, 17–2 8).
4. R ecently, there has been much attention to group belief, e.g., Schmitt (2014), and group
knowledge, e.g., De Ridder (2013).
3
4
doxastic responsibility of groups of people would not remove the doxas-

tic responsibility of individuals.
Second, we can appraise people for other people’s beliefs, e.g., in
cases of indoctrination or brainwashing, but also in situations that we
encounter more often in daily life, such as certain cases of manipulation.
Here, I limit myself to people’s responsibility for their own beliefs.
Third, one might think that we also bear responsibility for the degree
to which we hold our beliefs. Whether this is correct and, if so, how such
responsibility is to be understood, is a topic that I will not broach here.
Fourth, I bracket the issue of how responsible belief relates to
epistemic justification and treat it separately in the Appendix. 5 This is
because, as the above examples show, doxastic responsibility is a topic
that deserves philosophical attention no matter how it is related to epis-
temically justified belief—whether it is identical to it, similar to it, or
completely different from it. Nonetheless, the notion of epistemic jus-
tification will show up at several points in the ensuing chapters. For,
especially in the debate about the so-called deontological conception
of epistemic justification—a conception on which justified belief is
responsible or blameless belief—a crucial problem for an account of
responsible belief has been formulated. Paying attention to that debate,
though, will be entirely instrumental: I consider it to the extent that it is
relevant to responsible belief, no matter what it says about epistemically
justified belief.
Fifth, I take an account of responsible belief not to include any dox-
astic advice. It might not be too difficult to derive certain epistemic rec-
ommendations from my account of responsible belief, but I leave such
practical ramifications for another occasion—here, I only aim at getting
the theory right.6
5. For more on my view on epistemic justification, independently of how it relates to

responsible belief, see Peels (2010a).
6. P erhaps the most famous example of such doxastic advice is found in John Locke’s Essay
Concerning Human Understanding IV.xvii.24; see Locke (1975, 687–688). According to
Wolterstorff (1996, xv–x vii), it was even Locke’s main purpose in that book to give dox-
astic advice rather than to develop an account of responsible belief.
4
5
Let me stress that the aim of this book is to provide an account of

responsible belief, not (merely) an account of our concepts of such things
as responsibility and belief. This is because certain things may be true
about responsible belief that one might fail to realize, even if one has the
concepts of responsibility, belief, and doxastic responsibility. This is not
unusual: there may be certain truths about, say, knowledge that many
people who competently use the concept of knowledge are unaware of,
as Edmund Gettier famously showed.7 I would like to aver that this does
not mean that we can do without an analysis of words and concepts.8
After all, we use language and concepts in thinking and talking about
reality. As J.L. Austin famously said, ordinary language is not the last
word for philosophers, but it is nonetheless the first word.9
But why should we think that there are indeed necessary and suf-
ficient conditions for responsible belief? An alternative thought would
be that what unites cases of responsible belief is that they bear a certain
family resemblance to each other. Let me give two reasons in favor of my
approach. First, ethicists hardly ever analyze responsible action in terms
of family resemblances. This provides us with at least prima facie good
reason to adopt a similar method when it comes to responsible belief.
After all, both concern responsibility and it is widely believed that there
are necessary and sufficient conditions for responsibility. Second, I sub-
mit that my final account unites cases of responsible belief in a natu-
ral way, that is, in a way that meets (most of) our intuitions concerning
what it is to believe responsibly and what it is not. If this is correct, then
we have good reason to think that there are indeed necessary and suffi-
cient conditions for responsible belief. This is not to say that my account
matches our ordinary use of the expression “responsible belief.” In fact,
it seems that we hardly ever use that expression. Nevertheless, we do
have the concept of believing responsibly. By formulating necessary and
sufficient conditions for responsible belief I aim to provide an account of
responsible belief that is more precise than our ordinary, unarticulated
conception of what it is to believe responsibly.
7. See Gettier (1963).

8. For the same point, see Williamson (2007b, 45–47).
9. See Austin (1979, 185).
5
6
One may object that an analysis in terms of necessary and sufficient

conditions conflicts with the fact that there are not only paradigm cases
of responsible belief, but also boundary cases, that is, cases in which we
hesitate to say that someone believes responsibly, even if we are fully
informed. I think there are indeed such cases, but it seems to me that
this does not conflict with my approach to the issue. For those necessary
and sufficient conditions may, to a certain extent, themselves be vague.
On my analysis of responsible belief this is indeed the case. My analy-
sans contains such notions as responsibility, belief, and ignorance, and
my analyses of these notions are themselves vague. For instance, I ana-
lyze responsibility in terms of being the proper object of attitudes like
praise and blame. But there seem to be boundary cases: sometimes it is
not clear whether someone is the proper object of, say, a little blame or
no blame at all, even if we are fully informed about that person and her
situation. Hence, my approach to analyzing responsible belief by way of
formulating necessary and sufficient conditions is perfectly compatible
with there being paradigm and boundary cases of responsible belief.
I.3 W
HY WE NEED AN ACCOUNT
OF RESPONSIBLE BELIEF
I argued above that it matters whether or not we believe responsibly. But

why is it valuable to provide an account of responsible belief? This is an
important question, for not just anything that matters in our lives is in
need of a philosophical account. There are at least three reasons why we
need an account of responsible belief.
First, as William Alston famously argued, the idea that we believe
responsibly or culpably faces an important philosophical challenge.10
We can hold each other responsible, for instance, for what we say,
because we can choose to make an offensive comment or wisely decide to
make the point in a more subtle way. But it seems we do not and cannot
choose our beliefs. To believe something is simply to think or take it that
10. S ee Alston (1989d).
6
7
something is the case or that something is true. I cannot now choose to

believe that Nantes is the capital of France, that gender bias is a good
thing, or that the number of stars is even. But if we cannot choose our
beliefs, then how can we be responsible for them? We need an account
of responsible belief that meets this challenge by explaining how we can
believe responsibly or culpably despite the fact—i f it is a fact—t hat we
cannot choose our beliefs.
Second, as I indicated above, whether a person has acted respon-
sibly or culpably sometimes crucially depends on whether or not the
belief from which she acted was held responsibly or culpably.11 This is
reflected in law: certain laws explicitly state that one is to be punished
only if one acted from a mens rea (guilty mind), that is, a mental state
for which one is blameworthy. Here, we should think of such things as
malicious intent, but also culpable ignorance. Now, in certain cases it
might be obvious what a person should have believed, but in other cases
things are not so clear. Providing synonymous expressions, such as
“from an irrational belief ” or “she should have known better,” is of little
help in such circumstances. What we need is an account of responsible
belief.12
Third, maybe believing responsibly is of more importance to us than
knowing, as some philosophers contend.13 But even if it is not, it should
concern us all. For, as I argued above, what we believe often makes a
crucial difference to how we act. Thus, we should take our doxastic
household seriously by forming responsible rather than blameworthy
beliefs. However, if we want to believe responsibly, we need at least
some grasp of what it is to believe responsibly. An account of responsible
belief, therefore, provides us with the material from which we can derive
guidelines for our cognitive behavior. As I said above, I will not provide
11. Th is is rightly noticed by Levy (2007, 127–128), Montmarquet (1995, 46), Nottelmann
(2007, 3–10).
12. Slightly more precisely, what may suffice in such cases is an account of blameless rather
than responsible belief. In c hapter 1, however, I argue that responsible belief entails
blameless belief, so that giving an account of responsible belief sheds light on merely
blameless belief as well. Also, my first and third considerations provide reasons to
develop an account of responsible rather than merely blameless belief.
13. Th
us, for instance, Foley (2005a, 313), Wolterstorff (2005, 328).
7
8
such guidelines, but my account may well provide the theoretical basis
for such guidelines. It might also have certain implications for how we
judge other people’s beliefs. Like certain theories in ethics an account of
responsible belief might have inflationary or deflationary implications in
the sense that our practice of doxastic praise and blame may turn out to
be too conservative or too liberal. Maybe our social practice of holding
each other doxastically responsible needs revision. Whether it does is a
question that can be answered only if we have a solid empirical analysis
of our practice of holding each other responsible and a plausible account
of what it is to believe responsibly. This book aims to provide the latter.
I.4 A
NEW APPROACH: INTELLECTUAL
OBLIGATIONS AND DOX ASTIC EXCUSES
Unsurprisingly, I am not the first one to address the issue of what it is
to believe responsibly. Many philosophers have defended theses that
bear on an account of responsible belief, but a full-fledged account of
responsible belief has not yet been given.14 The account of responsible
belief that I articulate in the ensuing chapters differs to a smaller or
larger extent from all views in the field. At several junctures I highlight
these differences and defend my departure from other accounts. Let me
briefly mention the two main differences.
First, many philosophers working on the ethics of belief argue that
our lack of intentional control over our beliefs—we cannot choose
what to believe—is compatible with doxastic responsibility, because
we have compatibilist doxastic control. For instance, most of our beliefs
are responsive to reasons: we would not have held them if we had had
different reasons. In this book, I argue that approaches in terms of
14. The focus of Montmarquet (1993) is on epistemic virtues, Owens’ approach in Owens
(2000) is primarily historical, Adler (2002) seems to provide an account of rational
rather than responsible belief and embraces some kind of doxastic compatibilism (an
approach I reject), Conee and Feldman (2004) analyze rational belief and are not con-
cerned with responsibility in its full-blooded sense, and Nottelmann (2007) gives an
analysis of blameworthy rather than responsible belief.
8
9
compatibilist doxastic control are mistaken. We are indeed responsible

for our beliefs, but the fact that many of our beliefs are reason-responsive
is insufficient for such responsibility.
Instead, I provide a solution to the problem of doxastic involun-
tarism by arguing that we have influence on our beliefs. The basic idea
here is that we cannot choose our beliefs, but that we can choose to do
or refrain from doing certain things that make a difference to what we
believe, even though we hardly ever foresee to which specific beliefs it
will make a difference. Thus, we can gather further evidence on some
topic and we can work to cultivate intellectual virtues and counteract
intellectual vices, such as, respectively, open-m indedness and intellec-
tual laziness. If we have an obligation to do so and meeting or violating
that obligation makes a difference to what we believe, we have an intel-
lectual obligation. In this book, I provide a framework for thinking about
which intellectual obligations we have and how they relate to respon-
sible belief.
Second, I approach the topic of responsible belief from a particu-
lar angle, namely that of excuses. In ethics, excuses for morally wrong
actions have received considerable attention, although, as we will see,
certain lacunas remain. Many epistemologists acknowledge that, simi-
larly, we are often excused for believing something that we should not
believe.15 Remarkably, though, virtually no philosopher has attempted
to spell out what those doxastic excuses amount to. In this book, I fill
this gap by providing an account of what, I argue, are the two main dox-
astic excuses, to wit, those of force and ignorance. And I will argue that
a third proposed excuse, namely luck, is in each case reducible to either
force or ignorance.
Paying attention to doxastic excuses in formulating and defending
my account of responsible belief has at least three philosophical payoffs.
15. They use expressions such as “epistemic excuses” or “excuses for belief.” See, for
instance, Alston (2005, 59n), Feldman (1988b, 410), Heil (1984, 59, 61), Helm (1994,
172–174), Kornblith (1983, 35–36), Levy (2007, 127, 140), Montmarquet (2008,
381–382, 388), Russell (2001, 43), Schleifer-McCormick (2015, 18), Steup (1988, 78),
Vahid (1998, 297). Nottelmann (2007, 207–217) and Van Woudenberg (2009, 373–
386) elaborate on doxastic excuses, but both confine themselves to particular doxastic
excuses without explaining their relation to responsible belief.
9
10
First, as I pointed out above, without an analysis of doxastic excuses,

an account of responsible belief is incomplete. If we want to know
whether someone has acted responsibly and notice that she has not met
her obligations, we also want to know whether or not she is blameworthy
for that. We, therefore, want to know whether or not she is excused for
not meeting her obligations. Blame does not track violating an obliga-
tion, but rather violating an obligation when one has no excuse for doing
so. Knowledge of excuses, then, bridges the gap between knowledge of
obligations and knowledge of blameworthiness or blamelessness. I will
argue that this is true not only for action, but also for belief: we know
whether someone believes responsibly only if we know what that per-
son’s obligations vis-à-v is that belief were, whether she has met those
obligations, and, if not, whether she is excused for that.
Second, a concern with excuses quickly brings to light the distinc-
tion between what I call narrow and broad doxastic excuses. Narrow
excuses are conditions that merely remove blameworthiness. They do
not remove all responsibility. Thus, if one is merely narrowly excused,
one can still believe responsibly. Broad excuses, on the other hand, are
conditions that remove all responsibility—that is, praise, blame, and
neutral appraisal. Focusing on excuses, then, helps us to differentiate
between situations in which one believes responsibly and situations in
which one merely believes blamelessly because one is not responsible for
one’s belief at all.
Third, the exceptional often throws light on the normal.16 It seems
that to be excused is the exceptional situation. It is not the case that we
act responsibly because each moment we are excused for violating all
sorts of obligations. Rather, we act responsibly because we only inciden-
tally violate certain obligations and in a few cases are excused for doing
so. Thus, to be blameworthy or to be excused seems to be the abnormal
situation. A discussion of being doxastically excused (the abnormal situ-
ation) can teach us important lessons about what it is to believe respon-
sibly (the normal situation). More specifically, from my discussion of the
doxastic excuses of force, ignorance, and luck, I conclude respectively
16. Th is is rightly pointed out by Austin (1979, 179–8 0) and Houlgate (1968, 116).
10
11
that responsible belief requires that one could have failed to have that
belief, that whether one believes responsibly depends essentially on
one’s factive and normative beliefs about one’s situation and is in that
sense radically subjective, and that whether or not one believes respon-
sibly is inevitably to some extent a matter of luck but that that does not
undermine our responsibility for many of our beliefs.
I.5 O UTLINE
This book is structured as follows. In chapter 1, I provide analyses of

responsibility and belief. Getting clear on the basic issues will turn out
to be of great help in the trenches. I argue that to be responsible is to
be the proper object of praise, blame, or neutral appraisal, and that to
believe some proposition p is to occurrently, dormantly, or tacitly think
that p. Thus, whether or not one is responsible for a belief that p is a mat-
ter of whether one can be appropriately appraised for thinking that p.
But then what is it to believe responsibly? Is it to believe praiseworthily or
merely blamelessly? I argue for the view that responsible belief is blame-
less rather than praiseworthy belief.
Having laid the foundation, in c hapter 2 I give a detailed version
of the main problem that an account of responsible belief faces: it
seems both that doxastic responsibility requires doxastic control
and that we lack such control. I argue against four solutions to this
problem: that according to which we have direct doxastic control, that
according to which we have indirect doxastic control, that according
to which we have compatibilist doxastic control, and that according
to which having a doxastic obligation requires no doxastic control
whatsoever.
This leaves us with only one option, namely, to explain doxastic
responsibility in terms of influence. On this approach, we are doxasti-
cally responsible in virtue of our control over states of affairs that make
a difference to what we believe rather than over our beliefs themselves.
In c hapter 3, I defend this approach by giving an account of what belief-
influencing factors we control and what obligations we have concerning
those belief-i nfluencing factors. As I said above, I call these “intellectual
11
12
obligations.” Also, I argue that, roughly, one responsibly believes that p

if and only if one had control over belief-i nfluencing factors pertaining
to one’s belief that p and either one did not violate any intellectual obli-
gations in coming to believe that p or one is excused for doing so.
In chapter 4, I give an account of force as a doxastic excuse. I argue
that to be forced to something is for it to be the case that one’s will does
not make a difference to it. Our will can make a difference to many of
our beliefs, since we are able to perform or not perform certain belief-
influencing actions. Hence, there is ample room for doxastic responsibil-
ity. My account of force as a doxastic excuse implies that responsibility
for believing that p requires that one could have failed to believe that
p. I discuss three objections against this claim. First, one might think
that the adherent of this thesis overlooks our control over properties of
beliefs, such as whether or not they are reasonable. Second, one could
object that this thesis conflicts with our intuitions in doxastic analogues
of Frankfurt-t ype cases, that is, scenarios in which one cannot fail to do
what one does but in which one seems nevertheless responsible. Third, it
may be suggested that the thesis is true for doxastic blameworthiness, but
not for doxastic praiseworthiness. I argue that each of these objections is
wanting.
Chapter 5 provides an account of ignorance as a doxastic excuse.
First, I explore what ignorance is and what varieties of ignorance there
are. Next, I distinguish four kinds of ignorance that can provide a nar-
row doxastic excuse by removing only doxastic blameworthiness: igno-
rance of one’s having the relevant intellectual obligation, ignorance of
one’s ability to meet that obligation, ignorance of how to meet that obli-
gation, and lack of foresight regarding a future intellectual obligation.
Finally, I argue that only blameless ignorance provides a doxastic excuse.
This fact does not lead to an infinite regress, but we will have to meet
the regress worry in a different way than philosophers such as William
FitzPatrick and Michael Zimmerman do. It follows from what I argue
in this chapter that whether one believes responsibly is fundamentally a
subjective matter in that it crucially depends on what one believes about
one’s situation and about one’s intellectual obligations.
In c hapter 6, I consider the doxastic excuse of luck, for, as has been
widely acknowledged, luck sometimes excuses. I argue that for an event
12
13
to be lucky for some person is for it to be beyond her intentional control,

of significance to her, and such that it could easily have failed to occur.
Subsequently, I defend the claim that there are four kinds of luck that
imply or often lead to doxastic luck: mechanism luck, evidential luck,
aretaic luck, and consequential luck. I argue that to the extent that they
excuse, they are reducible to force or ignorance—t hat is, in those cases
luck is an instance of force or ignorance. I discuss consequential luck in
some more detail, because the possibility of consequential luck shows
that any analysis of responsible belief has to take into account that one
is blameworthy for a belief only if it is non-accidentally related to the
violation of one’s intellectual obligation. Finally, I discuss and provide a
solution to a problem that arises from what I have argued. For it seems
that we are sometimes blameworthy for what we believe, that doxastic
luck is incompatible with doxastic blameworthiness, and that what we
believe is virtually always a matter of luck. I argue that, even though
what we believe is to a significant extent a matter of luck, that does not
undermine our responsibility for our beliefs.
In the Appendix, I explore the relation between responsible belief
and epistemically justified belief. I argue that responsible belief is not
identical to epistemically justified belief, but that it is much more similar
to it than is often thought.
There is much more that I would like to say on responsible belief,
especially in its relation to commemoration, religious disagreement,
and fundamentalism.17 Hopefully, what I argue in this book provides a
good starting point to engage these other issues.
17. A further issue is whether responsible belief is permissible or obliged belief. With Tony
Booth, I have defended the former view. See Booth and Peels (2012) and Peels and
Booth (2014).
13
14
[1]
DOXASTIC RESPONSIBILITY
What Is It?
1.1 I NTRODUCTION
Some of the difficulties that plague an account of responsible belief

already show up at the foundations. For, contrary to what some people
might expect, the core notions of responsibility and belief are them-
selves highly elusive. The literature contains several competing accounts
of what responsibility and belief are and it is hard to articulate these
notions crisply. Now, as we will see in the ensuing chapters, especially
chapter 2, we can only accurately assess certain views and arguments
in the debate on doxastic responsibility if we have a fairly firm grasp of
both responsibility and belief. The main aim of this chapter is, there-
fore, to tighten our grip on important aspects of these notions, so that it
becomes clear what it is to be doxastically responsible.
The chapter is structured as follows. First, I argue that to be respon-
sible for φ-ing is to be the proper object of such normative attitudes as
praise, blame, and neutral appraisal for φ-i ng. I also respond to objections
against this Appraisal Account of responsibility and argue that it is more
convincing than its rivals (§1.2). Thus, to be doxastically responsible is
to be the proper object of praise, blame, or neutral appraisal for one’s
belief. Now, there is much to be said about what belief is. For the pur-
poses of this book, I only consider the different ways in which one can
hold a belief and argue that there are three ways in which one can believe
some proposition p: one can occurrently, dormantly, or tacitly think that
14
15
D oxastic R esponsibility: W hat I s I t ?
p. For reasons that will become clear, I call this claim the Combination
Account of Belief (§1.3). Finally, when applied to doxastic responsibil-
ity, my Appraisal Account of responsibility naturally gives rise to the
following question: since responsible belief clearly is not blameworthy
belief (belief for which one is blameworthy), is it praiseworthy or rather
blameless (neutrally appraisable) belief? I provide three theoretical con-
siderations for the view that responsible belief is blameless rather than
praiseworthy belief and respond to three arguments to the contrary
(§1.4). I conclude that to responsibly believe that p is to be the proper
object of positive or neutral appraisal for occurrently, dormantly, or tac-
itly thinking that p.
1.2 R
ESPONSIBILIT Y AND APPRAISAL
What is it to be responsible? Before answering this question, let me be

a bit more precise about the question itself. In asking what it is to be
responsible, I mean to ask which informative conditions are necessary and
sufficient for being responsible. A set of conditions is informative if we
have a fairly firm grasp of the terms used in them and if they tell us some-
thing about the analysandum that is important for the issue at hand, in
this case, what it is to be responsible. For example, being the property
that I am now thinking about in the actual world may be necessary and
sufficient for being the property of being responsible, but it is not infor-
mative, since it does not tell us much about what it is to be responsible.
My Appraisal Account in terms of blame, praise, and other attitudes that
I specify below is informative, for if responsibility is essentially related
to such attitudes, then that tells us something important about responsi-
bility. My account is similar, but not identical, to that of Peter Strawson.
Below, I explain how it relates to Strawson’s account of responsibility.
Let me explain why I need an account of responsibility. In chapter 2,
we will consider scenarios in which people have various kinds of control
over or influence on their doxastic attitudes. In each case, a crucial ques-
tion is whether the relevant kind of control suffices for being respon-
sible for the doxastic attitude in question. We will be in a much better
15
16
position to answer that question if we have first acquired a firmer grip

on the notion of responsibility. The conditions for being responsible that
my Appraisal Account provides give us this firmer grip.
As suggested already by some of the more methodological comments
in the Introduction to this book, analyzing responsibility may require us
to pay close attention to different concepts of responsibility and to the
meaning of the word “responsible.” I think this is indeed the case when
it comes to responsibility and that there are different reasons for this.
First, we have a number of different, but related, concepts of responsibility.
There is the concept of causal responsibility: roughly, bringing about or
contributing to the occurrence of some state of affairs. There is the con-
cept of ministerial responsibility: one’s being accountable for the pro-
fessional actions and omissions of certain officials in one’s company or
department. There is the concept of legal responsibility: one’s being the
proper object of legal treatment in the form of punishment or its absence.
And there is the concept of role responsibility: having certain tasks or
duties in virtue of our role or profession. In fact, people’s obligations are
often referred to as their “responsibilities.” Clearly, these concepts sig-
nify different, although interconnected, kinds of responsibility.1
These kinds of responsibility display interesting relations to but are
nevertheless distinct from what I call normative responsibility, where
I use the word “normative” in a somewhat narrower sense than it is often
used. It is hard, if not impossible, to specify this variety of responsibility
in a theoretically neutral fashion, that is, in a way that does not commit
one to a particular account of responsibility. Giving examples might be
the best I can do here. Florence Nightingale was normatively respon-
sible for saving thousands of lives in the Crimean War. Adolf Eichmann
was normatively responsible for the deportation of millions of Jews to
concentration camps around Europe. And I am normatively responsible
for the amount of work I do on an average day and for whether I take
coffee or orange juice for breakfast. Moral responsibility falls under—
is a variety of—normative responsibility, but it is not the only kind of
normative responsibility: there also seem to be such things as prudential
1. Thus also Hart (1970, 211–2 30).
16
17
responsibility and, especially important for our purposes, epistemic

responsibility. In what follows, I am concerned solely with normative
rather than causal, ministerial, legal, or role responsibility, unless indi-
cated otherwise. We will consider cases of moral, prudential, and epis-
temic normative responsibility for our beliefs.
Paying attention to different concepts of responsibility, then, is valu-
able, especially in singling out normative responsibility from other
kinds of responsibility. Considering the semantics of responsibility
is equally important. Let us call a judgment that says or implies that
someone is normatively responsible a “normative judgment.” Then, we
should notice that in normative judgments the word “responsible” is
ambiguous. We can use the word in a neutral, approbative, or pejorative
sense. That is, in saying that someone is responsible, one can mean that
that person is the proper object of (i) some kind of appraisal, whether
positive, negative, or neutral, (ii) merely positive or neutral appraisal, or
(iii) merely negative appraisal. Compare the following sentences:
(a) “You are responsible(i) for what you say.”

(b) “Students are encouraged to take responsible(ii) action in an
emergency.”
(c) “James is responsible(iii) for his own misery.”
Sentence (a) may well be uttered to express the idea that you are the
proper object of some sort of appraisal for your speech acts, without
specifying which sort of appraisal. Judgment (b) may be made to say
that students are encouraged to meet their obligations in emergency
situations and, thereby, act praiseworthily or at least not blameworthily.
Finally, (c) may be uttered to say that James is to blame for his misery.
In order to avoid confusion, let me be explicit that, henceforth, I use
the noun “responsibility” and the expression “responsible for X” only
in sense (i), that is, as expressing neutral appraisal, whereas I use the
adverb “responsibly” and the expression “responsible X” only in sense
(ii), that is, as positive appraisal. I will not use “responsible” in sense (iii),
that is, as negative appraisal.
To be responsible, I will argue, is to be the proper object of one or
more normative attitudes, such as praise, blame, and neutral appraisal.
17
18
But what kind of attitudes are they? Normative attitudes, perhaps with
the exception of neutral appraisal, are affective attitudes that we adopt
primarily toward people on the basis of their actions, desires, beliefs,
virtues and vices, and character. We adopt them toward other people,
but also toward ourselves. Thus, we can be angry at someone’s decision
to remain silent and I can feel remorse about what I said to my friend last
night. Normative attitudes are as varied as blame, praise, resentment,
outrage, gratitude, forgiveness, indignation, respect, compunction, and
remorse.
Normative attitudes are to be distinguished from another kind of
attitude that we sometimes adopt, an attitude that I call the (merely)
evaluative attitude. Varieties of the evaluative attitude are pity, certain
kinds of love, certain kinds of shame, deeming inappropriate, consider-
ing harmful, and so forth. Some evaluative attitudes are affective atti-
tudes, whereas others are not. The crucial difference between normative
and evaluative attitudes is that in adopting the former, we hold some-
one responsible, whereas in adopting the latter we merely take it that
the occurrence of some state of affairs was desirable, undesirable, or of
neutral worth, not that someone is responsible for it. This is not to say
that if we adopt an evaluative attitude toward someone, we do not hold
that person responsible. After all, we can adopt both a normative and an
evaluative attitude toward someone for something. It is to say, however,
that in adopting an evaluative attitude, we do not thereby hold someone
responsible, whereas if we adopt a normative attitude we do thereby hold
that person responsible.
Talk of attitudes such as praise and blame will probably remind the
reader of Peter Strawson’s account of responsibility in terms of what he
calls “reactive attitudes.” I take it that Peter Strawson’s reactive attitude is
identical to my normative attitude and that his objective attitude is identi-
cal or at least similar to my evaluative attitude.2 I prefer the expression
“normative attitude” to “reactive attitude,” since we can react to people in
all sorts of ways without holding them responsible. Using the term “nor-
mative” is more focused in that it restricts the relevant attitudes to those
2. See Strawson (1974, 4–13).
18
19
by which we hold people up to certain norms. I prefer the term “evalua-

tive” to “objective,” because evaluations are often highly subjective.
Let me bring the notion of normative attitude, which is crucial to my
Appraisal Account of responsibility, more sharply into focus by way of
the following four comments.
First, when I say that someone is the proper object of positive, negative,
or neutral appraisal for φ-ing, I mean that that person deserves or merits
positive, negative, or neutral appraisal for φ-ing. I take this to be an objec-
tive notion: if someone deserves praise, blame, or neutral appraisal, then
such appraisal will be the normative attitude that would or, at least, could
be adopted by a physically, mentally, and emotionally well-functioning
person who is fully informed about the situation, that is, who knows every
truth about it. This means that the notion of merit or desert is not to be
spelled out in terms of what a person with justified, rational, or warranted
beliefs about the situation would do. For a person with merely justified,
rational, or warranted beliefs may mistakenly hold someone responsible,
since justification, rationality, and warrant do not entail truth. It also
means that a fully informed and properly functioning human being need
not always actually adopt a normative attitude toward some person S for
something for which S is responsible. For instance, in considering the
Nazi crimes mentioned in Simon Wiesenthal’s monograph Justice, Not
Vengeance, she may be so overwhelmed by their sheer number that she
will not have particular emotions toward each of them. However, being
fully informed and properly functioning, she could adopt such normative
attitudes. Whether she will in fact adopt some normative attitude or not
depends on all sorts of situational factors. To be the proper object of one or
more normative attitudes for φ-ing, then, is to deserve appraisal for φ-ing.
Second, my Appraisal Account does not imply that it is appropri-
ate for anyone to express a normative attitude in a particular situation.
If I am fully informed about the crime that my brother and I jointly
committed, it may be appropriate for me to adopt the attitude of blame
toward my brother, but it might be inappropriate for me to condemn him
verbally. 3 After all, we jointly committed the crime and we are equally
3. Th is is also pointed out by Feinberg (1970, 127), Watson (1993, 145).
19
20
culpable. Also, that S is the proper object of normative attitudes does

not mean that anyone actually takes a normative attitude toward S. We
can be responsible for something, even if no one else is aware of it and if,
therefore, no one adopts a normative attitude toward us.4
Third, one might wonder how we should distinguish normative from
evaluative attitudes. It seems to me, though, that a criterion for distin-
guishing normative from evaluative attitudes is not needed. In principle,
we could provide an account of responsibility by including a long, pos-
sibly exhaustive, list of normative attitudes. And if that list correctly
picks out those attitudes that we adopt in holding people responsible,
then there is no theoretical need to add a criterion to distinguish them
from evaluative attitudes. For instance, it seems intuitively obvious that
in resenting or blaming someone, we hold her responsible, whereas in
pitying someone, we do not. And it is contradictory to say that someone
is blameworthy for her bad situation but that she is not responsible for it,
whereas it is not contradictory to say that someone is to pity for her bad
situation but that she is not responsible for it.
Fourth, how should we think of the normative attitude of neutral
appraisal? I am responsible for the order—or, I am afraid, disorder—i n
my study, for dressing as I do, and for having orange juice rather than cof-
fee for breakfast, but I am not the proper object of praise, blame, anger,
or any such strong normative attitude for these things. 5 It is usually
merely fine or all right if I take orange juice rather than coffee for break-
fast. This, however, raises the following question about the attitude of
neutral appraisal. Imagine that you know that, due to a mental disease,
such as Alzheimer’s, I take orange juice rather than coffee for breakfast,
say, because I always had orange juice for breakfast when I was young.
In this case as well, my taking orange juice rather than coffee is fine or
all right to you. After all, we assume, there is nothing wrong or harmful
about it. But then how are we to distinguish the normative attitude of
neutral appraisal from the evaluative attitude of neutral evaluation? It
seems to me that one way to do so is the following. If I take the normative
4. Th
us also Zimmerman (1988, 38–39).
5. It is sometimes overlooked that neutral appraisal is one of the normative attitudes. See,
for instance, Bennett (1980).
20
21
attitude of neutral appraisal toward someone for φ-ing, then I could in

principle also blame or praise her for φ-i ng. Thus, if I know that I am aller-
gic to orange juice and that I will need to be rushed to the hospital if
I drink orange juice, then I will, in normal circumstances, be blamed if
I do. But if I suffer from Alzheimer’s and I am allergic to orange juice,
then I am not blameworthy if I take it. Since the person who suffers from
Alzheimer’s is not in principle blameworthy for taking orange juice, she
is only the proper object of an evaluative rather than a normative attitude
for doing so. However, someone who is perfectly healthy is in principle
blameworthy for taking orange juice and, therefore, the proper object of
a normative attitude for doing so.
So far, I have argued for the following Appraisal Account of respon-
sibility: S is responsible for φ-i ng iff S is the proper object of at least one
of the normative attitudes of positive, negative, or neutral appraisal for
φ-i ng. This account is clearly similar to the widely influential account of
responsibility that Peter Strawson provided in his essay “Freedom and
Resentment.” In that essay, he defends an account of responsibility in
terms of what I call normative attitudes. 6 However, in two regards, my
Appraisal Account differs from that of Strawson.
First, Strawson seems to address the question of when it is legiti-
mate to hold someone responsible, but he does not specify what it is to
be responsible. This is because he is interested in the question whether
we should give up the normative attitudes if determinism turns out to
be true. He argues, among other things, that it is practically inconceiv-
able that we do, because our participation in interpersonal relation-
ships is simply too deeply rooted. My Appraisal Account, however, is
an account of what it is to be responsible. It says that one is responsible
not if someone adopts a normative attitude toward one, but if one is the
6. See Strawson (1974). For a similar view prior to Strawson’s essay, see Brandt (1958, 9–17).
After Strawson, we find a wide variety of appraisal views. See, for instance, Fischer and
Ravizza (1993a, 4–25; 1998, 5–8), Gibbard (1990, 126–150), Scanlon (1998, 248–294),
Wallace (1994, 2), Watson (1993, 119–148), Wolf (1993), Zimmerman (1988, 38–73).
However, most of these accounts are much less developed than that of Strawson. They
merely say that responsibility is to be understood in terms of normative attitudes like
praise and blame, without spelling out what these normative attitudes amount to and
when one is the proper object of these attitudes.
21
22
proper object of normative attitudes, and I have briefly spelled out when
someone is a proper object of normative attitudes.
Second, in addition to positive and negative appraisal, I distinguish
neutral appraisal as a normative attitude, an attitude that Strawson fails
to mention. It is important to distinguish this as a distinct normative
attitude in addition to such attitudes as praise and blame. For, in many
cases in which someone is responsible for something, such as my tak-
ing coffee for breakfast, there is nothing blameworthy or praiseworthy
about the action in question. Below, we will see that this also applies to
our beliefs: many of them are such that we are neither blameworthy nor
praiseworthy, but nonetheless responsible for them.
Let me stress that these are primarily elaborations of rather than
departures from Strawson’s view. One could, in principle, adopt the view
Strawson advocates and add that someone is responsible only if one is
the proper object of normative attitudes and that neutral appraisal is one
of those attitudes.
Unsurprisingly, accounts along the lines of the Appraisal Account
that I spelled out above have been criticized. Here, I will defend my
Appraisal Account against what I consider to be three important
objections to it.
First, some philosophers, such as Marina Oshana, have objected
that an appraisal view cannot offer an informative account of some per-
son S’s being responsible, since it focuses on some other person S*’s atti-
tudes toward the cognitive subject S rather than on S herself.7 I find this
objection unconvincing. The Appraisal Account tells us that S is respon-
sible for φ-ing just in case S is the proper object of some normative atti-
tude for φ-i ng. Surely, this tells us something important about S herself,
namely that she is properly appraisable for φ-i ng. In this regard, there is
nothing exceptional about responsibility. To give one other example, it
is hardly possible fully to explain someone’s being admirable without
appealing to the attitudes of others, namely their attitudes of admira-
tion, for it seems that to be admirable is to be deserving of admiration by
some person—whether an actual or imaginary person. Clearly, it does
7. See Oshana (1997, 80).
22
23
not follow that an analysis of being admirable in terms of one’s relation

to someone else is uninformative. If we spell out what it is to admire
someone and then say that someone is admirable just in case she is the
proper object of the attitude of admiration, then that tells us something
important about what it is to be admirable.
Second, one may object to appraisal accounts that certain emo-
tions, such as anger, are always improper and that more objective or
more detached attitudes are called for than normative attitudes, which
are typically affective attitudes after all. It seems to me that this objec-
tion fails for two reasons. First, even if some normative attitudes, such
as anger, are always inappropriate, it does not follow that all of them
are. Attitudes like praise and indignation may still be appropriate. The
range of legitimate normative attitudes would be smaller, but responsi-
bility could still be cashed out in terms of normative attitudes. Second
and more importantly, there is good reason to think that the claim that
certain emotions, such as resentment, are always improper is simply
false. A father of a raped daughter properly feels resentment toward the
offender, and if a husband cheats on his wife, while she is loving and
devoted, she properly feels anger toward him for cheating on her. In fact,
it seems not only that these attitudes are proper in the situations just
described, but that, unless special conditions hold, it would be improper
not to have them. I conclude that this second objection also fails.
Third, one might object that the account is circular. Someone is
responsible for φ-i ng just in case she is the proper object of one or more
normative attitudes for φ-ing. One is the proper object of a normative
attitude for φ-ing if a fully informed and properly functioning person
could adopt that attitude toward one for φ-ing. But when could such a
person adopt a normative attitude toward one for φ-i ng? Well, given that,
in adopting such an attitude toward someone for φ-i ng one assumes that
that person is responsible for φ-ing, it seems that a fully informed and
properly functioning person would do so only if the person in question
is responsible for φ-i ng. But then we are back where we started, namely
the question of what is it to be responsible for something. This objec-
tion fails as well. As I said in the Introduction to this book, an analysis
need not be reductive in order to be illuminating. If we have a fairly firm
grip on normative attitudes such as praise and blame and we can spell
23
24
out what it is to be responsible in terms of such attitudes, then such an

account provides valuable insight, even if it does not follow that respon-
sibility is reducible to such normative attitudes.
In the literature, we find four important views about responsibil-
ity that conflict at least prima facie with my Appraisal Account. Since,
as we shall see in chapter 2, at least some of them in certain scenarios
give different verdicts from those given by my Appraisal Account as
to whether or not someone in particular circumstances is doxastically
responsible, it is important to discuss them. Here, I will only single out
the main reasons for thinking that to the extent that they differ from my
Appraisal Account—t hey will agree on a large number of cases—t hey
are unconvincing.
First, on what has been called the behaviorist account, to be respon-
sible is to be the proper object of punishment or reward. On this view, one
is responsible to the extent that punishment and reward will increase the
likelihood of good behavior and decrease the likelihood of bad behavior
in the future. This account, which has been advocated most famously
by Moritz Schlick,8 seems implausible for at least two reasons. First, we
may punish and reward animals and young children if we believe that
we can thereby train them to perform certain beneficial actions or not
perform certain harmful actions. Yet we often do not hold them respon-
sible. Rather, we consider them as beings to be manipulated or treated
in order to bring about some good or to avoid some harm. Second, there
are circumstances in which we hold people responsible, but consider
punishment or reward inappropriate. We hold our friends responsible
for not telling what they consider to be the truth, but only in judicial
circumstances are people punished for such a thing. Blame rather than
punishment seems to be the appropriate reaction in such cases.9 Thus,
being the proper object of punishment or reward is neither sufficient nor
necessary for responsibility.
8. See Schlick (1962, 151–158).

9. Glover (1970, 50–5 4), Scanlon (1998, 263–2 67), and Wallace (1994, 52–62) also distin-
guish between being responsible on the one hand and being the proper object of punish-
ment and reward on the other.
24
25
Second, some philosophers think that to be responsible is to be

accountable in the sense that it is appropriate for one to explain one’s
motivation or reasons. The word “responsible” derives, after all, from
the Latin respondere, which means “to answer.” According to Marina
Oshana, for instance, “a person is responsible for an act if and only if
it ought to be the case that the person account for her behavior, where
doing so involves giving some statement of the person’s beliefs or inten-
tions regarding the act.” Unfortunately, etymology can be misleading.
The English verb “to believe” derives from the German belieben (“to
love”), but belief is clearly not essentially linked to emotions or passions.
Moreover, it seems that accountability in this sense is neither necessary
nor sufficient for responsibility. If I take orange juice rather than cof-
fee for breakfast—something for which I am clearly responsible—it
does not seem appropriate that I account for that to anyone, not even
to my own conscience. Oshana admits that sometimes it is not appro-
priate to call for an account, but rather to state one’s point of view on
one’s φ-i ng.10 But if I drink orange juice rather than coffee, am I really
required to state my point of view about that to anyone? That seems
rather implausible. The issue seems even more pressing for praiseworthy
rather than neutrally appraisable actions. It would often be inappropri-
ate to state one’s point of view on one’s excellent sport achievements or
brilliant academic accomplishments, even though one is responsible for
them. Nor is accountability sufficient for responsibility. We sometimes
need to account for our actions in cases in which we are not responsible.
Imagine that a sudden epileptic seizure makes me kick my colleague,
while I am the only one who knows the cause of this movement. Then,
it seems, I need to account for hurting my colleague, even if I am not
responsible for doing so.11
On a third account of responsibility, the so-called ledger view, to
be responsible for φ-ing is to have a negative or positive mark on one’s
record as a person. Positive marks add to one’s credit, whereas negative
10. S ee Oshana (1997, 77–8 0).

11. It is not helpful to claim, as Oshana does (1997, 80), that the accountability view runs
for the usual sorts of cases, for we are looking for conditions that are necessary and suf-
ficient for responsibility.
25
26
marks add to one’s debit. To be responsible is to have positive or negative

moral, prudential, or epistemic value as a person. We find this view in
the works of, among others, Joel Feinberg and Jonathan Glover.12 This
account faces at least two problems. First, I am responsible for whether
I drink two or three cups of coffee a day, but doing so does not seem to
add anything to my moral, prudential, or epistemic debit or credit, or,
more generally, to my value as a person. Adherents of ledger views do
not address this worry, but there may be a way to meet this worry. For
one could adopt the same strategy that I adopted above in distinguish-
ing neutral appraisal as a normative attitude from neutral evaluation as
an evaluative attitude. Applied to the issue at hand, the slightly revised
version of the ledger view would then say that one is responsible for
φ-i ng just in case one’s φ-i ng is such that in principle it could add to one’s
credit or debit. There is a second and more important worry, though.
This account, even if it is correct, is not really informative. For what is it
to have moral value or disvalue? These notions are so vague that, unless
more is said, the ledger view cannot function as an informative analy-
sis of responsibility, and that is what we need in order to assess certain
views on doxastic control and doxastic responsibility, as we will do in
chapter 2. I do not say that some such thing cannot be done. All I am
saying is that, until this account is developed in more detail, we are bet-
ter advised to work with my Appraisal Account than with ledger views.
Fourth, one might think that to be responsible is to have a particular
kind of control or influence. Such control or influence can be spelled
out in terms of alternative possibilities, reason-responsive mechanisms,
second-order desires, or some such thing. Let us call this view the control
view. One might think that to be responsible for φ-ing is to have some
kind of control over φ-ing and that it is because we have such control
that we are sometimes the proper objects of praise, blame, or neutral
appraisal for φ-ing. In opposition to the behaviorist and accountabil-
ity views, the control view is perfectly compatible with the Appraisal
Account that I have defended. The control view and the Appraisal
Account are rivals only if they are understood as accounts of the nature
12. S ee Feinberg (1970, 30–31) and Glover (1970, 64–65).
26
27
or essence of responsibility.13 Let me explain what I mean by this by way

of an illustration. Being three-a ngled can properly be analyzed in terms
of being necessary and sufficient for being trilateral. Such an analysis
is also informative: it tells us something important about the property
of being three-a ngled. Nevertheless, it seems that the property of being
triangular is distinct from the property of being trilateral: being trian-
gular and being trilateral are not the same thing, even though these
properties are necessarily co-exemplified. Now, it may well be that to
be responsible is to have control and that that property is necessarily
co-exemplified with the property of being the proper object of appraisal.
Whether or not that is the case, my Appraisal Account is to be preferred
for the purposes of this book, for it provides us with informative neces-
sary and sufficient conditions for responsibility that can be used as an
independent criterion for establishing whether someone with a particu-
lar kind of control over or influence on some belief is responsible for that
belief or not. As I pointed out at the beginning of this section, we need
such an independent criterion for the assessment of certain views about
the relation between doxastic responsibility on the one hand and doxas-
tic control on the other that will be discussed in c hapter 2.
It follows, from the Appraisal Account of responsibility that I have
defended, that to be responsible for some belief is to be the proper object
of positive, negative, or neutral appraisal for that belief. This is crucial, for
many doxastic judgments—by that, I mean judgments on people’s beliefs—
are only evaluative rather than normative. Expressions such as “ought to” and
“should not” are used in such judgments merely to evaluate some person’s
belief, not to hold her responsible for it. Consider the following sentences:
(d) “When asked what she sees, an adult and properly function-
ing human being ought to believe that she is seeing a chair
when she is looking straight at one.”
13. To provide an account of the nature or essence of something is to give what Ralph
Wedgwood calls a “constitutive account” of that thing. See Wedgwood (2007, 135–
152). The distinction between the nature of responsibility on the one hand and mere
necessary and sufficient conditions for responsibility on the other has been noticed, but
not elaborated in any detail, by Bennett (1980, 20) and Wallace (1994, 1–17).
27
28
(e) “If Judith is to win the quiz, she should believe that Jupiter is
the largest planet in our solar system.”
(f) “Since Hercule Poirot did not return to Miss Burgess’ house,
he must have believed that she was not culpable.”
(g) “If she knows that the polar bear was reclassified as a vulner-
able species in 2005, then she has to believe that.”
In most circumstances, these will not be normative doxastic judgments,

that is, judgments by which we hold people responsible for some belief.
Normally, (d) will be uttered to describe proper doxastic functioning,
(e) to say what belief is instrumentally useful or necessary for attaining
some social good, in this case winning a quiz, (f) to tell what belief some-
one can reasonably be expected to have given one’s own evidence regard-
ing that person’s behavior, and (g) to characterize the conceptual relation
between knowledge and belief. In uttering these sentences, we need not
hold the cognitive subject in question responsible for her belief. In fact,
if the circumstances for (d)–(g) are as I just described, then in uttering
these sentences, one does not take a normative attitude toward the person
for holding or not holding a particular belief. This is important, for we will
see in c hapter 2 that normative doxastic judgments are all too easily con-
fused with merely evaluative doxastic judgments.14 My Appraisal Account
of responsibility is meant to help us in keeping the two apart.
1.3 O
CCURRENT, DORMANT,
AND TACIT BELIEF
Now that we have tightened our grip on the notion of responsibility that
we are working with, let us turn to the other crucial phenomenon, that
of belief. Like responsibility, however, belief seems to have certain vari-
eties. Let me, therefore, stress that I am concerned with propositional
14. I n this regard, normative and evaluative doxastic judgments are just a special case of
normative and evaluative judgments generally. The same point is made about the
latter—normative and evaluative judgments generally—by Anscombe (1963a, 64–65).
28
29
belief, that is, belief that something is true or that something is the case,
rather than such phenomena as believing my neighbor and belief in
democracy.15 We should also notice that the term “belief,” even if used
only for propositional belief, is ambiguous. Depending on the context,
it denotes either believing something or that which is believed. The former
can be rational, unjustified, hastily formed, or reasonable, whereas the
latter can be true, credible, or in contradiction with some other propo-
sition.16 In the ensuing chapters the context of the discussion should
sufficiently indicate which sense of “belief ” I have in mind. Where the
discussion is likely to get confused, I state expressis verbis how the term
is meant to be understood.
The literature on belief is, of course, vast, but, fortunately, most of
it need not concern us here. I take it that it is fairly widely accepted that
to believe a proposition p is to think that p—a nd, often, to think that p
is the case or that p is true. At least all the philosophers whose ideas on
doxastic control I discuss in c hapter 2 share this view. Things become
more controversial when we ask under which conditions one is sup-
posed to think that p in order to count as believing that p. It is clear that,
normally, if I now consciously consider p and think that p, I believe that
p. But such occurrent belief does not seem to exhaust the possibilities
for holding a belief.
Before we explore further ways in which one can hold a belief, how-
ever, let me say a few words on why I think it is important to spell out
different ways of thinking that p that plausibly count as believing that
p. The main reason is, of course, that it is helpful for a theory of respon-
sible belief to get a firmer grip on the conditions under which one holds
a belief. But, apart from that, there are at least three additional reasons
to consider this issue.
15. Th is is not to deny that both believing someone or something and belief in someone
or something may entail certain propositional beliefs. In fact, elsewhere—see Peels
(2010b, 97–103)—I have argued that certain kinds of belief-in entail propositional
belief.
16. Thus also Alston (1996, 3), David (2005, 297), Hieronymi (2006, 67n), Schellenberg
(2005, 40).
29
30
First, it is widely thought that there are three main doxastic atti-
tudes: belief, disbelief, and suspension of judgment (that is, suspen-
sion of both belief and disbelief). Most philosophers take disbelief
simply to be belief that ~p. If that is correct, then disbelief is a variety
of belief, a particular kind of belief. But it is not directly clear how we
should think of suspension of judgment. On the one hand, we do not
suspend judgment on just any proposition that we fail to believe and
fail to disbelieve. An average human being does not suspend judg-
ment on the Brouwer fixed point theorem, which states that every
continuous function from the closed unit ball Bn to itself has at least
one fixed point—it seems that most people have never even consid-
ered this proposition. On the other hand, saying that one suspends
judgment on p only if one has considered p seems too demanding. If
I suspend judgment on whether or not Sally is at her work, I will nor-
mally also suspend judgment on whether or not Sally is at her desk,
even if I have not explicitly considered that proposition. As we shall
see below, a plausible analysis of belief also sheds light on suspension
of judgment. It is important to get a firmer grip on suspension of judg-
ment as well, because we not only hold each other responsible for our
beliefs and disbeliefs, but also for suspending judgment. Sometimes,
we think that someone is blameworthy for suspending judgment, for
instance, because she has not informed herself while she should have.
And sometimes suspension of judgment is the responsible doxastic
attitude to take, as when we suspend judgment on the proposition that
the number of planets is even.
Second, we do not only hold each other responsible for our occur-
rent beliefs. If someone holds a heinous racist belief, I may properly
hold her responsible for that—she may be blameworthy for having that
belief—even if she is not currently considering a proposition. It may
even be plausible to argue that one can responsibly or culpably believe
propositions that one has never explicitly considered. If I responsibly
believe that John is less than seven feet tall, then it seems that, unless
special considerations hold, I also believe that John is less than eight feet
tall and responsibly so, and also that John is less than eighteen feet tall
and that I responsibly believe that. (For those who have qualms about
30
31
whether I truly believe these things rather than having a disposition to

believe these things, I return to this issue in more detail below.)
Third, distinguishing dormant and tacit belief in addition to occur-
rent belief will turn out to be of crucial importance in solving certain
problems that plague an account of responsible belief. To give just one
example: in chapter 5, I discuss the so-called regress problem for igno-
rance as a doxastic excuse. Briefly, the problem is this. It seems that one
is blameworthy for φ-ing only if one believes that one should not φ or
if one is culpably ignorant that one should not φ. Now, it seems one is
culpably ignorant that one should not φ only if there was something one
did and should not have done or if there was something one failed to do
but that one should have done such that if one had not done that (respec-
tively if one had done that), one would not have been ignorant. But these
same principles apply to the earlier occasion that led to one’s culpable
ignorance. It follows that one is blameworthy only if at some point one
φ-d while believing that one should not φ. If there is only occurrent
belief, then it seems that we will hardly ever be blameworthy for any-
thing whatsoever, since we hardly ever act from such clear-eyed akrasia
(performing an act while occurrently believing that one should not per-
form that act). If, however, there are also dormant and tacit beliefs, then
the regress might be avoided—or, at least, so I will argue.
The first way in which one can believe a proposition p is to occur-
rently believe it. Some person S occurrently believes that p at some time
t if and only if S at t is entertaining p, considering p, or thinking about p
(these expressions are meant as equivalents) and S at t thinks that p. This
is a common phenomenon. If I now think about whether Libya is larger
than Sudan and think that Libya is indeed larger than Sudan, then I now
believe that Libya is larger than Sudan. Thus, we can define Occurrent
Belief (OB) as follows:
OB: S occurrently believes that p at some time t iff S at t consid-

ers p and thinks at t that p.
A few philosophers, such as Radu Bogdan, have suggested that this is not
only sufficient for believing that p, but also necessary. It would follow
31
32
that all our beliefs are occurrent.17 I think that such a view faces insur-
mountable difficulties. For one thing, when I am sound asleep, I can
still correctly be said to believe that eating two pieces of fruit a day is a
healthy habit for an adult human being or that Napoleon died in 1821.
And I can also rightly be said to believe these things when I am teaching
a class on moral realism or canoeing down a wild river, thinking about
neither Napoleon nor healthy habits. Occurrent thought, then, is suf-
ficient, but not necessary for belief.
The second way in which one can believe that p at some time t is
to have a dormant belief (DB) that p. By that I mean, roughly, that one
has thought about p at some time in the past, that one thought at that
time that p is true, and that that belief is somehow stored in one’s mind.
Even if one does not currently consider p and, hence, does not occur-
rently believe that p, the belief is stored and can easily be activated, that
is, become occurrent. My beliefs that eating two pieces of fruit a day is
a healthy habit and that Napoleon died in 1821 are good examples of
this: I believe these things even when I do not think about them. In fact,
this seems true for a large number of our beliefs, given that there are
many things that we believe and that we have thought about, but that we
do not consciously consider all the time. This suggests that the second
way in which one can have a belief can be understood as follows:
DB: S dormantly believes that p at t iff (i) S at t does not consider

whether p, and (ii) the last time t* at which S considered
whether p, S thought that p.
As we will see below, DB needs further refinement. Before we provide

such refinement, though, let us first turn to the third and final way in
which, some philosophers have suggested, one can believe a proposition.
According to many philosophers, some of our beliefs are neither
occurrent nor dormant, at least, if we understand “occurrent” and
“dormant” along the lines of OB and DB. People believe that they are
17. S ee Bogdan (1986).
32
33
less than two miles tall, that they are not bats, and that 2 is a smaller
number than 999, even if they have never thought about these propo-
sitions. If you are not sure whether people believe these things, ask
whether they know these things. There, it seems, the answer is surely
positive. My neighbor knows that he is not a bat, even if has never
considered that proposition. But knowledge, most philosophers
claim, entails belief. Hence, people believe these things, even if they
have never entertained the propositions in questions. Let us call such
beliefs tacit beliefs.
One might try to take such beliefs into account by suggesting
that there is a third and final way in which one can believe a propo-
sition, namely if a proposition obviously follows from other proposi-
tions that one believes. But when does a proposition obviously follow
from other propositions one believes? I think that in explaining what
such “obvious entailment” amounts to, one would have to take at least
two important facts into account. First, it should be the case that the
proposition in question is entailed or rendered sufficiently likely by
other propositions one believes on one’s standards rather than objec-
tively. The mathematical propositions I believe may entail all sorts
of new mathematical theorems that nobody has ever thought of, but
it seems false to say that I believe these new mathematical theorems.
Thus, what follows obviously from certain propositions may differ
from person to person. Second, sometimes things follow from other
things that we believe, but we are somehow blind to them: we fail
to see them. Imagine that I read a detective novel. If I were to ask
myself whether the butler could have left the house, I would immedi-
ately realize that he could not, because the doors were locked and he
did not have the keys. But since that is something I do not ask myself,
I somehow fail to realize this. I do not believe that the butler could not
have left the house, even though it follows, even on my own standards,
from other things I believe. What we should say, then, is that some-
thing obviously follows by other things one believes if the truth of the
proposition that follows would not come as a surprise or discovery.
In what follows, when I say that some proposition obviously follows
from other propositions, it will be this rather stipulative meaning of
33
34
“obvious entailment” that I have in mind. This results in the following

analysis of Tacit Belief (TB):
TB: S tacitly believes that p at t iff (i) S has never considered

whether p, and (ii) p obviously follows from propositions that
S occurrently or dormantly believes at t.
Unfortunately, DB and TB are both problematic for at least two reasons.

First, imagine that one day, as a six-year-old child, Alfred occurrently
believes that his father is the strongest man in the world. Alfred does
not consider this proposition again for the next thirty years. It follows
from DB that when Alfred is, say, twenty-six, he believes that his father is
the strongest man in the world. That seems false. Moreover, it conflicts
with TB, for it follows from TB that Alfred, when he is twenty-six, does
not believe that his father is the strongest man in the world, because that
obviously follows from other propositions that he believes, such as the
proposition that his father’s health has always been pretty bad. This is
also a problem for TB, for on DB, Alfred now believes that his father is
the strongest man in the world, and since any proposition entails itself,
it also follows from clause (ii) of TB that Alfred as an adult believes that
his father is the strongest man in the world.
A second problem for TB is that the conditions given in the analysans
of DB seem to a certain extent arbitrary (and, given that DB returns in
TB, this problem also plagues TB). Imagine, for instance, that Rosemary,
upon considering some piece of evidence, finds herself with the belief that
her son betrayed her. However, when Rosemary encounters her son the
next day and looks him in the eye, she believes that he is innocent. Upon
rethinking the issue the day after, she again finds herself with the belief that
he betrayed her. Imagine that one Saturday night, when her attitude on this
issue has been wavering for months in the way just described, she watches
a late-night show and does not think of her son at all. Should we say that
Rosemary then believes that her son has betrayed her, just because that was
the most recent thought she had on the issue, even if she were to disbelieve
it upon considering whether he did betray her? That seems arbitrary.
In order to avoid such problems, some philosophers, such as William
Alston and Jonathan Cohen, have suggested that we should think of
34
35
belief as a disposition.18 The metaphysics of dispositions is, of course, a

matter of fierce controversy. Fortunately, we need not go into the details
here, for the following will do for our purposes: S has a disposition to φ if
and only if there are certain conditions such that if they obtain, S will φ.
Dispositions, then, are properties of some entity X that account, at least
on a surface level, for X’s φ-i ng in certain circumstances.19 The disposi-
tional view has often been wedded to some kind of behaviorism, where
behaviorism is the view that belief is a matter of how one acts or would
act in certain circumstances. There is no reason to think, however, that
a dispositional account of belief has to be behavioristic. One can just as
well provide a dispositional view of belief in terms of what one would
think in a certain counterfactual scenario.
But what should that counterfactual scenario look like? On the
one hand, it should not be identical to the situation in the actual world,
for then we would not need to appeal to dispositions: considering the
actual world would suffice to establish whether someone holds a par-
ticular belief. On the other hand, the relevant scenario should not be
altogether different from that in the actual world. If I had perceptual
evidence of being ambushed by a tiger, I would come to believe that a
tiger ambushes me. But I do not now believe that I am ambushed by
a tiger merely because I have such a disposition. Clearly, the relevant
scenario should be much more similar to the situation in the actual
world. Most adherents of the dispositional view, therefore, say that
the scenario should be identical to that in the actual world except for
the fact that in the counterfactual scenario one considers whether the
proposition in question is true. Of course, there are exceptional situa-
tions, circumstances in which one is distracted, in which one has severe
brain damage, or in which too many belief dispositions are activated at
the same time, so that even in the counterfactual scenario one does not
18. For accounts of belief in terms of a phenomenal disposition, that is, a disposition to
think, to feel, or to be appeared to in a certain way, see Alston (1996, 4–7), Cohen (1992,
4–6), Schellenberg (2005, 39–50). For dispositional accounts of belief that are partly
or entirely non-phenomenal, see, for instance, Ramsey (1931, 169–184), Braithwaite
(1967, 30–4 0), Ryle (1949, 128–130), Schwitzgebel (2002).
19. F
or a similar understanding of dispositions, see Prior (1985, 5–10).
35
36
occurrently believe the proposition in question. The word “normally” in

the following analyses is supposed to exclude such situations from the
relevant counterfactual scenarios:
DB′: S dormantly believes that p at t iff (i) S does not consider p

at t, but has considered p at some earlier time t*, and (ii) if
S at t were to consider whether p, S would normally think
that p.
TB′: S tacitly believes that p at t iff (i) S has never considered
whether p, and (ii) if S at t were to consider whether p, S
would normally think that p.
Now, let us see whether these revised versions of DB and TB fare better
than DB and TB themselves. The first thing to notice is that they do not
face the first problem that I mentioned above. Twenty-six-year-old Alfred
clearly lacks a disposition to believe that his father is or was the strongest
man in the world. Hence, he neither dormantly nor tacitly believes that.
Unfortunately, like DB and TB, DB′ and TB′ seem unable to meet the sec-
ond challenge. Imagine that the following is true. The last time Rosemary
considered whether her son is reliable, she occurrently believed that he
is not because he betrayed her. However, even though she has acquired
no additional evidence, if she were to consider that proposition now, she
would believe that he is reliable and did not betray her. Now, would we
really say that she now believes that he is reliable and did not betray her?
Her attitude seems too unstable to say without any qualification that she
(dormantly) believes that her son did not betray her.
Moreover, TB′ faces a problem of its own. For, as Robert Audi has
pointed out, having a disposition to believe that p upon considering p
does not suffice for believing that p. 20 Let me give an example to illus-
trate this point. Gottlob Frege famously believed that for any property
P, there is a set whose members are all and only those objects that have
P. This is his so-called basic law V. As Bertrand Russell showed, how-
ever, this law is false: for the property of being non-self-membered there
is no set whose members are all and only those objects that have that
20. See his example of Tom offending Norma: Audi (1994, 419–420).
36
37
property. Now, it seems reasonable to suppose that, given his impres-

sive intellect and logical skills, if Frege had so much as considered the
proposition q that the property of being non-self-membered provides a
convincing counterexample to this basic law V, he would have believed
it. But if he had believed it, he would not have accepted basic law V. But
he did. Therefore, he did not believe that q. TB′, however, falsely implies
that he did believe that q, for (i) Frege presumably never considered q,
and (ii) if he had considered q, he would have thought that it is true.
Where does this leave us? We saw that both DB and TB, cashed out
in terms of obvious entailment, and DB′ and TB′, cashed out in terms of
dispositions, are problematic.
A natural solution, which I will defend here, is to combine the vir-
tues of both approaches by spelling out dormant and tacit belief in terms
of both obvious entailment and certain dispositions. For obvious rea-
sons, I call it the Combination Account. On the Combination Account,
dormant belief can be understood as follows:
DB″ S dormantly believes that p at t iff (i) S at t does not consider

whether p, (ii) the last time t* at which S considered whether
p, S thought that p,21 and (iii) at any time from t* and t, if S
were to consider whether p, S would normally think that p.
And tacit belief can be understood as follows:
TB″ S tacitly believes that p at t iff (i) S at t does not occurrently

or dormantly believe that p, (ii) from S’s perspective, p
obviously follows from propositions that S occurrently or
dormantly believes at t, and (iii) if S at t were to consider
whether p, S would normally think that p.
It seems to me that the Combination Account—which I take to include

OB—provides a plausible analysis of at least three different ways in
21. L et me stress that (ii) should be non-vacuously true. That is, one dormantly believes
that p only if there actually was some time at which one considered whether p.
37
38
which we can believe a proposition. It also seems natural in that it cuts

belief at its joints. There are beliefs that we have because we consciously
entertain propositions and think that they are true or that they are the
case. There are also beliefs that we have because we have consciously
entertained propositions in the past and what we thought on those
occasions has somehow remained part of how we see the world. And
there are beliefs that also constitute a part of how we see the world,
not because we have ever consciously considered the propositions in
question, but because they obviously follow from propositions that we
have considered and that we think are true or are the case. OB, DB″,
and TB″ are simply somewhat more detailed statements of these differ-
ent ways of believing a proposition. 22 I do not claim that these are the
only ways of believing a proposition, nor do I claim that OB, DB″, and
TB″ are immaculate as they stand; it may well be that they need further
explanation and refinement on certain points. All I need for my theory
of responsible belief is that to believe that p is to occurrently, dormantly,
or tacitly think that p, and that there are plausible ways of spelling out
these different ways of believing a proposition.23
The Combination Account does not face the problems that I identi-
fied for the accounts merely in terms of obvious entailment and those
merely in terms of dispositions. First, on the Combination Account,
twenty-six-year-old Alfred does not dormantly believe that his father is
the strongest man in the world, for he has no disposition to believe that.
In fact, on the Combination Account he tacitly believes that his father
is not the strongest man in the world, for if he were to consider the issue
now, he would believe that his father is not strongest man in the world
and that is also something that obviously follows from other proposi-
tions he believes, such as that his father’s health is rather bad.
22. Disjunctive theories are not unusual in philosophy; we find them in the philosophy of
perception, the theory of practical reason, accounts of knowledge, and the philosophy
of action. See, for instance, Haddock and Macpherson (2008).
23. E lsewhere, Nikolaj Nottelmann and I have argued that a plausible ethics of belief implies
that there is not only occurrent, but also dormant and tacit belief; see Nottelmann and
Peels (2013). Here, I defend this view without appealing to a plausible ethics of belief
and I spell out in more detail how we should understand dormant and tacit belief.
38
39
Second, on the Combination Account, on Saturday night Rosemary

does not believe that her son betrayed her, because, although this is what
she thought when she last considered the issue, it is not what she would
believe if she were to think about it now. In fact, on the Combination
Account, Rosemary on Saturday night neither believes nor disbelieves
that her son betrayed her, and that squares well with our intuitions
about this case. Rosemary’s attitude toward the proposition that her son
betrayed her is simply not stable enough to count as a belief or to count
as a disbelief, at least not when she does not consider the issue. (She may
very well, at certain points, occurrently believe that her son betrayed
her, or that he did not). Imagine the following slight variation on the
case. The last time Rosemary considered the issue, she occurrently
believed that her son betrayed her. Also, she now has the disposition to
believe that. It follows from the Combination View that I propose that
she dormantly believes that her son betrayed her in this case. This seems
to chime well with our intuitions, for in this alternative case, Rosemary’s
mental state about her son is more stable than in the original version.
Third, the Combination Account is also able to meet the problem,
based on the example of Frege’s basic law V, that I identified for TB′. On
the Combination Account, Frege did not believe the proposition q that
the property of being non-self-membered provides a refuting counterex-
ample to his basic law V. For, even though Frege would have believed q if
he had considered it, q did not obviously follow from other propositions
that he occurrently or dormantly believed in the sense of “obvious entail-
ment” that I specified above: without any doubt, q’s truth would have
come as a surprise and, in fact, a discovery to him. The Combination
Account, then, is able to meet the worries faced by the other two views.
Even if the Combination Account meets the worries I identified ear-
lier, it may face other problems. Let us, therefore, consider what might
be thought to be five problems for the Combination Account as I have
spelled it out. First, one might think with William Lycan that what one
believes should not depend on one’s intelligence.24 This, however, con-
flicts, with DB″ and TB″. After all, how intelligent one is will often make
a difference to whether or not a proposition is obviously entailed by other
24. See Lycan (1988, 60).
39
40
propositions one believes. This would, of course, also be a problem for

DB′ and TB′, not only for DB″ and TB″. But it seems to me this is actually
a consequence of the Combination Account that we can gladly embrace.
Consider the following scenario. There is a mathematical proposition p
that a mathematician and I have both considered and think to be true.
There is also some other proposition q which both of us have never con-
sidered. For the mathematician, but not for me, q obviously follows from
p: he would think it to be true as soon as he were to consider it without
being surprised by it, whereas it would take me several minutes to see
that q follows from p. It seems perfectly plausible to say that, prior to con-
sidering q, the mathematician believes that q, whereas I do not.
Second, imagine that I once visited the beautiful ruins of the castle
Chalusset in the Haut-Limousin (France). If you ask me the ruins’ name,
I cannot tell you. But if you were to ask me whether it is “Chalusset,”
I would immediately remember it and occurrently believe that it is
“Chalusset.” It follows from DB″ that I believe that the ruins’ name is
“Chalusset,” even if I cannot retrieve it from my memory before being
presented with it.25 Again, if this is a problem for DB″, it is also a problem
for DB′. But I do not think it is. A plausible interpretation of this scenario
is that I believe—or know and, hence, believe—t hat the ruins’ name is
“Chalusset,” but that I do not remember it. Upon being presented with
its name, I immediately recognize it and occurrently think it is true that
that is its name. My immediate assent without any further evidence is
best explained by the fact that all along I believed it, even though I could
not remember it.
Third, one might claim that the Combination Account seems to give
the wrong verdict on my attitude toward the proposition p that I am cur-
rently considering a proposition. If the Combination Account as I have
spelled it out were correct, I would always believe that p, because I always
satisfy the conditions of DB″: (i) the last time I considered whether p,
I thought that p, and (ii) at any time between the last time I considered p
and now, if I were to consider p, I would think that p. Obviously, though,
I do not always believe that I am currently considering a proposition. 26
25. Th is objection might be inspired by Goldman (1978, 526).

26. F or this kind of objection, see Foley (1978, 312n).
40
41
In response, I would like to point out that it seems that there is no one
proposition involved in this example. The proposition p that I consid-
ered when last time considering whether I was then believing a proposi-
tion is not the same proposition as the proposition q that I now consider
when I consider whether I am now considering a proposition, and that
is because p and q have different time indices. Therefore, p does not sat-
isfy the condition that I would now think it to be true if I were to enter-
tain it, and q does not meet the condition that last time I considered it,
I thought that it was true. Hence, the scenario provides no convincing
counterexample to the Combination Account.
Fourth, one might object that TB″ is too liberal in that it counts
as tacit beliefs what are really dispositions to believe. As we saw, the
Combination Account already excludes some dispositions to believe
from being beliefs, such as Frege’s relation to the proposition that the
property of being non-self-membered provides a good counterexample
to his basic law V. But one might think that it still counts too many dis-
positions as beliefs. According to Audi, for instance, I have a disposition
to believe that 124 is larger than 98, not a belief that that is the case. For,
when I consider this proposition, it seems I form the belief that 124 is
larger than 98.27 In response, let me point out that I do indeed come to
occurrently believe that 124 is larger than 98 when I consider that propo-
sition. It does not follow that I did not believe this before I considered it.
(Even Audi acknowledges that one need not have considered a proposi-
tion in order to believe it.28) In fact, it seems that any educated adult
human being knows that 124 is larger than 98. However, since knowl-
edge entails belief, it follows that every educated adult human being
believes that 124 is larger than 98.
Fifth, one may object that TB″ suffers from an embarrassment of
riches, since it would follow that we have an infinite number of beliefs.
This, one might think, conflicts with materialism: if we are material
beings, then, given that we consist of a finite number of material enti-
ties, we cannot have an infinite number of beliefs. However, we should
note that accounts in terms of obvious entailment and dispositional
27. S ee Audi (1994, 422).

28. S ee Audi (1994, 421).
41
42
accounts face the same problem: they also imply that we hold an infi-
nite number of beliefs, for there is an infinite number of things that are
obviously entailed by what we believe and we have an infinite number
of dispositions. Moreover, the idea that we have an infinite number of
beliefs conflicts with materialism only if we make several further con-
troversial assumptions. Embracing both materialism and, for instance,
functionalism—on which beliefs are certain functions (causal rela-
tions) of the brain—is perfectly compatible with the claim that we have
an infinite number of beliefs.
As I said above, I take it that one disbelieves that p if and only if one
believes that ~p. The Combination Account that I have defended also
provides us with the material to construe a plausible account of suspen-
sion of judgment. For we can now say that one suspends judgment on p
if, roughly, one has considered p or p obviously follows from proposi-
tions that one has considered, and one neither believes nor disbelieves
that p. I say “roughly,” because as Jane Friedman has pointed out, it also
seems necessary that one has actually adopted an attitude toward p.29
The simpler definition will do for our purposes, though.
This means that the three doxastic attitudes of belief, disbelief, and
suspension of judgment are mutually exclusive, 30 but not jointly exhaus-
tive: one can stand in a relation to a proposition that is not an instance of
belief, disbelief, or suspension of judgment. For instance, propositions
that I have never considered and that do not obviously follow from prop-
ositions that I believe, disbelieve, or suspend judgment on, are proposi-
tions toward which I do not have a doxastic attitude. In chapter 5, I argue
that our relation to such propositions is best considered as an instance
of ignorance. I think this squares well with how we would describe the
cognitive situation of people who are in such a state. We would not say
that small children suspend judgment on the proposition that glucose is
composed of C6H12O6. Rather, we would say that—t hrough no fault of
their own, of course—t hey are deeply ignorant of it.
29. S ee Friedman (2013).

30. C hisholm (1977, 6), is one of the first philosophers to clearly present the idea of there
being three doxastic attitudes.
42
43
Let me summarize. I have argued that to believe a proposition p is,

roughly, to occurrently, dormantly, or tacitly think that p, and I have
argued that dormant and tacit beliefs are best understood in terms of
both obvious entailment and certain dispositions. Again, it may be that
OB, DB″, and TB″ still need further refinement on certain points. I take
it, though, that what I have argued above shows that there is not only
occurrent, but also dormant and tacit belief, and that there are plausible
ways to spell out what these different ways of believing a proposition
amount to.
1.4 R
ESPONSIBLE BELIEF
IS BL AMELESS BELIEF
On my Appraisal Account of responsibility, to be responsible for hold-

ing a belief is to be the proper object of praise, blame, or neutral appraisal
for one’s belief. Precisely which of these normative attitudes, however,
maps onto responsible belief? In other words, which of these norma-
tive attitudes would be the proper kind of normative attitude toward
responsible belief? Clearly, to believe responsibly is not to believe blame-
worthily. And, clearly, if one believes praiseworthily, then one believes
responsibly. But is praiseworthiness necessary for responsible belief or
does blamelessness suffice? For economical reasons, in this section I will
use the term “blameless” as shorthand for “praiseworthy or the proper
object of neutral appraisal.” As I understand it here, then, blamelessness
entails responsibility. Let us distinguish two views: the Blamelessness
Thesis (BT) and the Praiseworthiness Thesis (PT):
BT: S responsibly believes that p iff S blamelessly believes that p.

PT: S responsibly believes that p iff S praiseworthily believes
that p.
BT and PT are mutually exclusive, since they give different verdicts

on those cases in which one is neutrally appraisable for one’s belief.
For instance, if my belief that communism is bankrupt is the proper
object of neutral appraisal, BT says that I believe responsibly, whereas
43
44
PT denies this. This means that an account of responsible belief has to

take a stance on this issue. To my knowledge, even though BT is often
tacitly assumed in the literature, no arguments have been provided for
it. Let me, therefore, start by giving three arguments for BT. After that,
I will discuss three arguments that Brian Weatherson has put forward
in defense of PT.
First, it seems that one’s account of responsibility for actions and
that of responsibility for beliefs should be structurally similar, unless
we have good reason to think that there are important differences
between the two realms of responsibility that account for structural
dissimilarity. However, to act responsibly is not to perform a praise-
worthy action, but to perform an action that is blameless. Imagine
that Alastair, a teacher, is not particularly fond of his job but decides
to meet all his professional duties by preparing and teaching his classes
and correcting the students’ exams. Imagine also that Alastair decides
not to do anything that, as a teacher, he is not required to do. Thus,
he corrects the exams within the required period of time, but not as
quickly as he could, and he prepares his classes sufficiently well, but not
as thoroughly as he could if he put his best effort into it. He spends the
rest of his time on his family life, charity work, hobbies, and so forth. It
seems that in doing his job the way he does it Alastair acts responsibly,
although in this regard there is nothing particularly praiseworthy about
him. To act responsibly, then, is to act blamelessly. It is not required
that one act praiseworthily. But if this is true, then why should we think
that things are different in the doxastic domain? Of course, there are
important dissimilarities between the two domains. For instance, it is
widely thought that belief, in opposition to action, is involuntary. But
it is not clear at all how that would favor PT over BT. At least, I know of
no argument to that effect. 31
Second, BT accords much better than PT with certain episte-
mological principles that are highly popular among epistemologists.
Epistemic conservatism is, roughly, the view that so long as S’s belief
31. Th is is not to deny that there is a quite different argument for PT from the idea that PT,
in opposition to BT, is able to solve the problem of doxastic involuntarism. I discuss this
argument below.
44
45
that p is undefeated, S’s belief that p is justified. 32 And on the Principle

of Credulity, in the absence of special considerations, it is justified
to believe that p if it epistemically seems to one that p. 33 Now, what
at least many of those who champion these principles, such as Kevin
McCain and Richard Swinburne, seem to have in mind with “justi-
fied belief ” is responsible belief: one is blameless for believing as one
does, there is nothing culpable about it. But if it is possible to believe
responsibly merely on the ground that one’s belief is undefeated, then
it is surely too much to require praiseworthiness; one is not praise-
worthy simply for having an undefeated belief. Hence, adherents of
epistemic conservatism or the Principle of Credulity who take these
principles to concern responsible belief, then, should embrace BT
rather than PT.
Third, unlike BT, PT sets the standard too high and thereby makes
it virtually impossible to meet certain philosophical challenges. Take
the Cartesian skeptical challenge that, since we have no better reason
to believe that there is an external world than a world in which an evil
demon systematically cons us into thinking that there is, we should not
believe that there is an external world. Nearly all philosophers believe
that there is something wrong with such skeptical arguments, since
we take it that we responsibly believe that there is an external world.
BT seems to have at least some resources to explain why: there is noth-
ing blameworthy about our belief that there is an external world, for
instance, because we cannot help but believe that there is such a world
or because we have no good defeaters for thinking that there is an exter-
nal world. The adherent of PT, however, would have to say that we are
praiseworthy for our belief that there is an external world. But that seems
false: we all automatically believe that there is an external world; since it
is inevitable, there seems nothing praiseworthy about it.
I take it that these considerations do not demonstrate BT’s truth, but
that they clearly count in its favor. This means that we should accept BT,
unless there are good reasons for accepting PT that trump or balance the
32. F or a defense of this thesis, see McCain (2008).

33. S ee Swinburne (2001, 140–142).
45
46
arguments in favor of BT. Let us, therefore, discuss three putative advan-
tages of PT over BT that have been brought forward by Weatherson. 34
First, according to Weatherson, PT can handle the problem of dox-
astic involuntarism, whereas BT cannot. As I show in more detail in
chapter 2, the problem is that, since doxastic responsibility seems to
require control over our beliefs but we lack such control, we cannot be
said to bear doxastic responsibility. One might think that praisewor-
thiness, in opposition to blameworthiness, does not require control.
Consider Weatherson’s cricket captain, who comes up with a particu-
larly imaginative field placing during a match. While we may want to
praise him, we do so despite the fact that he had no control over that.
After all, claims Weatherson, coming up with a particular field place-
ment is hardly something one can set out to do. And we deem this cap-
tain more praiseworthy than his colleague, who works equally hard, but
does not come up with such an imaginative field placement. So, if we
can justly praise S for φ-i ng, irrespectively of whether φ-i ng is under S’s
control, then PT is immune from the voluntarism problem.
I think that this strategy is unconvincing for at least three reasons.
First, if only praiseworthy beliefs escape the problem of doxastic invol-
untarism, all our beliefs are either praiseworthy or such that we are not
at all responsible for them. But, clearly, certain beliefs for which we are
responsible are blameworthy; in the Introduction and this chapter,
I have already given several examples of blameworthy belief. The argu-
ment is, therefore, unconvincing.
Second, I argue in c hapter 3 that we have indirect influence on our
beliefs. We can, for instance, train ourselves to be more critical of gossip
and that will make a difference to which beliefs we hold. Similarly, we can
train ourselves to be more imaginative, for instance, by trying to conceive
of innovative solutions to everyday problems. In bringing the notion of
indirect influence into play, we can see Weatherson’s example in a different
light. It now seems that the praise we ascribe to the captain is in fact due to
his capability to engage and train his imagination. Thus, there is no reason
34. A rguments for PT are hard to find. The only ones that I know of are found in Weatherson
(2008). Other philosophers, such as Levy (2007, 141), merely assume the truth of PT.
46
47
to think that being the proper object of the normative attitude of praise
for φ-ing does not require some kind of control over or influence on φ-ing.
Third, it seems that my distinction between evaluative and nor-
mative attitudes also applies to praise. Praise can be an evaluative atti-
tude: we can praise S for φ-ing if we value S’s φ-ing, without thereby holding
S responsible for φ-ing. Thus, I might praise Gloria for her beauty or my
recently bought Chevrolet for its speed, without holding them responsible
for these things. This sort of praise ought to be clearly distinguished from
praise as a normative attitude. If we adopt the normative attitude of praise
toward someone for φ-ing, we can in principle also, say, blame or resent that
person for φ-ing. Now, if, contrary to what one would expect, the imagina-
tive captain did not exercise his capacity to train his imagination, but none-
theless happened to come up with imaginative field placements, we would
praise him only in the first way, that is, express an evaluative attitude rather
than a normative attitude. But this kind of praise has nothing to do with
responsibility: we just value the captain’s imaginative field placements and
admire him for that. The example would then be irrelevant to the issue of
doxastic responsibility and, hence, would not count in favor of PT.
A second argument for PT is based on our intuitions in New Evil
Demon scenarios. Such scenarios are usually presented in arguments
against reliabilist theories of justification, theories on which one’s
belief is justified if it is produced by a reliable cognitive mechanism
that aims at truth. Assume, for instance, that there are three victims
of equal demonic deception: Alex, Bert, and Caroline. Alex frequently
uses faulty reasoning procedures to arrive at his beliefs, something
he would see if he were a little more careful, as he should be. Bert also
often uses faulty procedures to arrive at his beliefs, but his faults are the
product of bad upbringing and the mistakes are so subtle that we can-
not expect Bert to notice them. Caroline displays the paradigm of good
reasoning, but still has mostly false beliefs because she is the victim of
devilish deception. 35
Weatherson claims that the notion of blamelessness cannot capture
the intuition that Caroline is epistemically better than Bert. If both are
35. For his presentation of such scenarios, see Pryor (2001, 117).
47
48
blameless, then, according to BT, both Bert and Caroline believe respon-
sibly. However, there is an important distinction to be made between
Alex and Bert on the one hand and Caroline on the other, namely that
only Caroline is a good evidence processor: she processes her evidential
input excellently. Thus, Caroline is epistemically praiseworthy in a
way that Alex and Bert are not. Only PT makes sense of the difference
between Bert and Caroline.
I agree that we need something like the notion of praiseworthi-
ness to distinguish between Bert and Caroline—a lthough, again, it
seems that the kind of praise involved here is an evaluative rather than
a normative attitude, since Bert may be not be responsible for being
a worse evidence processor than Caroline and Caroline may not be
responsible for being a better evidence processor than Bert. However,
we also need the notion of blamelessness to distinguish between Alex
and Bert: while Alex is blameworthy for using faulty reasoning pro-
cedures, Bert is not. It seems impossible to articulate this difference
if we have only the concept of praiseworthiness at hand, since neither
is worthy of praise. In order to make sense of all the intuitions in the
above scenario, then, we need both the concept of praiseworthiness
and that of blamelessness. This means that our intuitions here do not
favor PT over BT.
The final argument that Weatherson provides for PT is rather com-
plicated. In the argument, Weatherson uses the expression “justified
belief,” but by that he simply means responsible belief (whether that
amounts to praiseworthy belief or blameless belief; if it did not mean
responsible belief, the argument would be irrelevant to the issue at
hand). The basic idea is that, unlike PT, BT in conjunction with some
plausible principles leads to a contradiction and that we should, there-
fore, reject it. Here are the two principles that Weatherson takes to be
plausible:
(1) It is possible for S to have a justified but false belief that her
belief that p is justified.
(2) If S blamelessly believes that she is justified in believing that
p, and on the basis of that belief comes to believe that p, then
she is blameless for believing that p.
48
49
That BT, in conjunction with (1) and (2), leads to a contradiction can,

according to Weatherson, be shown as follows:
(3) S justifiedly, but falsely, believes that she is justified in believ-

ing that p. [prem.; from (1)]
(4) On the basis of this belief, S comes to believe that p. [prem.]
(5) S blamelessly believes that she is justified in believing that p.
[from (3), BT]
(6) S blamelessly believes that p. [from (2), (4), (5)]
(7) S is justified in believing that p. [from (6), BT]
(8) S is not justified in believing that p. [(from (3)]
(9) S both is and is not justified in believing that p. [from (7), (8)]
(10) Therefore, BT is false. 36
According to Weatherson, (1) “is extremely plausible” and to give up

(2) is to commit oneself to externalism about justified belief. Thus, if we
believe that we bear doxastic responsibility and want to be internalists
about justified belief (i.e., responsible belief), we have to give up BT. 37
Elsewhere, Anthony Booth and I have argued that if this is a problem
for BT, then PT does not fare any better. 38 Here, I will merely point out
where I take the argument to go wrong.
We should note that if the argument is to refute BT, (1) should be
understood as saying that it is possible for someone blamelessly to believe
that her belief that p is blameless, while her belief that p is in fact blame-
worthy. But it is not at all clear why we should buy into (1), thus under-
stood. It is, of course, widely acknowledged that, generally, justification
does not entail truth. Thus, blamelessly believing that p is no guarantee
that p is true. This is perfectly compatible with saying that blamelessly
believing that one blamelessly believes that p entails that one blame-
lessly believes that p. Contrary to what Weatherson claims, this move
is not ad hoc. If one believes that p and one blamelessly believes that it
36. Actually, (9) is not part of Weatherson’s version of the argument. I have added it to make
the contradiction that he aims at explicit.
37. S ee Weatherson (2008, 567–569).
38. S ee Booth and Peels (2010, 262–2 64).
49
50
is perfectly all right or responsible to believe that p, then how could one
possibly be blameworthy for believing that p? The very fact that one’s
meta-belief is blameless seems to guarantee that one’s first-order belief
is equally blameless, at least if one has the first-order belief because one
has the second-order belief. This means that if (3) is true, (4) cannot be
true. Weatherson disagrees and suggests that the following provides a
good counterexample:
S has a false but justified belief that e is part of her evidence. S

knows both that anyone with evidence e is justified in believing
p in the absence of defeaters, and that there are no defeaters pres-
ent. So S comes to believe, quite reasonably, that she is justified in
believing that p. But S does not have this evidence, and in fact all
of her evidence points toward ~p. So it is false that she is justified
in believing p. 39
Here, Weatherson claims that there are situations in which (a) S has a
false but justified belief that e is part of her evidence, (b) S knows that
anyone with evidence e is justified in believing p in the absence of defeat-
ers, (c) S knows that she has no defeaters for e, (d) S, therefore, comes to
believe that she is justified in believing that p, (e) S lacks e, (f) S’s evi-
dence strongly points toward ~p, and, therefore, (g) S is not justified in
believing p.
Weatherson’s argument, however, trades on the ambiguity of the
word “justified,” which he uses synonymously with “blameless” in (a),
(b), and (d), but purely externalistically—a matter of (not) believing in
accordance with one’s evidence—i n (g). It is because of this ambiguity
that prima facie the situation seems possible. If, however, we understand
all instances of “justified” in this quote as equivalent to “blameless”—
as we ought to, if BT is the thesis under investigation—we see that the
scenario is impossible. It does not seem possible that someone (i) has
a blameless belief that e is part of her evidence, (ii) knows that anyone
with evidence e is blameless in believing that p in the absence of defeaters
39. W
eatherson (2008, 567–568).
50
51
and that she has no defeaters for e, (iii) on the basis of that comes to
believe that p, and (iv) is blameworthy for believing that p. If conditions
(i)–(iii) obtain, then for S to believe that p is clearly to believe respon-
sibly, whether or not S’s belief that p is justified in any externalist sense
of the word. In other words, when (i)–(iii) are met, it cannot be that (iv)
is also met, although S’s belief that p may well be epistemically unde-
sirable in some way or other. Weatherson’s example is incoherent, for
the conditions that he specifies cannot be met simultaneously. Hence,
Weatherson’s third argument also fails to count against BT.
1.5 C ONCLUSION
We can now draw three conclusions that will provide the framework for
my account of responsible belief that I develop in the ensuing chapters.
First, to be responsible for believing that p is to be the proper object of
one or more normative attitudes like praise, blame, and neutral appraisal
for believing that p. Being the proper object of a merely evaluative atti-
tude for believing that p does not suffice for being responsible for believ-
ing that p. Second, to believe that p is to occurrently, dormantly, or
tacitly think that p. Accounts merely in terms of obvious entailment or
merely in terms of dispositions have trouble giving a convincing analy-
sis of dormant and tacit belief, but an account in terms of both obvious
entailment and dispositions seems to have the resources to do so. Third,
to believe responsibly is to believe blamelessly, that is, to be the proper
object of either neutral appraisal or praise for one’s belief. Responsible
belief excludes blameworthy belief, but one can believe responsibly even
if one does not believe praiseworthily.
51
52
[2]
THE PROBLEM
Doxastic Control and Doxastic Obligations
2.1 I NTRODUCTION
In c hapter 1, I argued that to believe responsibly is to be the proper

object of neutral or positive appraisal for occurrently, dormantly, or
tacitly thinking that p. But exactly when does one believe responsibly?
One acts responsibly, it seems, roughly, if one meets or at least does not
violate any obligations in performing that action. As William Alston
has argued, though, this gives rise to an important philosophical prob-
lem. We can act irresponsibly or responsibly—v iolate or not violate our
obligations—because our actions are under our voluntary control: we
normally can choose to do or not to do something. With belief, however,
things are different: it seems that we cannot choose our beliefs, that we
lack control over our beliefs, and that we, therefore, do not have any obli-
gations (not) to have particular beliefs. The literature displays four main
attempts to solve this problem. The aim of this chapter is to argue that
they fail. In c hapter 3, I develop and defend an alternative response to
Alston’s argument.
Let me give a short outline of this chapter. First, I provide a sum-
mary and analysis of Alston’s argument (§2.2). Next, I consider in
more detail the relation between having an obligation to φ and having
control over φ-i ng (§2.3). This will turn out to be helpful in assessing
responses that have been given to Alston’s argument. Subsequently,
I assess the four main responses to Alston’s argument that we find in
52
53
T he P roblem : D oxastic C ontrol and D oxastic O bligations
the literature and argue that none of them are convincing. First, it has
been thought that we do have direct control over our beliefs (§2.4).
Second, several philosophers argue that, even though we lack direct
doxastic control, we do have indirect doxastic control (§2.5). Third,
doxastic compatibilists claim that, although we lack intentional dox-
astic control, we have compatibilist doxastic control and that we have
doxastic obligations in virtue of that (§2.6). Finally, some philoso-
phers argue that having doxastic obligations does not require any kind
of doxastic control whatsoever (§2.7). Since these seem to be the only
possible games in town, I conclude that if we bear doxastic responsibil-
ity, it is not to be explained in terms of doxastic control and doxastic
obligations.
2.2 T
HE ARGUMENT FROM DOX ASTIC
INVOLUNTARISM
Alston’s argument is directed against what he calls the deontological con-

ception of epistemic justification. On this conception, a belief ’s being epis-
temically justified is to be understood in terms of obligation, permission,
requirement, responsibility, praise, and blame. Obviously, epistemically
justified belief, thus conceived, and responsible belief are closely related
to each other, if not identical. Here, however, we are not concerned with
epistemic justification, but only with responsible belief, no matter how
the two are related to each other. For, as I said in the Introduction to this
book, whether or not one believes responsibly is a matter that deserves
independent philosophical attention. In the Appendix, I discuss in
detail how responsible belief relates to epistemic justification deonto-
logically conceived.
Here, I will therefore focus on that part of Alston’s argument that
is relevant for doxastic responsibility. Alston does not present a formal
version of the argument, but it seems it can be formalized as follows:
(1) For any person S and proposition p, S responsibly (blame-

lessly) believes that p only if S has not violated an obligation
53
54
not to believe that p or met her obligation to believe that

p. [ass.]
(2) For any S and p, S has an obligation (not) to believe that p
only if S has sufficient voluntary control over whether or not
S believes that p. [prem.]
(3) For any S and p, S has no voluntary control over whether or
not S believes that p. [prem.]
(4) For any S and p, S has no obligation (not) to believe that p.
[from (1)—(3)]1
With many other philosophers, I call this the argument from doxastic
involuntarism. This is because, as we shall see below, the idea that we lack
control over our beliefs is the most contested premise of the argument.
Now, we should note that Alston does not actually say how we are to
infer from (4) that (1) is false. It seems to me that there are two options,
each of which spells trouble for the idea that we bear doxastic respon-
sibility. First, if we should understand (1) in such a way that one can
believe responsibly only if there are doxastic obligations that one has
met or at least not violated, then the conditions for believing respon-
sibly are never met. It would follow that we never believe responsibly.
Second, if we should understand (1) in such a way that one can believe
responsibly even if there are no doxastic obligations, then the conditions
for believing responsibly are always met. It would follow that we always
believe responsibly. Both options are, clearly, highly problematic. This
is because on any plausible theory of responsible belief, sometimes we
believe responsibly and sometimes we believe irresponsibly.
Alston does not defend premise (2)—t he idea that doxastic obliga-
tions require doxastic control—i n any detail. He thinks it follows from
the principle that ought implies can. It is not trivially true, though. Below,
in section 2.7, I defend (2) against several accounts of doxastic responsi-
bility that imply the falsehood of (2).
1. For this argument, see Alston (1989b, 91–93; 1989d, 115–136; 2005, 58–73). For ease of
exposition, I have cast the argument in a strong form: it makes a claim about any person
and any proposition. Obviously, if there are a few exceptions, that fact will not save the
widespread doxastic responsibility that I referred to in the Introduction.
54
55
However, Alston’s defense of (3), the idea that we lack doxastic vol-
untary control, is extensive. He argues that, as a matter of psychologi-
cal fact, we lack control over our beliefs. Other philosophers, such as
Bernard Williams, have made the stronger claim that it is conceptually
impossible to believe at will.2 Elsewhere, I have argued that there is good
reason to think that the conceptual claim is false. 3 Here, I will sidestep
that issue and focus on Alston’s argument for the contingent, empirical
claim. Alston’s argument proceeds by way of enumerative induction. He
distinguishes between different varieties of control and asks us, for par-
ticular propositions, whether we have control over our doxastic attitude
toward them. Can we choose to believe that the United States is still a
colony of Great Britain? Can we choose not to believe that it is raining
when we look out our window and see rain falling? Can we choose to
believe that God exists or that materialism is true? In each case, it seems
the answer has to be negative.
One of the distinctions he makes is that between direct and indirect
control. There is no clear boundary between these two kinds of control,
but they can roughly be characterized as follows. I have direct control
over φ-i ng if I can choose to φ either simply by an act of will or by per-
forming a single action over a relatively short period of time. Thus, I have
direct control over nodding and over opening the door. I have indirect
control over φ-ing if I can choose to φ by interruptedly performing a
series of actions over a considerable period of time. Thus, I have indi-
rect control over my body weight, the color of my house, and my level of
education. A few philosophers have claimed that we have direct control
over our beliefs and even more philosophers have claimed that we have
indirect control over our beliefs. I discuss their views below. Before I do
so, however, let us first tighten our grip on the argument from doxastic
involuntarism by having a closer look at the relation between obliga-
tions and control.
2. See Adler (2002, 55– 6 4), Buckareff (2014, 33– 50), Church (2002, 367– 374),
O’Saughnessy (1980, 21–2 8), Scott-K akures (1994), Williams (1973, 148–149).
3. See Peels (2015b).
55
56
2.3 O BLIGATIONS AND CONTROL
It is important to note that philosophers use the word “obligation” in

different ways. Let me introduce these different uses by way of the fol-
lowing example. Michael is sitting on the beach. Suddenly, he witnesses
simultaneously an old man being robbed and a young girl drowning. He
realizes that he should attempt to help the man, but that it is even more
important that the girl be rescued. Unfortunately, because of a physical
handicap, Michael is unable to walk, let alone help the robbed man or
rescue the girl. Here are three views on what Michael’s obligations are
in this case. First, Michael has an obligation to help the robbed man and
an obligation to rescue the girl. The latter obligation trumps the former,
but because of his disability he is excused for not meeting his obligation
to rescue the girl. Second, Michael has a reason to help the robbed man
and a reason to rescue the girl. The latter reason weighs heavier, so that
all things considered he only has an obligation to rescue the girl, but he is
excused for not meeting that obligation because of his handicap. Finally,
on the third view, since Michael is clearly blameless for not helping the
man and for not saving the girl, he may have a reason to help the robbed
man and a reason to rescue the girl, but all things considered no obliga-
tion to do so. One might think, for instance, that having an obligation to
φ requires that one is able to φ, and that, since Michael is unable to do
anything, he has neither an obligation to help the man nor an obligation
to rescue the girl.
It seems to me that the difference between these three views is to a
significant extent a verbal dispute. I think it is important, though, to be
explicit on my use of “obligation,” in order to avoid confusion. I adopt the
first view. Michael has an obligation to help the man and an obligation to
rescue the girl. Since the latter obligation trumps the former, the former
is merely a pro tanto obligation, while the latter is an all-things-considered
obligation. However, he is excused for not meeting that all-things-
considered obligation in virtue of his inability to do so.4 Also, it seems
that we are only ever excused for the violation of all-things-considered
4. F
or a similar position, see Brandt (1959, 356–368), Ross (2002, 19).
56
57
obligations. Michael is not excused for not helping the man. Rather, he is
justified in not doing so, for in these circumstances it would not be wrong
to rescue the girl instead of helping the man. He is excused for not rescu-
ing the girl, because that was his all-t hings-considered obligation and he
was blamelessly unable to meet that obligation. I think this understand-
ing of “obligation” fits in well with the argument from doxastic involun-
tarism, as long as we add a clause about excuses. Proposition (1) would
then say that one responsibly believes that p only if one has not violated
an obligation not to believe that p or one is excused for doing so.
Let us now turn to “control.”5 I said that we are blameworthy for fail-
ing to φ if we have an all-t hings-considered obligation to φ and are not
excused for failing to φ. Let us consider the relationship between blame-
worthiness and control. I think the following scenario can help us to
tighten our grip on the relevant kind of control. Imagine that my child
is starving. There is a box in front of me that I know to contain a fresh
tuna sandwich. The box has ten buttons, numbered 1 to 10. I have no
idea which button I have to press in order to open the box. I am allowed
to press only one button. Only pressing button 7 will open the box. Now,
there is a sense in which I have control over getting the sandwich: I only
have to press 7. Since in this case, I have the causal power to open the box
and the causal power to not open the box, let us call this kind of control
causal control:
S has causal control over φ-i ng iff S can φ and S can ~φ.6

(5)
One might think that this unduly stretches the meaning of “control.”
I am not convinced that is true; there seems nothing incoherent or even
strange about saying that the person in the scenario that I described has,
5. A s I said, Alston makes some fine-g rained distinctions between different kinds of con-
trol. Unfortunately, he says little on what it is to have control over something. One of the
few things he says is that control necessarily extends to contraries: one has control over
φ-i ng only if one also has control over ~φ-i ng; see Alston (1989d, 123). This may be true,
but, as I will argue, it hides important complexities.
6. Here, the word “can” should be understood in such a way that one’s having causal control
is compatible with determinism, where determinism is the view that the past and the
laws of nature jointly determine which possible future will be actual.
57
58
unbeknownst to him, control over whether the box is open. In any case,
my argument will not rely on it; those who disagree can simply under-
stand (5) as a stipulative definition.
It is clear that my causal control is insufficient for being blameworthy
for not opening the box. Suppose I press the wrong button and thus fail
to secure a sandwich for my child. Clearly, since I did not know which
button will open the box, from the mere fact that I had causal control
over opening the box, it does not follow that I am blameworthy for not
opening it.7 Is having causal control nonetheless necessary for being
blameworthy? This seems intuitive, but compatibilists might think that
there are good reasons to think otherwise. Here, I will assume that it
is necessary and return to compatibilist worries in section 2.6 and in
chapter 4.
We saw that having causal control over φ-i ng is insufficient for being
blameworthy if one violates an obligation to φ. What else is necessary?
What also seems necessary is that S can φ intentionally. Usually, φ-i ng
intentionally includes some true belief about how to φ. Thus, I can open
the box intentionally only if I hold a true belief about how to open the
box. In other cases, all that seems required is knowledge-how. Most
people can intentionally twitch their eyelids or cough, but it is not clear
that they hold true beliefs about how to do so. Rather, they know how to
do so. Thus:
S has intentional control over φ-i ng iff (i) S can φ and S can ~φ,
(6)
and (ii) S can φ intentionally. 8
In opposition to causal control, intentional control is the kind of control

that renders one blameworthy for violating an obligation. If in the above
case I can intentionally open the box and the box remains closed if I do
not press 7, then I am blameworthy if I fail to do so.
7. Here, I assume that my inability to φ or my ignorance about how to φ is not itself

blameworthy.
8. For a distinction that is similar to my distinction between causal and intentional control,
see Goldman (1970, 203–2 07), Zimmerman (1986, 205; 2008, 132–133).
58
59
Henceforth, by “control” I mean “intentional control,” unless explic-

itly indicated otherwise. Now that we have tightened our grip on the
notions of obligation and control, let us consider how the two relate to
each other. Since Sharon Ryan is one of the few philosophers who in
their discussion of the argument from doxastic involuntarism mention
obligations, control, and excuses, I will use some of her examples and
ideas to see how obligations relate to control.9 One might think that, in
the spirit of the famous Kantian Dictum—whether or not Kant himself
actually endorsed this principle—“ought” implies “can” and that obliga-
tions, therefore, require control. As Ryan points out, though, this is not
convincing. For example, a bride has an obligation to attend her wed-
ding, even if she cannot do so because of a snowstorm.
Some philosophers might want to bite the bullet and say that if the
bride really cannot attend her wedding, then she does not have an obliga-
tion to do so. It might be true, they could say, that she ought to attend
her wedding in the sense that it is desirable for her to do so or undesirable
for her not to do so. It does not follow that she has an obligation to do so.
I do not find this convincing. The bride properly feels that she is under
an obligation to make it to the wedding. In this regard, her attending
the wedding is different from all sorts of other desirable or undesirable
states of affairs that she could actualize. Of course, one might just stipu-
late that having an obligation entails having control and thereby define
the word “obligation” as a philosopher’s term of art. However, I prefer
to stick closely to how we ordinarily use the word. And it seems that in
the bride’s case, we say that she has an obligation to attend her wedding,
but that she is excused for not doing so in virtue of her inability to do so.
Having an obligation, then, does not entail that one has control.
Given the exceptionality of such situations as that of the bride, one
might think that:
S at t has an obligation to φ only if φ-i ng is normally under S’s

(7)
control.10
9. See Ryan (2003, 49–57).

10. Th
us Alston (1989d, 118n).
59
60
Under normal conditions, the bride could have attended her wed-
ding. Unfortunately, as Ryan rightly points out, (7) is false. My friend
has an obligation not to steal my CDs, even if she cannot but steal them.
Of course, her kleptomania may excuse her, but it is not at all implausible
to say that she is excused for violating her obligation not to steal those
CDs. The following analysis avoids this problem:
(7′) S at t has an obligation to φ only if at t people normally have

control over φ-i ng.
Again, counterexamples are easy to come by. Ryan asks us to imagine that
Kristin can calculate numbers much faster and better than anyone else in
the world. If a crazed man barges into a coffee shop, where Kristin hap-
pens to be, and threatens to blow it up if no one solves a particular complex
math problem in less than five seconds and Kristin is able to do that, she
has an obligation to do so. Nor will it do, I think, to make a disjunction of the
right hand-sides of (7) and (7′) to the effect that S at t has an obligation to φ
only if at t φ-ing is normally under S’s control or people’s control. Imagine
that Kristin normally lacks the relevant mathematical talent, but that she
knows that she has it on this particular occasion. Then, it seems Kristin has
an obligation to solve the problem, even if she does not meet the disjunctive
condition just spelled out. I therefore propose the following alternative:
(7″) S at t has an obligation to φ only if (i) at t people normally have

control over φ-i ng, or (ii) S normally has control over φ-i ng, or
(iii) S at t has control over φ-i ng.
This analysis gives the correct verdict on the above examples. It may
need some further refinement in order to make sense of all possible cases
of obligations, but (7″) will do for our purposes.11
Before we move on to consider the four main responses that have
been given to the argument from doxastic involuntarism, I would like
11. Of course, (7″), with its three disjuncts, is not as theoretically elegant as, for instance,
the principle that having an obligation requires having control. However, that is because
it takes into account an important fact of our lives, namely that we excuse each other for
violating certain obligations by appealing to our inability to meet them.
60
61
briefly to consider Ryan’s own response to the argument. On the basis

of the examples she gives and those of others, such as Walter Sinnott-
Armstrong and Michael Stocker,12 she claims that having an obligation
does not require having control. She subsequently applies this idea to
the argument from doxastic involuntarism and concludes that we should
reject premise (2), which says that one has an obligation (not) to believe
that p only if one has sufficient voluntary control over whether or not
one believes that p.13 My discussion in this section shows where Ryan’s
reasoning goes wrong. Even though her examples are convincing, hav-
ing an obligation, as we saw, still requires that most people have control,
that one normally has control, or that one has control on the occasion in
question. This means that we frequently have doxastic obligations only
if we frequently have control over our beliefs. But, if Alston is right, we
hardly ever have control over our beliefs. Counterexamples to the gen-
eral principle that having an obligation requires having control, there-
fore, do not defeat the argument from doxastic involuntarism.
2.4 F IRST RESPONSE: WE HAVE DIRECT

INTENTIONAL DOX ASTIC CONTROL
Carl Ginet claims that he sometimes has the ability to believe at will, but
he provides no argument for that view.14 And Philip Nickel argues that
in exceptional circumstances, namely situations in which our evidence
does not favor one doxastic attitude over the other, we can believe at
will.15 As far as I know, though, the only detailed defense of the thesis
that our doxastic responsibility is to be (nearly) entirely explained in
terms of our direct intentional control is provided by Matthias Steup.16
According to Steup, we intentionally form our beliefs in a way that is
not significantly different from how we intentionally perform most of
12. S ee Sinnott-A rmstrong (1984); Stocker (1971).

13. S ee Ryan (2003, 50–53).
14. S ee Ginet (2001).
15. See Nickel (2010).
16. M atthias Steup also provides another kind of response to Alston’s argument, namely
that we have compatibilist doxastic control. I discuss that response in section 2.6 below.
61
62
our actions. In believing that p, we can and often do carry out an inten-
tion to believe that p. Steup gives the following example in support of
this claim:
Suppose that, having returned from a trip and taken a shuttle to

the airport parking garage, I am now where I thought I left my car.
To my surprise, it is no longer there. I wonder whether it has been
stolen. There is of course the possibility that I don’t accurately
remember where I parked it. So I retrieve the paper slip which
states the exact location of my parking spot. According to the slip,
I am at the right spot. Considering my evidence—t he parking slip
and the absence of my car—I conclude that it was stolen.17
Steup interprets this case as follows: I consider my evidence, I decide

to believe that my car was stolen—where the causal relation between
this belief and the decision to hold it is non-deviant—a nd then come
to believe, as a result of that decision, that my car was stolen. According
to Steup, my decision to believe that my car was stolen is analogous to
deciding, say, to take a walk: one considers one’s reasons for taking or
not taking a walk and then decides to take a walk. Steup discusses and
replies to two arguments against this interpretation of the scenario.
Here, I will level two objections of my own against Steup’s interpreta-
tion. For it seems to me that there are at least two important differences
between coming to believe that my car was stolen and deciding to take
a walk.
First, I think Steup is right that, before considering my reasons, it
seems to me that there are two options available to me; namely, respec-
tively, believing that p and not believing that p, and taking a walk and not
taking a walk. However, once I have considered my reasons, the scenarios
are different. When I have considered my reasons for believing or not
believing that the car was stolen and find myself with sufficient evidence
for believing that it was, I cannot but believe that the car was stolen.
But when I have considered my reasons for taking or not taking a walk
17. S teup (2012, 157). See also Steup (2016).
62
63
and find myself with equally good reasons to take a walk as to not take a
walk—say, because the weather is really nice, but I also do not have that
much time left to prepare a lecture that I should deliver this evening—
it seems it is up to me whether or not I take a walk. This is confirmed
by how we interpret such scenarios afterward. We think that, given the
exact same practical reasons, we could have decided not to take a walk.
But we do not think that, given the exact same evidential reasons, we
could have failed to believe that the car was stolen.
This is true even when we make the two scenarios more similar to
each other by imagining that my evidential reasons for and against p are
balanced, in the same way as my reasons for and against taking a walk
are balanced. Here, one could think of the kind of evidence regarding
such propositions as that the number of stars is even or that the next
time I flip this coin, it will turn up heads. Once I have considered all my
reasons, I cannot but suspend judgment on the proposition that the
number of stars is even and on the proposition that the coin will turn
up heads next time I flip it. Similarly, if my evidence on the proposition
that my car was stolen is balanced, I cannot but suspend judgment on
whether it was stolen. However, again, once I have considered all my
reasons for and against taking a walk and they are on a par, it seems that
it is still up to me whether or not I take a walk.
In a recent paper, Steup has replied to this Up-to-me-ness Argument
of mine, as he calls it.18 Steup points out that the argument can be spelled
out either in a compatibilist way or in a libertarian way. He thinks it is
plausible that in the car scenarios, my belief is up to me in a compatibilist
sense. After all, my beliefs about the car are properly responsive to the
evidence I have. Of course, we are now no longer talking about direct
intentional doxastic control. Below, in section 2.6, I argue that having
compatibilist doxastic control is insufficient to have doxastic responsi-
bility. As to spelling it out in a libertarian way, Steup thinks it is hard to
assess the Up-to-me-ness Argument, until we have been given a detailed
account of when libertarian free will is possible. Now, it would, of course,
be nice to have such an account, if, indeed, such an account is possible.
18. See Steup (2016).
63
64
However, there is no reason to think that such an account is needed at

this point. For all I am pointing out is that if my reasons for and against
believing that the number of stars is even are balanced, I cannot choose
my doxastic attitude; normally, I inevitably suspend judgment in that
case. However, when my reasons for and against taking a walk in the
park are balanced, I can choose to take a walk, but I can also choose not
to take a walk. My action is normally not psychologically determined
by my reasons in such a case, while my doxastic attitude is normally
psychologically determined by my reasons in the case I just described.
I conclude that Steup’s reply to my first argument against his view that
we have direct intentional doxastic control is unconvincing.
Second, it is an important characteristic of voluntary actions that
we can decide when we perform them. I can now decide to take a walk in
five minutes, or tonight, or tomorrow. In deciding whether or not to per-
form an action, I also decide when I will perform it or, at least, whether
or not I will perform it right now.19 I might even decide to perform it at
some time in the future, without deciding on when I am going to per-
form it. When I decide to go home, to buy a bicycle, or to raise my arm,
I thereby also decide whether to do it straightaway or to do it later, and
if the latter, we often decide when we will do it. If I decide to purchase
a car tomorrow, then from that moment onward—u nless I change my
mind—I explicitly or implicitly intend to purchase a car tomorrow. All
this is absent in the case of belief. Once I have considered my evidence
regarding p, I find myself automatically and immediately believing that
p, disbelieving that p, or suspending judgment on p. I cannot decide to
believe that p later—i n, say, five minutes or tonight. Nor can I have the
intention for an extended period of time to believe that p—which is not
to deny that I may have the intention to try to acquire the belief that p
during a certain period of time. This is a second, important difference
between the process of coming to believe that my car was stolen and
deciding to take a walk.20
19. W hen I say each of these is possible, I mean that they are both psychologically possible, not
necessarily also metaphysically possible, because I do not want to commit myself to the
falsehood of determinism here. Matthias Steup rightly raises this worry: Steup (2016).
20. A n exception may be intentional actions with regard to once-in-a-lifetime opportu-
nities, such as shaking the hand of a celebrity. Such actions can be intentional, even
64
65
Steup has also responded to this second argument of mine, which

he calls the Delay Argument. 21 Steup’s reply is that the ability to form
at t an intention to φ at a later time t* is not an essential feature of inten-
tionality. Here are two examples of his that are meant to illustrate
the point:
(A) Sue is a surgeon. She is stationed with US marines in a war

zone. When she receives patients with life- t hreatening
injuries, she follows a standard surgical protocol, consist-
ing of a long list of steps each of which is necessary to save
the patient’s life. In the particular case under consideration,
Sue is clamping an artery. If she were to delay clamping the
artery, the patient would die.
(B) Approaching a red light at a busy intersection with heavy
traffic, Ralph steps on the brakes. If he were to delay stepping
on the brakes, he would run into oncoming traffic and cause a
serious accident, thereby putting in jeopardy his own life and
that of other motorists. 22
I will say three things in reply. First, it seems to me that in situation

(A) Sue could but normally would not postpone clamping the artery,
given that she is aware of the serious consequences of doing so. Mutatis
mutandis, the same can be said about Ralph in (B). I do not see any rea-
son to think that they literally could not postpone these actions, even
though there is plenty of reason to think that, unless they are mal-
functioning, they would not postpone these actions. Second, given the
special circumstances, it might be the case that Sue and Ralph cannot
postpone the actions in question for a very long time, but it seems they
can postpone them at least somewhat. It seems, for instance, that Sue
could decide to clamp the artery half a second later and that Ralph could
though one cannot decide to do them at a later point in time—it is now or never. Most
intentional actions are actions we can decide to do at some later point in time, though,
whereas we can (virtually) never decide to form a particular belief at some later point
in time.
21. S ee Steup (2016).
22. Steup (2016).
65
66
decide to step on the brakes half a second later. Third, even if I am mis-
taken about these two points, and clamping an artery and stepping on
the brakes are actions that one cannot delay in these circumstances, they
are still actions that one can normally delay. They are the types of actions
that in many circumstances one can decide to do right now or at a later
point in time. With belief, however, things are different: for almost all
beliefs, it seems that one can (virtually) never decide to form them at
some later point in time.
I conclude that the Up- to-
me-ness Argument and the Delay
Argument provide us with enough reason to think that, even though
we decide to consider our evidence regarding certain propositions and,
therefore, decide to weigh our epistemic reasons, we do not decide to form
a belief or come to hold a belief as the (non-deviant) direct result of an
intention to do so. We lack direct intentional control over our beliefs.
2.5 S
ECOND RESPONSE:
WE HAVE INDIRECT INTENTIONAL
DOX ASTIC CONTROL
The second main response to the argument from doxastic involun-

tarism is that we have indirect control over at least some of our beliefs
and that we have at least some doxastic obligations in virtue of that.
Henry Price, John Heil, and Nikolaj Nottelmann have championed this
view. 23
In order to assess this response to the argument, let us first clearly
distinguish indirect control over beliefs from influence on beliefs. 24 The
main difference between the two is that indirect control over φ-i ng
requires that one foresees that one will φ as a result of one’s voluntary
23. See Heil (1992, 51), Huss (2009, 256–2 61), Nottelmann (2007, 157–159), Price (1954,
16–21). Naylor (1985, 427–436) has also argued that we have indirect control over a
significant body of our beliefs, but she does not relate this putative fact to the issue of
doxastic responsibility.
24. Unfortunately, this important distinction is often overlooked. See, for example,
Buckareff (2004, 179–182), Dretske (2000, 600), Huss (2009, 256–2 61), Mele (1987,
110), Tidman (1996, 275).
66
67
actions or omissions, whereas influence on φ-i ng requires that one lacks

such foresight.25 Thus, I foresee that eating less and running for half an
hour a day will result in my losing weight. It follows that I have indirect
control over it and that I can, therefore, do it intentionally. I do not fore-
see which particular experiences I am going to have or which particu-
lar beliefs I will acquire as a result of investigating a murder. It follows
that I have mere influence on my experiences and beliefs in this case and
that, therefore, I do not have those experiences intentionally and that
I do not acquire those beliefs intentionally. Slightly more formally:
S has indirect control over φ-i ng iff (i) S can intentionally φ by

(8)
performing a series of different actions over a considerable
period of time, and (ii) S can ~φ.
S has influence on φ-ing iff there is some action or series of
(9)
actions χ such that (i) S has control over χ-i ng, (ii) if S χ-s, S will
φ, and if S does ~χ, S will ~φ, (iii) S cannot φ intentionally. 26
Let me first explain why Alston thinks we lack indirect doxastic control.
According to Alston, we only incidentally succeed in getting ourselves
to believe something, while what is needed for indirect doxastic con-
trol and for doxastic obligations is being generally successful or reliable in
producing the belief that we intentionally set out to acquire.27 One may
have one’s doubts about Alston’s argument, though. Imagine that I can-
not reliably open the sandwich box, because it malfunctions in a large
25. Henceforth, I mostly talk about actions, but mutatis mutandis the same thing can be
said about omissions.
26. Notice that the fact that S cannot φ under one description of φ-i ng does not entail that
S also cannot φ under another description of φ-ing. Imagine, for instance, that David
reads a Wikipedia article on Germany in order to find out whether the proposition p
that Germany has more than 60,000,000 inhabitants is true. Then, it seems, he intends
to form the correct doxastic attitude toward p, that is, belief in case p is true and disbe-
lief in case p is false. In such a scenario, he has intentional control over forming the cor-
rect doxastic attitude toward p, but not over believing that p, even though belief that p is
the correct attitude toward p. In what follows, I will be concerned with indirect control
over doxastic attitudes under the description of “belief that p,” “disbelief that p,” and
“suspending judgment on p.”
27. S ee Alston (1989d, 134–135).
67
68
portion of cases. I know that the only way to open the box is to press
7, even though I also know that often it does not open upon pressing
7. However, as a matter of fact, if I press 7, the box will open. It seems that
one’s intuitions can go two ways here. On the one hand, one could think
that I lack control over opening the box and that I only have an obliga-
tion to try to open the box. On the other hand, one could think that I do
have control over opening the box, since I will be successful if I try, and
that I therefore have an obligation to open the box. If this is correct, then
we have doxastic obligations in those cases of indirect control in which
we would be successful if we were to try. Since it is not entirely clear
what our intuition should be in this case and since, as we shall see, there
are cases in which I can reliably exercise indirect doxastic control, my
assessment of an approach in terms of indirect doxastic control will not
rely on this controversial issue.
I said that there are cases in which we can exercise indirect doxastic
control. In fact, I think there are at least two kinds of such cases. First,
as Richard Feldman has pointed out, it seems that we sometimes have
indirect control over our beliefs in virtue of the fact that we have indi-
rect control over the actualization of certain states of affairs in the world
and the fact that generally our beliefs automatically and correctly track
those states of affairs.28 For example, I have control over my belief that
I am running in virtue of my control over whether I run or not. One may
object that this control comes in at the wrong juncture: what is relevant
for doxastic responsibility is what our beliefs are given the way the world
is. Below, I argue that this is correct for some, but not all, cases of dox-
astic responsibility. Instead, I provide a different argument against the
view that we bear doxastic responsibility in virtue of indirect doxastic
control.
Second, Alston’s examples of both actions and beliefs are cases in
which a person intentionally actualizes a certain state of affairs that was
not actualized, e.g., my coming to weigh 150 pounds or my coming to
believe that my sister loves me. Alston’s argument seems convincing
for such cases. However, there are also cases of indirect control over
28. S ee Feldman (2001, 81–82).
68
69
maintaining a certain state of affairs. There are certain things I can do

to keep my weight roughly, say, 160 pounds. One could argue that, simi-
larly, we sometimes intentionally maintain a doxastic attitude and that
we do so much more often than intentionally acquiring a doxastic atti-
tude. It even seems that intentionally maintaining a doxastic attitude is
generally successful, since, say, avoiding all evidence that may contra-
dict one’s belief is a successful and frequently practiced way of maintain-
ing a belief. A neo-Nazi might successfully intentionally maintain his
belief that Jews are inferior to Aryans by attending neo-Nazi meetings
and shunning all personal contact with Jews. 29
Thus, these two points show that we have a certain amount of indi-
rect control over some of our beliefs. Nevertheless, I think that such con-
trol is largely irrelevant to doxastic responsibility, because they cannot
explain why we sometimes believe responsibly and sometimes culpably.
Let us first consider Feldman cases, that is, cases in which we have
indirect control over the (non-)actualization of certain states of affairs
that our doxastic attitudes automatically track. If I decide to run, I will
normally acquire the belief that I am running, and, since I simply see and
feel that I am running, that will be a true belief, a rational belief, a piece of
knowledge, and so forth. If I decide not to run, I also normally acquire or
maintain a true belief, a rational belief, and a piece of knowledge, namely
that I am not running, for I simply see and feel that I am not running.
I say “normally,” because I may, of course, malfunction or I may be in a
Gettier scenario or some such exceptional situation. It seems, though,
that if I do not meet my obligation in virtue of some such fact, there are
two options. Either I am not culpable at all—for instance, because, due
to no fault of my own, I am in a Gettier scenario and fail to form a war-
ranted belief—or I am culpable, but not because I failed to exercise indi-
rect doxastic control, but because I failed to meet some other obligation,
such as working on a malfunctioning cognitive mechanism. Thus, to the
extent that we have an epistemic obligation to hold a particular belief,
such as holding a rational belief, in virtue of our indirect control over the
29. Wolterstorff (2010b, 75–76) also points out that we sometimes intentionally maintain
certain beliefs, but his account of doxastic responsibility is largely cashed out in terms
of doxastic influence. See Wolterstorff (2010a).
69
70
world, we can hardly fail to meet that obligation, and to the extent that
we fail to meet such an obligation, it is not our (not) exercising indirect
control that explains why we do.
Might we then have a moral or prudential rather than an epistemic
doxastic obligation in some Feldman cases in virtue of our indirect
doxastic control? The problem is that those cases in which believing or
not believing a proposition might add significant moral or prudential
value to our lives or those of others are usually not cases in which we
have indirect doxastic control over the states of affairs those beliefs are
about. We may enhance our happiness by coming to believe that God
exists and we may increase our chances of survival if we believe that
we will not die from a serious illness that we have, but the existence of
God and our dying from a serious illness are precisely things over which
we—normally, at least—lack control. 30 Thus, it seems, in most Feldman
cases, we do not have a moral or prudential doxastic obligation, and to
the extent that we have an epistemic doxastic obligation, we can hardly
fail to meet it. But if in Feldman cases we either have no obligation or
inevitably meet it, then Feldman cases are unable to explain why we
sometimes believe responsibly and sometimes culpably.
Let us, therefore, turn to the second kind of cases in which we have
indirect doxastic control, namely certain cases of belief maintenance
rather than belief formation. As I said, we can exercise indirect control
over maintaining our beliefs by avoiding all kinds of evidence to the
contrary.
Now, it seems problematic to say that we could have an epistemic obli-
gation to maintain a belief by shunning all evidence to the contrary. If the
30. Here is one example that has been suggested to me by Anthony Booth. If for a moral or
prudential reason I want to come to believe that the next appointment with my dentist
is on April 29, I can write that in my agenda, knowing that I will have forgotten by then
that it is actually on April 22. However, it is not easy to construe examples like these
in such a way that the point comes through. I have to have an obligation to come to
hold this belief, I should genuinely forget that I have an appointment on April 22, my
belief that I perform such self-deceptive actions should not prevent me from forming
the belief that the appointment is on April 29, and so forth. Specifying such scenarios
in this way will make clear how highly exceptional they are. For some such examples,
see Kelly (2002, 168–171).
70
71
obligation is to be an epistemic one, we should not concentrate on part of

the evidence, but take all the relevant evidence into account. One may
object that there might be situations in which one has good reason to think
that the evidence one does not possess is misleading and that from an
epistemic point of view one should, therefore, shun any further evidence.
I think there are indeed such cases. We should note, though, that in order
for these cases to count as cases of indirect doxastic control, it should—on
(ii) of (8), our definition of indirect control—be the case that I can also not
believe that p. In other words, it should be the case that if I do gather the
misleading evidence, I come to believe that p. After all, if I would still hold
the belief that p if I did gather the misleading evidence on p, then, in shun-
ning any evidence against p, I do not exercise indirect control over holding
the belief that p. (If one thinks this is not necessary, then one embraces
some kind of doxastic compatibilism. I return to that in section 2.6.) It
seems that in most cases in which I believe that p and I know that there is
misleading evidence against p that I should shun, I would not abandon my
belief that p if I were to acquire the relevant evidence, precisely because
I know that it is misleading evidence. This is not to deny that it is possible
that there are such scenarios, but from what I have said it should be clear
that if there are such cases, they are very few. Thus, situations in which we
have indirect doxastic control will hardly ever be situations in which we
have an epistemic obligation to exercise that control.
Might we then have a moral or prudential obligation to maintain a
belief by exercising indirect control over that belief in shunning evi-
dence against it? Well, what would a situation look like in which we have
such an obligation? Presumably, there would be something valuable
from a moral or prudential point of view about having the belief that p, or
something morally or prudentially harmful or bad about not having the
belief that p. Surely, that seems possible. Sometimes, it is morally or pru-
dentially valuable to believe that one is not ill and shun all evidence to
the contrary, because it makes one’s life and that of others much better.
The problem is that it is not clear that in such scenarios we can exercise
indirect doxastic control. For in order to have indirect doxastic control,
it should be the case that if one does not shun the evidence, one no longer
maintains one’s belief that p, and knowing that one shuns all evidence
against p because it is morally good to believe that p usually undermines
71
72
one’s belief that p, for any rational agent will realize that the fact that it
is morally or prudentially good to believe that p does not count in favor
of p’s truth. Knowing that one is intentionally shunning evidence that
bears on the truth of p and that might alter one’s belief that p suffices,
in normal circumstances, for automatically suspending judgment on p
(or at least no longer believing that p), no matter how much one might
want to believe that p for moral or prudential reasons. This means that
in such cases we virtually always lack indirect doxastic control and that
means—as we saw above and as is captured in (7″)—t hat we do not have
doxastic obligations in virtue of being in such scenarios.
In summary, it seems that we do sometimes have indirect doxastic
control, namely in Feldman cases and some cases of belief maintenance,
but such scenarios do not explain why we bear doxastic responsibility.
This is because in such scenarios we either have no obligation to exercise
that control or, to the extent that we do, we can hardly fail to meet it.
2.6 THIRD RESPONSE: WE HAVE

COMPATIBILIST DOX ASTIC CONTROL
A third strategy to meet the argument from doxastic involuntarism is

doxastic compatibilism. 31 Compatibilism in the theory of action is cru-
cially different from compatibilism in epistemology. Whereas the com-
patibilist in the theory of action claims that the inability to act otherwise is
compatible with having an obligation to act and with being responsible
for one’s action, the doxastic compatibilist believes that the absence of
intentional doxastic control is compatible with having a doxastic obliga-
tion and with being responsible for what one believes. According to the
doxastic compatibilist, having the kind of control necessary for having
a doxastic obligation does not require that one is able to intentionally
acquire or maintain a belief. 32 This means that, depending on precisely
31. W hat I argue in this section and the next section is based on Peels (2014a).
32. Thus, compatibilism in the theory of action and doxastic compatibilism are not strictly
analogous. The compatibilist in the theory of action might very well maintain that
one is responsible for φ-i ng only if one can φ intentionally. All she denies is that being
72
73
how control is spelled out, she replies to the argument from doxastic
involuntarism by denying either the premise of doxastic involuntarism
or the premise that doxastic obligations require doxastic control.
If doxastic obligations do not require intentional doxastic control,
then what is required? Doxastic compatibilists answer this question dif-
ferently, although their accounts show important similarities. Let me
briefly characterize three important compatibilist accounts of doxastic
responsibility that we find in the literature. First, according to Mark
Heller, people have an epistemic nature, that is, second-order desires to
form beliefs in accordance with certain dispositions rather than others.
What is required for doxastic control is that our beliefs reflect our epis-
temic nature, that is, that we form our beliefs due to the epistemic nature
we have. We are responsible in such cases because our beliefs manifest
who we are from an epistemic point of view. 33
Second, according to Sharon Ryan, what is required for doxastic
responsibility is that we can appreciate evidence and form beliefs in
accordance with it, in the same way as we can weigh various practical
considerations and act in accordance with them. We are responsible for
our beliefs if they are unlike typical coerced actions and like actions
such as typing the letters that we type and moving our limbs when we
have been running for a while. These are actions that are responsive
to reasons, but that do not seem to involve the formation of any inten-
tions. We mean to perform such actions and in that sense we perform
them intentionally. This does not require an explicit intention to perform
them. It follows that intentional doxastic control is not necessary for
having a doxastic obligation. 34
Third, according to Steup, some person is responsible for a belief if
that belief is the outcome of a process that is responsive to epistemic
reasons, i.e., evidence, and if that person’s belief is weakly intentional.
responsible for φ-i ng entails that one could ~φ. Doxastic compatibilism is not the view
that responsibility for belief does not entail that one can believe otherwise. That view
will be discussed in c hapter 4.
33. S ee Heller (2000, 132–137).
34. See Ryan (2003, 70–74). An account similar to that of Ryan is Owens (2000, 115–129).
His account is different in that he is not willing to describe reasons-responsiveness as
control; he simply denies that doxastic responsibility requires any kind of control.
73
74
Something is weakly intentional if it is non-accidental and if one has a

pro-attitude toward it. Thus, my stepping on the clutch is weakly inten-
tional if it does not result from such things as a sudden cramp in my leg
and if I mean to step on the clutch. A belief is weakly intentional if it is
not due to cognitive malfunction and if one endorses it or if one is com-
fortable with it. Beliefs, then, are like actions such as starting a car by
inserting the ignition key, engaging the clutch, shifting into reverse, and
stepping on the gas. These automatic, unthinking, and habitual actions
are not performed as the result of an explicit intention, but are neverthe-
less under one’s control. 35
Let us consider these accounts in this order. As to Heller’s account,
imagine that Nagoni is raised in a culturally isolated community. From
an early age onward, she is indoctrinated in the tradition of the tribe.
She therefore thinks that she should believe anything that the tradition
teaches and so she desires to believe anything that is part of her tradi-
tion. Thus, upon considering the proposition that humans are fallen
angels, she immediately believes this proposition, for this is one of the
core teachings of her tribe. She believes this as strongly as anything
she could possibly believe. It is completely irresistible. Since she is con-
vinced that this is what the tradition teaches, she believes in accordance
with her epistemic nature. It is clear, however, that she has no control
over her belief: she has been indoctrinated to such an extent that her
belief is genuinely irresistible. This shows that Heller’s account in terms
of epistemic natures is untenable. The reasons-responsiveness accounts
of Ryan and Steup do not face this problem, for Nagoni’s belief-forming
mechanism clearly is not reasons-responsive: she would hold that belief
no matter what her evidence is.
Ryan’s and Steup’s accounts face another problem, though. Imagine
a possible world in which there are creatures who are like us in that their
belief-forming mechanisms are largely functioning properly: upon hav-
ing the experiences and beliefs we have, they roughly form the same
beliefs as we do. In one regard, however, they are crucially different from
35. See Steup (2012). For other, less developed compatibilist accounts of doxastic control
in the same spirit, see Hieronymi (2008, 362–363), Jäger (2004, 217–227), Shah (2002,
443), Smith (2000, 240–2 46; 2005, 236–271), Weatherson (2008, 546).
74
75
us: they cannot influence what they believe. Thus, they cannot gather evi-
dence, work on their intellectual virtues and vices, improve the func-
tioning of their cognitive mechanisms, and so forth. For instance, they
cannot decide to think about something or reflect on their reasons to
believe something. Would we hold those creatures responsible for their
beliefs in this recherché scenario? It seems clear to me that we would not.
Their belief formation is not up to them. Their beliefs are simply the
deliverances of their cognitive mechanisms in combination with certain
inputs. But, we have assumed, neither the functioning of their cogni-
tive mechanisms nor the scope or quality of their evidence base is up to
them. It seems clear that if these are not up to them, the output is not up
to them either and it would be unfair to hold them responsible for their
beliefs.
I submit, then, that the plausibility of Ryan’s and Steup’s doxastic
compatibilism derives from the fact that in the scenarios they sketch,
people have control over all sorts of factors that influence what the
subjects believe. It is because we can intentionally perform such belief-
influencing actions and not just because our belief-forming mechanisms
respond differently to different inputs that we can be held responsible
for our beliefs in such cases. After all, if we have not been able to work on
our belief-forming mechanisms (which would be an instance of doxastic
influence), we are simply stuck with certain mechanisms that respond to
a certain input by automatically producing a certain output. It is hard to
see how we could ever be responsible for the beliefs—t he outputs—t hat
such mechanisms produce.
Doxastic compatibilist control, then, does not seem to be the kind
of control in virtue of which we can have doxastic obligations. However,
we should prefer the view that doxastic obligations require intentional
rather than compatibilist doxastic control only if it does not face equally
insurmountable difficulties. Let me, therefore, discuss the three main
arguments that compatibilists have leveled against the idea that doxastic
obligations require intentional doxastic control.
First, Steup and Ryan object that we are responsible for and have
control over actions such as my typing the letters that I type, my mov-
ing my legs when I have been running for a while, and my stepping on
the clutch, even though we often do not perform these actions as the
75
76
result of a preceding intention to do so. Hence, forming an intention to

φ is not necessary for having control over φ-ing, for being responsible
for φ-ing, or for having an obligation to φ. In response, I would like to
point out that beliefs are crucially different from such actions as mov-
ing my legs and typing the letters I type. I can type the letters on my
computer without explicitly intending to do so, only because I once did
intend to type these letters, and something similar applies to actions like
running. I do not think that this is true for all actions that are under
our control, though. I can exercise control over breathing or not breath-
ing, even if I have never intentionally done so. However, breathing or not
breathing—or, at least, breathing or not breathing at some particular
moment—is something that I could in principle do intentionally. With
belief, things are different. I do not now unintentionally form certain
beliefs because I once intentionally formed them. Nor is it the case that
I could in principle form a belief as the result of an intention to do so.
Rather, I have always found myself with certain beliefs, given the evi-
dence I had. Beliefs may be quite different from compelled actions, but
they are not very much like the unintentional actions just mentioned
either: we not only do not form them as the result of an intention, we
normally cannot do so.
Second, Steup, Ryan, and Bruce Russell argue that what is neces-
sary for control and responsibility, both when it comes to action and
when it comes to belief, is reasons rather than intentions. Action and
belief are similar in that if one takes oneself to have sufficient reason to
perform act A or hold belief B, one will normally perform A or hold B.
For instance, if I have convincing reasons to slow down and stop when
I approach an intersection, I will stop, and if I have convincing reasons
not to stick a knife in my arm, I will not do so. Of course I could decide
to act differently if I had an overriding reason to do so, but similarly
I could believe differently if my epistemic reasons were different. 36 I find
this reply unconvincing. First, if I take myself to have good reasons to do
A, I will not do A as long as I do not also intend to do A (unless doing A
36. For this argument, see Russell (2001, 42–43), Ryan (2003, 63–6 4), Steup (2000, 46,
54; 2001, 17n).
76
77
is a habitual action, as we saw above). Reasons alone, then, will not suf-
fice to explain action: we also need intentions. Second, there are many
situations in which my practical reasons for doing A and doing ~A are
balanced. In such situations I can choose to do A or ~A by forming an
intention to do A or ~A. For instance, I can choose to put cheese or pea-
nut butter on my bread. However, in all or most situations in which my
evidence for and against p is balanced, I simply cannot decide to believe
or disbelieve that p. Rather, I automatically find myself with a particular
doxastic attitude toward p, normally that of suspension of judgment.
The third and, as it seems to me, most promising objection has been
leveled by Steup and Conor McHugh. 37 It runs as follows. We lack con-
trol over our intentions, but our intentions are still free; if that is the case,
we have good reason to think that our beliefs are also free. To say that
φ-i ng freely requires the ability to form an intention to φ is to measure
the realms of action and belief with the same yardstick, whereas they
should be measured with different yardsticks. Freedom and responsibil-
ity in the realm of action should be measured in terms of state-directed
reasons, that is, considerations in the light of which the subject’s being
in that state would be a good or bad thing. Freedom and responsibil-
ity in the realm of belief and intention, however, should be measured
in terms of object-directed reasons, that is, considerations pertaining to
the world’s being as it is represented in the state’s content. After all, our
intentions are free, but, according to Steup and McHugh, to say that
they are free only if they are under our intentional control leads to two
problems.
First, as Gregory Kavka’s famous Toxin Puzzle shows, we lack control
over our intentions. In his puzzle case, I am offered a reward for form-
ing the intention to drink a mild toxin tomorrow. It seems that I cannot
form the intention, since I know that as soon as I have formed the inten-
tion, when the time comes, I no longer have any reasons to drink the
toxin. 38 We cannot form intentions to φ as a result of deciding to intend
to φ upon believing that intending to φ would be good. We can intend
37. S ee McHugh (2014); Steup (2008, 387–390; 2016).

38. F or a description of the case, see Kavka (1983).
77
78
to φ only if we believe that φ-i ng itself is good. As Jonathan Owens and

Pamela Hieronymi point out, we are deluded into thinking that we have
intentional control over our intentions, because an intention is usually
good to have just in case the action intended is good to perform. Kavka’s
Toxin Puzzle and other cases show that we lack intentional control over
our intentions. 39
Second, according to Nishi Shah, the idea that responsibility for
φ-ing requires the ability to form an intention to φ leads to an infinite
regress. One’s intention to φ is under one’s control only if one could form
an intention to intend to φ. But that intention would be under one’s con-
trol only if it could be formed by a further intention, and so on. In order
to avoid this infinite regress, we should deny that control over φ-i ng
requires the ability to φ intentionally.40
I think that these two points are convincing. Normally—we will see
below that there are some exceptional circumstances—I cannot con-
trol my intentions, and if intentional control over φ-i ng were necessary
for responsibility for φ-ing, we would not be responsible for our inten-
tions. But it seems we are responsible for our intentions. Imagine that
I am an adult and properly functioning human being and that I decide
to attempt to murder the British queen. Thus, I form the intention to
murder her. I have not yet had any time to make plans, let alone attempt
to murder her. We can even stipulate that the queen is so well guarded
that it is impossible for me to murder her. Still, it seems that in this situ-
ation, I am responsible and in fact blameworthy for intending to murder
the queen and that in forming the intention I violate an obligation not to
form such an intention.
Thus, the view that one is responsible for φ-ing only if one can φ
intentionally is unconvincing. There is, however, a plausible alterna-
tive that does not commit us to the idea that intentions and beliefs are
equally free. For one could say that one has control over φ-ing and one
is responsible for φ-ing only if φ-ing, in the situation at hand, can be the
intended product of our uncoerced will or the exercise of our uncoerced
39. Th is point has been made by Owens (2000, 81), Hieronymi (2006, 56–57; 2008,
368–371).
40. F or this line of reasoning, see Shah (2002, 440–4 42).
78
79
will itself. Thus, many of our actions are under our control because they
are the uncoerced products of our will: we perform them because we will
to do them and are not coerced in willing to do so. For actions, this means
that we form an uncoerced intention to perform the action. Intentions
themselves are usually also free and we are usually responsible for them,
because they are the exercise of our uncoerced will: to will something
without being coerced to do so is to intend to do something without being
coerced to form that intention. Beliefs, however, are neither the product
of our uncoerced will nor the exercise of our uncoerced will: we do not
believe as the result of willing to believe something nor is believing some-
thing identical to willing something without being coerced to do so.41
Hence, the fact that our intentions are under our control and such
that we are responsible for them and have obligations with regard to
them provides us with no reason to think that our beliefs are also under
our control and that we have doxastic obligations. However, is there
anything we can say against the idea, advocated by Steup and McHugh,
that intentions and beliefs are equally free because they are both equally
responsive to object-d irected (rather than state-d irected) reasons?
According to Anthony Booth, there is: intentions, he suggests, are
free by definition: doing something intentionally logically entails that
one does it freely.42 This seems unconvincing to me. Imagine that some-
one suffers from a brain tumor and that as a result of that serious inflic-
tion he decides he will attempt to murder the president. He carefully
plans the assassination for weeks and on the day the president visits his
town, he leaves his house in order to kill the president. It seems entirely
unproblematic and true to say that he intends to murder the president,
even though we all know that his intention is not free and that he is not
responsible for having it, because it is due to a brain tumor. According to
Booth, our intuition that such an agent has an intention can be explained
away by acknowledging that the subject falsely believes that he acts freely
and falsely believes he has an intention to perform the action in question,
in this case a murder.43 But this seems misguided to me: not only the
41. F
or a similar thought, see Booth (2009a, 9–11).
42. See Booth (2014, 1874).
43. S ee Booth (2014, 1875).
79
80
deluded subject, but we as well, who are fully informed, would say that
the subject intends to murder the president. Imagine that we found out
about his plans on the morning of the president’s arrival. Then, it seems,
there would be nothing incoherent about saying: “We need to find him
as soon as possible, because, due to a brain tumor, he intends to murder
the president.”
Yet it seems to me that the idea that intentions and beliefs are equally
free because they are equally responsive to object-d irected reasons is
problematic for at least two other reasons. First, if something counts as
free and we are responsible for that thing merely in virtue of the fact
that that thing is responsive to object-d irected reasons, then we will be
responsible for all sorts of things for which we are clearly not respon-
sible and that are clearly not free.44 For instance, my fear is responsive
to reasons: when a wild, large, hungry grizzly bear suddenly breaks into
my office, I am scared, but I am not scared when my small and hungry
cat enters my living room. But, clearly, my fear is normally not under my
control or free when a large, wild, hungry grizzly bear breaks into my
office.
Second, if I have equally good reason to do A as to do ~A and equally
good reason to intend to do A as to intend to do ~A, as is often the case,
then I can equally well intend to do A as intend to do ~A. For example,
if I do not mind whether I take coffee or orange juice for breakfast, I can
equally well intend to take coffee as I can intend to take orange juice.
Whether I intend to do some action A or ~A is in such cases up to me. If
you offer me a reward for intending to do A in such a case, I can easily
choose to intend to do A. However, my doxastic attitude is never up to
me, not even in cases in which my evidence regarding p is balanced, as
we saw above. Thus, it seems that I have a kind of control over my inten-
tions that I lack over my beliefs, however precisely this kind of control is
to be spelled out.
Before we turn to the fourth and final response to the argument from
doxastic involuntarism, there is an important question that I would
like to address: does the falsehood of doxastic compatibilism imply
44. B
ooth (2014, 1872–1873) also makes this point.
80
81
the falsehood of compatibilism in the realm of actions? That would be

quite a disturbing conclusion, given the large number of philosophers
who embrace the latter thesis. Fortunately, no such thing follows.45
Compatibilism in the realm of actions, such as that advocated by Harry
Frankfurt, John Martin Fischer, and Mark Ravizza, only denies that
responsibility for φ-i ng implies the ability to ~φ. That view is clearly per-
fectly compatible with the thesis that one is responsible for φ-i ng only if,
in the situation at hand, φ-i ng can be the intended product of our unco-
erced will or the exercise of our uncoerced will itself.
2.7 F
OURTH RESPONSE: DOX ASTIC
OBLIGATIONS WITHOUT ANY KIND
OF CONTROL
The final strategy to maintain the idea that we have doxastic obligations
is to argue that doxastic obligations belong to a special kind of obliga-
tions that do not require any sort of control whatsoever. In this section,
I discuss four proposals along these lines.46
First, according to Richard Feldman, doxastic obligations are role
obligations. Parents ought to take care of their children, teachers ought
to explain things clearly, and cyclists ought to cycle well, whether they
are able to do so or not. Similarly, judgments on belief prescribe the right
way to play the role of a believer, even if one has no control over one’s
belief.47 Feldman confines his account to epistemic doxastic obligations,
that is, “evaluations that have more to do with epistemologically cen-
tral matters such as knowledge and rationality.”48 The right way to play
the role of a believer, according to Feldman, is to believe in accordance
45. According to Ryan (2003, 47, 70) and Steup (2008, 375, 390), compatibilists about
action have good reason to be or are even committed to being doxastic compatibilists.
And Jäger (2004, 217–223) wrongly assumes that the argument from doxastic involun-
tarism relies on the idea that control entails the ability to believe otherwise.
46. There are further proposals along these lines, e.g., Engel (2009). The main problem that
I raise for these four proposals applies to such other proposals as well, though.
47. S ee Feldman (2000, 676–677).
48. F eldman (2008, 346).
81
82
with one’s evidence. One might object that there is a crucial difference
between the role of believer on the one hand and the roles of, say, teacher
and cyclist on the other. For, in opposition to the former role, one has
the latter roles voluntarily. Feldman agrees, but points out that there
are many other roles that are as involuntary as that of believer, such as
the roles of eater and breather. That these roles are involuntary does not
imply that there are no correct ways to eat or breathe.49
I think that the analogy fails. We do have control over the way
we eat or breathe. To the extent that we do not, we are not the proper
object of positive, negative, or neutral appraisal for eating or breathing.
It seems unfair to hold me responsible for the way I breathe if there is
nothing I can do to change it. Perhaps we can hold a teacher responsible
for teaching badly, even if she could not do any better. If so, however,
that would be because the role of teacher is a role she has voluntarily
accepted. Thus, to the extent that role obligations imply responsibility,
they require some kind of control. Such control is remarkably absent in
the case of being a believer: I have control neither over my beliefs nor
over my being a believer. Feldman points out that we sometimes praise
a person for things that are not under her control, such as her beauty. 50
As I argued in c hapter 2, though, some praise is merely evaluative rather
than normative. We would not praise a girl for her beauty in the norma-
tive sense, in the same way as we would not blame another person for
her ugliness. 51
Second, Hilary Kornblith construes doxastic obligations in terms
of ideals that take into account human limitations. More specifically,
he says that what he has in mind are epistemic ideals. Epistemic ideals
take into account what humans can believe, but they are not confined to
what a particular human can believe: sometimes one ought to believe a
49. See Feldman (2008, 351). Feldman (2000, 674) distances himself from his earlier
view on which doxastic obligations are contractual obligations; see Feldman (1988a,
240–2 43).
50. S ee Feldman (2000, 676).
51. Feldman (2001, 77) includes praise and blame for beliefs among the phenomena to be
explained. In Feldman (2008), however, he defends what he calls “modest deontolo-
gism,” according to which we have doxastic obligations, but are usually not blamewor-
thy or praiseworthy for our beliefs. Here, I have argued that the fact that there are role
oughts provides us with no reason to think that we have doxastic obligations. That
82
83
proposition that one cannot believe. Epistemic ideals can be reached by

at least some human beings. Doxastic obligations, therefore, provide the
middle ground between epistemic ideals that are insensitive to human
capacities and ideals that are so constrained by a particular individual’s
capacities that they are unworthy of pursuit. Since being subject to epis-
temic ideals does not require control, this approach would save doxastic
obligations. 52
It seems to me that this account is flawed. The problem with our
beliefs is that they seem under no one’s control. Thus, even if we confine
ourselves to epistemic ideals that can be reached only by those who do
epistemically well, there will not be any doxastic obligations. No mat-
ter what the ideal attitude is toward a proposition, if there is no way I or
anyone else can voluntarily acquire that attitude, it seems that neither
I nor anyone else has an obligation to acquire that attitude. In fact, this
follows from proposition (7″), which concerns the relation between con-
trol and obligations that I defended in section 2.3 above. 53 Again, we
may well be the proper objects of evaluative attitudes for not reaching
certain epistemic ideals. It follows neither that we are the proper objects
of normative attitudes nor that we have doxastic obligations. Of course,
one could say that we have an obligation to believe that p if and only if
one ought to believe that p and that one ought to believe that p if and
only if the epistemic ideal is to believe that p. Perhaps the word “obliga-
tion” is sometimes used in this contrived sense. But then, as I argued in
chapter 2, we will no longer be talking about doxastic responsibility.
epistemic oughts to believe in accordance with one’s evidence are not to be understood
in terms of responsibility is also argued by Dougherty (2012b, 534–538).
52. See Kornblith (2001, 238–2 39). In a previous article—Kornblith (1983, 33)—he
distinguishes between doxastic obligations as epistemic ideals and doxastic obliga-
tions that imply responsibility. In an even earlier paper, he understands doxastic
justification in terms of the absence of epistemic culpability; see Kornblith (1982,
243). Unfortunately, the distinction between ideals and responsibility is absent from
Kornblith (2001) and (2002, 137–161).
53. Proposition (7″) says that S at t has an obligation to φ only if (i) at t people normally have
control over φ-ing, or (ii) S normally has control over φ-ing, or (iii) S at t has control
over φ-i ng. We saw, though, that people’s beliefs are normally not under their voluntary
control and that this is true for people generally.
83
84
Third, Matthew Chrisman argues that we should think of doxastic

obligations as what Wilfrid Sellars calls rules of criticism that materi-
ally imply rules of action. Whereas rules of criticism (ought-to-be’s)
concern ways of being, rules of action (ought-to-do’s) concern actions.
One can be subject to a rule of criticism for φ-i ng, even if one has no
control over φ-ing, whereas rules of action do require control. The
material implication that Chrisman has in mind can be spelled out as
follows:
(10) If X ought to be in state Φ, then, other things being equal and

where possible, one ought to bring it about that X is in state Φ.
Now, Chrisman rightly notices that (10)’s consequent can be interpreted

in at least three different ways: (i) on the conditional view, X herself ought
to do what she can to bring it about that X is in Φ, (ii) on the universal view,
everyone ought to do what she can to bring it about that X is in Φ, and (iii)
on the existential view, someone ought to do what she can to bring it about
that X is in Φ. According to Chrisman, which view is correct depends on
which kind of ought is involved. For instance, the rules of criticism:
(a) people ought to feel outrage about genocide;

(b) this child ought to be able to tie his shoes by age four;
materially imply respectively a universal and an existential rule of

action:
(a′) everyone ought to do what she can to bring it about that peo-
ple feel outrage about genocide;
(b′) someone ought to do what she can to bring it about that this
child is able to tie his shoes by age four.
According to Chrisman, the epistemic ideal for humans is to be good

information-tracking and -transmitting beings. Ought-to-do’s implied by
these doxastic ought-to-be’s are interpersonal and sometimes intraper-
sonal forward-looking and backward-looking rules of action on the part of
one’s epistemic community, which sometimes includes oneself. Thus,
84
85
(c) you ought to disbelieve that the earth is flat;
is an ought-to-be that materially implies the following correlative exis-

tential ought-to-do:
(c′) your parents and teachers ought to have taught you that the
earth is not flat. 54
If this account of doxastic obligations is correct, then they do not require

any doxastic control on the cognitive subject’s part. However, there is good
reason to doubt that it is. Chrisman rightly acknowledges that there are
many situations in which S in some sense ought to believe that p, without
S’s herself being subject to a materially implied ought-to-do. But then it is
incorrect to say that S is subject to an obligation to believe that p or that she
is blameworthy if she fails to meet that obligation. Assuming that there are
such things as epistemic obligations on the side of one’s epistemic com-
munity, one’s epistemic community may have an obligation to convince
one of the truth of p and may even be blameworthy for failing to do so, but,
clearly, that is something different from the subject’s being doxastically
responsible. The sense in which S ought to believe that p in such situations,
then, will not be normative, but merely evaluative. Let me be explicit. I do
not deny that some normative doxastic judgments express rules of criti-
cism that imply rules of action. That may be right. But such judgments can
very well be correct without there being any doxastic obligations. Hence,
Chrisman’s account of doxastic oughts in terms of rules of criticism and
rules of action does not meet the argument from doxastic involuntarism.
Finally, according to Philippe Chuard and Nicholas Southwood,
normative doxastic judgments are like judgments such as:
(d) Oscar ought to feel guilty for what he did to his sister, and
(e) Judy ought to understand what Nicole is going through,
in that these judgments make demands on us. They demand that we

respond in certain ways. But making a demand to φ on some person S
54. S ee Chrisman (2008, 358–363, 369–370).
85
86
does not presuppose that S has control over her φ-ing. It only presup-
poses that S can φ. The idea that S can φ can be spelled out in terms of
(i) there being the logical possibility that S φ-s, (ii) having the alternate
possibilities of φ-ing and ~φ-ing, or (iii) S’s having the capacity to φ.
But none of (i)–(iii) requires that S have intentional control over φ-i ng.
Alston’s argument goes wrong, then, in conflating the idea that S can φ
with the idea that S has control over φ-i ng. Hence, doxastic obligations
do not require doxastic control. 55
It seems to me that this approach faces a serious difficulty. For it is not
clear what it is to be the proper object of a demand. If it does not have to
do with being responsible, then Chuard’s and Southwood’s strategy does
not even address the argument from doxastic involuntarism. If it has to do
with being responsible, then it follows from what I argued in chapter 1 that
if a demand is made on someone, then that person is blameworthy if she
fails to meet that demand without being excused.
Chuard and Southwood, however, explicitly reject the idea that dox-
astic obligations are in any way relevantly related to blameworthiness.
They do so because they do not find any plausible interpretation of blame
on which blameworthy belief requires doxastic control. As they see it,
blame can be interpreted in terms of being criticizable or in terms of other
people having certain legitimate expectations toward one. As they rightly
point out, being criticizable does not require control. And someone else’s
having a legitimate expectation about what one will believe is an implau-
sible way of spelling out doxastic blame, since people often do not have
any expectations about what other people will believe.56
However, they overlook the option that I defended in chapter 1,
namely that one is blameworthy just in case one is the proper object of
a negative normative attitude, such as resentment or blame. I think sen-
tences (d) and (e) can plausibly be understood along these lines, that is, as
expressing normative attitudes toward Oscar and Judy. But if they do, then
the speaker, it seems, will at least tacitly assume that there is something
Oscar and Judy could have done about respectively not feeling guilty and
55. See Chuard and Southwood (2009, 601, 614–619).

56. See Chuard and Southwood (2009, 620–623).
86
87
lacking understanding. If there is nothing they could possibly do or have

done in order to feel guilty or to understand, then it is hard to see how we
could properly blame them. How, for instance, can we rightly blame Oscar
for not feeling guilty for what he did to his sister if he has tried everything
he possibly could, but still fails to feel guilty? Thus, a person is properly
subject to some demand only if that person has control over meeting that
demand. However, we saw that people have virtually no control over their
beliefs and that to the extent that they do such control is irrelevant to dox-
astic responsibility. The only option left, then, is to argue that we can influ-
ence what we believe and that we are doxastically responsible in virtue of
such influence. I defend this option in chapter 3.
In this section, I have criticized four attempts to argue that doxastic
obligations belong to a special kind of obligations that do not require any
kind of control. For each of the four kinds of obligations that I discussed—
role obligations, epistemic ideals, rules of criticism, and demands—I agree
that there may be some sense of the word “obligation” in which we have
doxastic obligations along these lines or at least that it can be true that we
ought to hold some belief. In all four cases, however, it turned out that
such oughts have nothing to do with responsibility. And that is because
they have nothing to do with praise, blame, neutral appraisal, or any of the
other normative attitudes that I distinguished in chapter 1. To the extent
that they do have to do with these attitudes, they do seem to require con-
trol. Of course, this does not establish that there is no way that one could
possibly argue that doxastic obligations are of a special kind that does not
require control. I take it, though, that my treatment of these four proposals
provides sufficient reason to seek an alternative solution to the problem
of doxastic involuntarism until a better version of this strategy shows up.
2.8 C ONCLUSION
Let me draw the threads of this chapter together. If one is to meet the
argument from doxastic involuntarism, one has to deny one or more of
its three premises. First, one could deny premise (3), the premise that
we lack doxastic control. I showed that the attempts to argue that we
are doxastically responsible in virtue of our presumed direct doxastic
87
88
control, our indirect doxastic control, and our compatibilist doxastic

control fail. Second, one could deny premise (2), the premise that dox-
astic obligations require doxastic control. One could do so by arguing
that certain kinds of obligations, of which doxastic obligations are a
special instance, do not require any kind of control. I showed that four
proposals along this line are seriously wanting. This leaves us with only
one strategy to meet the argument. That strategy is to deny premise (1),
which says that S responsibly (blamelessly) believes that p only if S has
not violated an obligation not to believe that p or if S met her obligation
to believe that p. 57 I will defend that option in detail in c hapter 3.
57. One might think that there is another option, namely to deny (2) by arguing—against
virtually all views about obligations and control—t hat one can have an obligation to φ
in virtue of one’s influence on rather than control over φ-i ng. Thus, all that doxastic obliga-
tions would require is doxastic influence. Since we can only assess this view if we first
have a firmer grip on doxastic influence, I return to this suggestion in c hapter 3.
88
89
[3]
THE SOLUTION
Doxastic Influence and Intellectual Obligations
3.1 I NTRODUCTION
In c hapter 2, I argued that doxastic responsibility and doxastic obliga-

tions are not to be understood in terms of control over our beliefs. We
have virtually no direct or indirect intentional doxastic control, compat-
ibilist doxastic control cannot account for doxastic responsibility, and
views on which doxastic obligations require no control whatsoever do
not seem to concern doxastic responsibility. In this chapter, I defend the
only option that seems left to meet the argument from doxastic invol-
untarism. I argue that to believe responsibly is a matter of not having
violated an obligation to influence what we believe rather than an obliga-
tion to have or maintain some belief. Other philosophers, such as Robert
Audi, John Heil, Anne Meylan, and Miriam Schleifer-McCormick, have
also adopted this position, but they have not spelled out exactly which
intellectual obligations we have to influence our beliefs and how they
relate to responsible belief.1 The aim of this chapter is to do precisely
that. Doing so will lead us, in section 3.7 and in the ensuing chapters,
into the territory of excuses for belief.
1. See Audi (2001, 93–111), Axtell (2011a; 2011b), Clarke (1986, 39–49), Dretske (2000,
602), Heil (1984, 60), Kornblith (1983, 39), Leon (2002, 421–424), Meylan (2013,
64–126; 2015), Schleifer-McCormick (2015, 91– 127), Stocker (1982, 398– 417),
Wolterstorff (2010a).
89
90
The chapter can be outlined as follows. First, in order to get a

fix on what intellectual obligations are, I argue that the main belief-
influencing factors that we often control are our doxastic mechanisms,
cognitive situatedness, and intellectual virtues and vices (§3.2). Next,
I defend the thesis that an obligation to perform a belief-influencing
action is an intellectual obligation only if not performing it leads to
beliefs that are in some sense bad—epistemically, morally, or pruden-
tially (§3.3). Using the conclusions of sections 3.2 and 3.3, I provide an
account of what kinds of intellectual obligations we have. As it turns out,
some intellectual obligations are contingent upon certain roles or pro-
fessions, whereas others are obligations that all people have, including
moral and epistemic intellectual obligations (§3.4). Next, I defend my
account, especially that of epistemic intellectual obligations, against six
objections that might be leveled against it (§3.5). Subsequently, I show
precisely how responsible belief relates to whether or not one has met
one’s intellectual obligations (§3.6). Finally, I suggest that even if one
has violated an intellectual obligation, one can still responsibly hold a
belief that issues from that violation, namely if one is excused for that
in a particular way (§3.7). This suggestion is elaborated and defended in
the ensuing chapters.
3.2 B ELIEF-I NFLUENCING FACTORS
In c hapter 2, in contrasting indirect control with influence, I defined

influence as follows:
(1) S has influence on χ-ing iff there is some action or series of

actions φ such that (i) S has control over φ-i ng, (ii) if S φ-s, S
will χ, and if S does ~φ, S will ~χ, (iii) S cannot χ intentionally.2
2. See proposition (9) on p. 67. For expository purposes, I have changed the order of the
variables φ and χ. This is because, whereas in chapter 2 the focus was on control over
believing, in this chapter the focus is on control over factors that influence what we believe.
90
91
T he S olution : D oxastic I nfluence and I ntellectual O bligations
In asking by which factors we influence what we believe, then, I mean

to be asking which states of affairs there are such that (i) we frequently
control whether or not they are actualized, (ii) whether or not they are
actualized makes a difference to whether or not we believe a particu-
lar proposition p, and (iii) in exercising control over them, we do not
intentionally acquire, maintain, or abandon a belief that p. Let me pres-
ent what I consider to be the main belief-influencing factors that we
frequently control. Doing so will help us to get a firmer grip on our intel-
lectual obligations:
A. Doxastic mechanisms
A1. The functioning of doxastic mechanisms
A 2 . The creation or elimination of doxastic mechanisms
B. Cognitive situatedness
B1. Being in a situation in which one acquires evidence of a
certain kind
B2 . Being in a situation in which one loses evidence of a
certain kind
C. Intellectual virtues and vices
C1. The quality of intellectual virtues and vices
C2 . The creation or elimination of intellectual virtues
or vices
C3. Intellectually virtuous or vicious behavior in particular
processes of belief formation or belief maintenance3
This scheme calls for substantial terminological elucidation. First, by

“doxastic mechanisms,” I mean mental faculties that produce or main-
tain doxastic attitudes: belief, disbelief, and suspension of judgment.
They have certain doxastic attitudes or experiences (or both) as their
3. Most authors overlook one or more of these categories of belief-i nfluencing factors. Heil
(1983, 361–362) focuses on A and B at the expense of C. Clifford (1901) and Levy (2007,
144–148) focus on B at the expense of A and C. Kornblith (1982, 253) focuses on C at
the expense of A and B.
91
92
input and doxastic attitudes as their output.4 For example, belief forma-
tion on the basis of introspection and belief formation on the basis of
visual perception count as doxastic mechanisms. There is, of course, an
important debate in epistemology about how finely grained these doxas-
tic mechanisms are supposed to be. We need not launch into that debate
here. All that my argument requires is that we have doxastic mecha-
nisms and that we sometimes indirectly control their functioning. Thus,
I can improve my reasoning faculties by taking a course in modal logic.
One cannot only increase the reliability of one’s doxastic mechanisms;
one can also eliminate some of them. By blinding myself, I remove the
doxastic mechanism of visual perception. If doxastic mechanisms are
sufficiently fine-g rained, I can also acquire some of them. Thus, people
can acquire the doxastic mechanisms of forming beliefs by way of car-
diogram reading and using modal logic. Clearly, exercising control over
such factors often makes a difference to what we believe.
Second, we frequently control our cognitive situatedness. I use the
expression “cognitive situatedness” to make clear that what we indi-
rectly control in many situations is not our having particular evidence,
but our having evidence of some kind or other. For instance, by critical
reflection on a moral issue or by reading an article on herons, I may influ-
ence what I believe about that moral issue or about herons. But I do not
thereby control the particular evidence I acquire, because in advance
I have no idea what particular evidence I will acquire by reflecting or by
reading the article and, therefore, do not intentionally acquire a particu-
lar piece of evidence or a particular belief.
Third, by “intellectual virtues” I mean such cognitive dispositions
as open-mindedness, intellectual courage, carefulness, precision, dili-
gence, and thoroughness that influence processes of belief formation
and maintenance. And by “intellectual vices” I mean cognitive disposi-
tions like epistemic conformity, intellectual laziness, and epistemic self-
indulgence. Thus, whether one does something impulsively, meticulously,
lazily, thoughtlessly, or thoroughly often makes a difference to our beliefs.
I can train and, thus, indirectly control myself to be more intellectually
4. Th is understanding of “doxastic mechanism” is highly similar to Alston’s understanding

of what he calls “doxastic practice”; see Alston (1991, 153).
92
93
thorough and thereby strengthen that intellectual virtue or even acquire

it if I was not intellectually thorough at all. Moreover, on specific occa-
sions, I can behave virtuously or viciously. I can be intellectually virtuous
but fail to exercise those virtues on a particular occasion. And I can be
intellectually vicious but nevertheless act intellectually praiseworthily in
exceptional circumstances. Virtues are dispositions and having a disposi-
tion does not guarantee that one always acts as one is disposed to.5
A–C often overlap. Thus, increasing the reliability of your reasoning
faculties by taking a course in modal logic is also to exercise control over
your cognitive situatedness. For you know that by doing so, you acquire
some kind of evidence on certain modal propositions. And if you train
yourself to be open-m inded by seeking to meet people with views that
differ from your own views, you intentionally strengthen your intellec-
tual virtue of open-m indedness, but also purposely change your cogni-
tive situatedness.
Nevertheless, they remain distinct categories, because there are situ-
ations in which these belief-influencing factors come apart. Imagine that
I intentionally undergo an ear operation in order to improve my audi-
tory perception. Ten days later, on my way to work, when passing by an
open window, I hear someone sing the song, famously first recorded by
Louis Armstrong, “What a Wonderful World.” I would not have heard it
if I had not had the operation. But I did not intentionally change my cog-
nitive situatedness on this particular issue. It merely happened to me
because I had worked on my doxastic mechanisms. Nor did I change
any intellectual virtues or vices in this case. Next, I can decide to con-
sult an encyclopedia article on herons. That makes a difference to my
cognitive situatedness, but usually it does not change the functioning
of my doxastic mechanisms or the quality of my intellectual virtues.
Finally, imagine that by training myself, I become more intellectually
focused and concentrated and that I exercise that virtue when I am
presented with a riddle. I have not improved my doxastic mechanisms
5. Of course, this is not to say that all doxastic mechanisms, cognitive situations, and intel-
lectual virtues and vices are under our indirect control. Many of them are not, such as the
fact that we can form beliefs by way of mathematical intuition. Yet A–C are often under
our control, and in cases in which they are, we can influence our beliefs.
93
94
or voluntarily changed my cognitive situation. Rather, because I have

worked on my ability to concentrate, I find the solution to the riddle that
otherwise I would not have found and thereby form a belief that other-
wise I would not have formed.
Some philosophers might think that a substantial category of belief-
influencing factors is missing from the above list. Let me, therefore, dis-
cuss two other phenomena that have been taken as belief-influencing
factors that we sometimes control. First, there are belief-forming dispo-
sitions.6 For instance, I have the disposition to acquire the belief, upon
hearing the doorbell, that there is someone at the door. However, many of
these dispositions are not dispositions over which we have control. Take
the belief-forming disposition that I just mentioned. It seems that already
as a child, I noticed that every time the doorbell rang, there was someone
at the door. At a certain point, I had experienced this often enough to auto-
matically acquire the corresponding belief-forming disposition. Certain
belief-forming dispositions, then, seem to be beyond our control. Other
belief-forming dispositions may be under our control or, at least, under
our influence. For instance, it seems that certain people have strong rac-
ist belief-forming dispositions and that they could have avoided acquir-
ing such dispositions. In fact, it seems that we blame them because they
could have avoided acquiring such dispositions. What we should notice,
though, is that to the extent that people have control over or influence
on such things as racist dispositions, it is in virtue of their control over
the three belief-influencing factors that I distinguished above. Someone
who purposely develops racist dispositions can only do so by shunning
certain bodies of evidence and by developing intellectual vices, such as
narrow-mindedness. Thus, such belief-forming dispositions can well be
explained in terms of the control that I spelled out above.
Second, there are belief policies, that is, epistemic standards to which
we subject our beliefs. According to Paul Helm, for instance, we control
our belief policies. Among the epistemic standards that we could adopt
are the policies to hold a belief if we have sufficient evidence for it and
6. For the notion of belief-forming dispositions or belief-forming habits, see, for instance,
Alston (1989d, 141).
94
95
that of holding a belief only if it is the outcome of scientific research.

According to Helm, our evidence can never all by itself establish a belief
policy, so that it has to be adopted at least partly on non-evidential
grounds. We are responsible for our beliefs to the extent that we are
responsible for our belief policies and to the extent that those belief poli-
cies make a difference to what we believe.7
It seems to me that this approach is flawed. 8 According to Helm,
belief policies are themselves beliefs.9 This seems correct for at least
some belief policies, because sufficiently sophisticated cognitive sub-
jects hold beliefs as to which beliefs they ought to have. But if they are
beliefs, then how can they be under our control? I hold beliefs about
what counts as sufficient evidence to hold some beliefs, but I cannot
intentionally change those meta-beliefs. I agree that it is not clear what
constitutes my evidence for such beliefs, but it does not follow that
I can voluntarily change such beliefs. Helm could change gears and say
that belief policies are not believed but accepted, so that one can decide
to reason from and act on a certain belief policy.10 This would solve the
problem of involuntarism, for reasoning from and acting on proposi-
tions is often voluntary. Unfortunately, this move gives rise to another
problem. If belief policies are merely accepted, then it is hard to see how
they could influence beliefs—e xcept, of course, beliefs about those
acceptances. If I do not believe, but merely accept that I ought to believe
something only if it is the result of scientific research, then that will
usually not make a difference to what I find myself believing. For, as is
widely accepted among epistemologists, if I take some consideration
only to count in favor of believing that p and not in favor of p’s truth,
then I normally cannot hold that belief for that reason.11 Thus, if belief
policies are accepted rather than believed, they are causally inert in the
7. See Helm (1994, 58–8 4).

8. E lsewhere, I have argued this in detail. See Peels (2013a).
9. See Helm (1994, 115).
10. Buckareff (2004, 179–182) adopts this position. Something like this view is suggested
by Helm (1994, 58) himself when he says that a belief policy is a “strategy,” “project,”
or “programme.”
11. P amela Hieronymi argues this in detail in Hieronymi (2005; 2006).
95
96
sense that they do not cause first-order beliefs and, therefore, cannot
account for our doxastic responsibility.
I conclude that doxastic mechanisms, cognitive situatedness, and
cognitive virtues and vices are the main belief-influencing factors over
which we have control and in virtue of which we have influence on our
beliefs.
3.3 W HAT ARE INTELLECTUAL

OBLIGATIONS?
Many philosophers agree that having control is necessary for having an

obligation (“ought” implies “can”), but not sufficient. I have control over
the order in which I eat the peas on my plate, but I do not have an obliga-
tion to eat them in a particular order. Thus, having control over a belief-
influencing factor does not imply that one has an obligation to exercise
that control. What, in conjunction with a control condition, is sufficient
for having an intellectual obligation?
Some moral philosophers have argued that we have an obligation to
φ if φ-i ng is objectively good or best, that is, such that a fully informed
and morally conscientious agent would pursue it,12 where the objective
good can be spelled out in a utilitarian, deontological, or virtue theo-
retical way. Others have argued that we rather have an obligation not to
perform any objectively bad actions.13 Thus, I merely have an obligation
not to kill, not to bully, not to rape, and so forth, because all these actions
would actualize an objectively bad state of affairs. Again, badness can
be spelled out in terms of bad consequences, not acting in conformance
with rules, and exemplifying vices. It does not follow from this rather
negative characterization of obligations that we never have an obligation
to actually do something. For example, if I promise you to wash your car,
then I have an obligation to do so, for not washing your car would be to
do something bad, viz., to fail to fulfill my promise.
12. E
.g., Ross (1939, 30–45), Wedgwood (2007, 126–132).
13. Th
us, for instance, Russell (2001, 35).
96
97
The discussion between adherents of these two kinds of views is

intricate and entering this debate would lead us too far afield. For two
reasons, I will assume that we have an obligation merely to avoid objec-
tive badness rather than to do the objectively best. First, we often do
not know what the best option is that is available to us. We hardly ever
know the best way to maximize pleasure and minimize pain in particu-
lar situations and we hardly ever know what a perfectly virtuous and
fully informed agent would do in the circumstances we find ourselves in.
One worry with regard to the idea that we have an obligation to do the
very best or one of the best actions we could possibly do is that it has the
implausible implication that we are almost always blameworthy for what
we do or excused for the vast majority of our actions. But blameworthi-
ness is relatively rare. And, as I said in the Introduction to this book,
one’s being excused should be the exceptional rather than the normal
situation. Second, adherents of the view that we should do the objec-
tively best agree that we have at least an obligation not to perform any
objectively bad actions. Hence, although they are likely to distinguish
more obligations, they can at least in principle agree that we have an
obligation to avoid objective badness and, hence, the intellectual obliga-
tions that I specify in section 3.4.
Let me add that it is widely agreed that we have not only an obliga-
tion to avoid objective badness, but also an obligation to avoid subjective
badness.14 For instance, it seems that if I believe on the basis of good
evidence that a girl is drowning in the canal in front of me, I have an obli-
gation to take steps toward what seems to me the act of saving her, even
if, unbeknownst to me, the girl is not drowning, but just playing the role
of a drowning girl in a television program. There is some disagreement
on how subjective obligations are to be spelled out—in terms of what
one believes about one’s obligation or in terms of what one believes about
objective badness, in terms of what one believes or in terms of what one
should believe.15 Again, this debate need not concern us here. I just want
to point out that the badness relevant for obligations can be understood
14. W
olterstorff (2008, 257), for instance, points this out.
15. See, for instance, the discussion in Smith (2011b).
97
98
both objectively and subjectively. As I argue in section 3.4, the fact that
we have an obligation to avoid not only objective, but also subjective
badness, is crucially relevant when it comes to intellectual obligations.16
One might think that if what I have argued so far is true, then we
have an intellectual obligation to φ if and only if the following three
conditions are met. First, the control condition is satisfied: most people
have control over φ-i ng, one normally has control over φ-i ng, or one has
control over φ-i ng on the occasion in question. Second, whether or not
one φ-s makes a difference to what one believes. Third, ~φ-i ng is objec-
tively or subjectively bad.
However, the following example shows that this cannot be correct.
Imagine that I freely rob an old man and thereby find out that he has
thirty-five dollars in his wallet. In such a case, I meet the three condi-
tions. First, I exercise control by robbing the old man. Second, robbing
the man is to perform a belief-i nfluencing action, for by robbing the man
I change my cognitive situatedness. If I had not violated my obligation
not to rob him, I would not have acquired the belief that he has thirty-
five dollars in his wallet. Third, on virtually any account of obligations
I have an obligation not to rob the man, for doing so is clearly to actual-
ize an objectively bad state of affairs. However, it seems false that I have
an intellectual obligation not to rob the man and it would surely be incor-
rect to say that my belief that he has thirty-five dollars in his wallet is
blameworthy. The case can easily be generalized. We have many obliga-
tions and virtually all of those obligations are such that whether or not
we meet them makes a difference to what we believe. But then pretty
16. A few other epistemologists also distinguish between objective and subjective doxas-
tic and intellectual obligations. See, for instance, Goldman (1986, 73–74), Hall and
Johnson (1998, 130), Plantinga (1990, 53), Pollock and Cruz (1999, 141–142), Vahid
(1998, 287).
One might think that control over φ-i ng in conjunction with the objective or subjec-
tive badness of failing to φ does not suffice for having an obligation to φ. Some philoso-
phers think that what, in addition to this, is necessary is that φ-ing is required of one,
where requirement can be understood in terms of, say, social expectations (e.g., Adams
(1999, 231–276)) or human worth and divine command (e.g., Wolterstorff (2008, 285–
310, 362–384)). Since this is a controversial issue that can be discussed separately, I will
not treat it here. Those who think that requirement is also necessary can add a require-
ment clause to my account of (intellectual) obligation.
98
99
much any obligation would count as an intellectual obligation and that

would result in the denial of responsible belief in many cases in which
we clearly believe responsibly.
I suggest that the solution to this problem is to narrow down the third
condition that I mentioned above. More specifically, I have an intellec-
tual obligation to φ only if the badness that will be brought about by fail-
ing to φ is at least partly constituted by my acquiring or failing to acquire
beliefs that are in some sense bad. Thus, on the one hand, I clearly do not
have an intellectual obligation not to rob the man, because the badness
of doing so has solely to do with the well-being and physical integrity
of the man, not with the beliefs ensuing from my robbing him. On the
other hand, it seems that I do have an intellectual obligation to prepare
for my biology exam because the badness brought about by not doing so
is at least partly constituted by my remaining ignorant on certain top-
ics that I should not be ignorant about. And I do have an intellectual
obligation not to spy on my niece, because the badness of doing so is at
least partly constituted by my coming to know certain things about her
private life that I should not know. Of course, doxastic attitudes may be
bad in different ways. To be ignorant on, say, photosynthesis as a result
of failing to prepare for one’s biology exam is bad from an educational
point of view. To have knowledge of my niece’s love affairs that I have
acquired by way of spying on her is perhaps morally bad. Similarly, it
seems, one’s beliefs can in principle be bad from an epistemic, prudential,
or professional point of view (below, I will spell out in some more detail
what such badness might amount to).
If what I have argued in this section is correct, we should understand
intellectual obligations as follows:
(2) Some person S has an intellectual obligation to φ iff (i) S has

control over φ-i ng, (ii) whether or not S φ-s will make a dif-
ference to S’s beliefs, and (iii) ~φ-ing is objectively or sub-
jectively bad in that it leads to or maintains beliefs that are
objectively or subjectively bad.
One might wonder why we should call such obligations intellectual obli-
gations. After all, they are practical obligations, that is, obligations to act
99
100
or not to act in certain ways. I agree that they are practical in the sense
that they are obligations to perform or not to perform certain actions.
However, they are practical obligations of a particular kind, namely obli-
gations to perform a belief-influencing action. The word “intellectual” is
meant to single out those practical obligations that we have in virtue of
the fact that not performing them would lead to objectively or subjec-
tively bad beliefs. Calling such obligations merely practical obligations
would fail to do justice to the property that they have that is relevant for
the purpose of constructing a theory of responsible belief, namely that
of being an obligation to perform a belief-i nfluencing action. And, as we
saw in chapter 2, calling them doxastic obligations would be misleading,
since, as I argued, we have (virtually) no obligations to hold or not to
hold specific beliefs.
3.4 W HAT INTELLECTUAL OBLIGATIONS

DO WE HAVE?
Now that we have strengthened our grip on belief-influencing factors

and intellectual obligations, let us consider what kinds of intellectual
obligations we might have.
There are certain intellectual obligations that we have contingently.
By that I mean that we have them in virtue of such contingent and vol-
untary acts as accepting a certain job, committing oneself to some task,
and making a promise. As a philosophy student, Sam has an obligation,
in preparing for his modal logic exam, to work on his logic skills. And
the police have an obligation to carry out meticulous investigation after
the assassination of the minister of foreign affairs.
Virtually all adults are subject to such contingent intellectual obliga-
tions. First, of course, many of us have voluntarily committed ourselves
to certain professions that bring along obligations to investigate certain
events, gather data, check documents, and prepare meetings. Arthur
Coningham, for instance, had an intellectual obligation carefully to
investigate the situation before he gave the order to attack all German
ships in the Baltic (for this example, see the Introduction). Second,
100
101
large numbers of people have committed themselves to specific roles

that bring along certain intellectual obligations. Commitment to par-
enthood, for instance, comes with an obligation to acquire information
about pregnancy and infant care. Finally, we often choose to bear cer-
tain tasks that bring along specific intellectual obligations. For instance,
I promise you to check the train timetables and thereby subject myself
to an intellectual obligation to check the train timetables.
These are intellectual obligations because whether or not we meet
them makes a difference to what we believe. Whether the police inves-
tigate things carefully can make a difference to their beliefs about the
culprit. These are intellectual obligations because the subjects in ques-
tion have control over performing these belief-influencing actions and
because not to perform them would be to actualize a bad state of affairs.
For a policeman not to investigate the murder would be to fail to do what
he has committed himself to doing in voluntarily accepting the profes-
sion of policeman. That makes the violation of that obligation profession-
ally bad. Not to acquire information about pregnancy upon deciding to
have children is morally bad in that failing to be informed might put the
child’s health at risk. In both cases, the badness is at least partly consti-
tuted by the badness of certain false beliefs or ignorance.17
A question that many philosophers will be interested in is whether
there are any non-contingent intellectual obligations, obligations that we
have, not in virtue of a voluntary commitment of some sort, but simply
in virtue of being human. I think we do. For instance, Julia has a non-
contingent moral intellectual obligation not to spy on her niece, for it is
morally bad to acquire knowledge about such private matters and mor-
ally wrong (at least prima facie) to do so intentionally. If Julia violates
this intellectual obligation, the moral badness involved is at least partly
constituted by Julia’s having certain knowledge (which entails certain
beliefs) about her niece’s private life that she should not have had.
Philosophers are likely to be especially interested in epistemic obli-
gations, so let us ask whether we have any epistemic non-contingent
17. C lifford’s essay “The Ethics of Belief ” is often interpreted as describing our moral duty
to inquire. See Clifford (1901).
101
102
intellectual obligations.18 In order to answer this question, let us first

consider what would count as epistemic badness. Many philosophers
believe that the hallmark of the epistemic is that it has to do with the
Jamesian goal—named after William James19—of believing the truth
and avoiding falsehoods.20 Assuming that this is correct, we can say
that the epistemic aim is twofold: to believe true propositions and not
to believe false propositions. This means that there are three main ways
in which one can reach the epistemic aim: one believes (i) more true
propositions and fewer false propositions, or (ii) more true propositions
and an equal number of false propositions, or (iii) an equal number of
true propositions and fewer false propositions.
It seems to me that epistemic intellectual obligations should not be
construed in terms of (i) and (ii). I know, for instance, that by reading a
book on herons I will most probably not acquire any false beliefs (or, if
so, only a few), probably acquire many true beliefs, and perhaps abandon
a couple of false beliefs. If we had an epistemic obligation each time we
can reach aims (i) and (ii), it would follow that I have an epistemic obli-
gation to read everything I can on herons, on French existentialism, and
even on just any individual, such as Justin Bieber. I would also have an
obligation to increase the reliability of almost any doxastic mechanism
and I would have an obligation to work on all my intellectual virtues and
vices. In fact, I would have an epistemic obligation to do almost any-
thing, for many actions will be likely to result in more true beliefs and
unlikely to result in more false beliefs. This counts as a reductio, for a
view that entails that we have an epistemic obligation to perform virtu-
ally any action is clearly untenable. One could retort that we have pro
tanto obligations to do each of these things. But that still seems to multi-
ply obligations far beyond plausibility: it seems false that I violated mil-
lions of obligations today by not reading books about herons or articles
on French existentialism and Justin Bieber.
18. Here, I assume with most epistemologists that there are indeed such things as purely
epistemic obligations. For a long, detailed pragmatist argument to the contrary, see
Schleifer-McCormick (2015, 16–17).
19. S ee James (1979, 24).
20. See, for instance, BonJour (1980, 54), Foley (2005a, 317), Nottelmann (2007, 55).
102
103
This means that several views on our intellectual obligations that we

find in the literature are mistaken. For instance, according to Richard
Hall and Charles Johnson, we have an intellectual obligation to gather
further evidence on p if p is less than certain on our evidence or if we are
not certain about p. And according to Laurence BonJour and Kihyeon
Kim, we have an intellectual obligation to critically reflect on our
beliefs.21 If one of these views were true, we would have far too many
obligations. We would be blameworthy or excused almost all the time.
It seems that we do not even have an epistemic obligation to reach
aim (iii). Each of us holds a substantial body of false beliefs. People hold
widely divergent views on religious, philosophical, moral, and political
issues. Since many of these views are mutually exclusive, it follows that
the majority of people hold false beliefs on these matters. Many of our
beliefs about what happened in the past, especially the distant past, are
false. We often misperceive things, we make mistakes in logical reason-
ing, and so on. Now, presumably, for many of those false beliefs there
is some action that we could perform, such as careful investigation or
consulting experts, which would rid us of those beliefs. It would follow
that we have numerous epistemic intellectual obligations that we fail
to meet. Again, it seems to me that a view on which we violate the vast
majority of our obligations and are blameworthy or excused for that is
implausible: to be excused should be an exceptional situation, not the
normal situation. Moreover, since all beliefs that a subject has, whether
they are true or false, are beliefs that some proposition is true, true and
false beliefs are indistinguishable from the subject’s perspective. But
then we would have no idea how to go about meeting our epistemic intel-
lectual obligations. We would have to check virtually all of our beliefs.
I conclude that epistemic intellectual obligations should not be spelled
out merely in terms of avoiding false beliefs.
Views on which we have an epistemic obligation to avoid other kinds
of epistemic badness face similar worries. Here, I think of irrational
belief (unreasonable belief, belief that does not fit one’s evidence), unre-
liably formed belief (belief produced by a mechanism that is aimed at
21. S ee Hall and Johnson (1998, 133), BonJour (1985, 42), Kim (1994, 284).
103
104
truth, but produces mostly false beliefs), and unwarranted belief (where
warrant is that which turns true belief into knowledge). These are, in
addition to falsehood, the prime epistemically undesirable properties of
beliefs that philosophers distinguish. The problem here is, again, that it
seems that many of us hold large numbers of beliefs that are irrational,
unreliably formed, or unwarranted, too large a number to have obliga-
tions to remove each of them. Moreover, from our own perspective, we
often cannot distinguish these beliefs from our other beliefs. In fact, for
many irrational, unreliably formed, and unwarranted beliefs that we
hold, we hold them precisely because we are unaware that they are irra-
tional, unreliably formed, or unwarranted. We only rarely believe that
some belief that we continue to hold is irrational, unreliably formed, or
unwarranted.
However, we should remember that, as I said in section 3.3, there
are not only objective, but also subjective obligations: obligations that
arise from the fact that something is bad from the subject’s perspective.
Applying this to the issue under consideration, we could say that we do
not have an epistemic obligation if we can avoid having a false belief,
but if we believe that we can. This solves both problems that I identified
above. First, it renders many epistemic obligations transparent to us—
we realize which ones we have. Second, it reduces the scope of our obli-
gations. Since, presumably, one can no longer believe that p if one comes
to believe that p is false, we only have an epistemic intellectual obliga-
tion to perform certain belief-influencing actions in cases in which we
believe that not performing a belief-influencing action is epistemically
bad in that not performing it is likely to lead to forming or maintaining
false beliefs.
Thus, the suggestion is that subjective epistemic obligations should
be understood in terms of what one believes. I prefer this to understand-
ing subjective obligations in terms of what one should believe. I cannot
defend this preference in detail here, but I would like to give at least one
reason to prefer the former over the latter approach. I argued that dox-
astic responsibility should be understood in terms of intellectual obliga-
tions. It follows that if one should believe that p, then one has violated an
intellectual obligation at some earlier time. But if subjective obligations
are understood in terms of what one should believe rather than in terms
104
105
of what one believes, then it seems that the chain of intellectual obliga-
tions would be endless. This is because an intellectual obligation would
have to be explained in terms of what one should have believed at some
earlier time and that would have to be explained in terms of what one
should have believed at an even earlier time, and so forth.
One might object that the strategy to cash out doxastic responsi-
bility in terms of what one believes about avoiding falsehoods is in two
ways ad hoc. First, one might think that it is ad hoc in that it claims that
we have an epistemic obligation only in those cases in which the epis-
temic aim can be reached in the third way, that is, in cases in which one
can avoid having a false belief. In response, let me grant that it would
perhaps be epistemically ideal to reach the epistemic goal in the first way
and epistemically better to reach it in the second way. But obligations are
sensitive to what humans are able to achieve in a way that ideals are not.
It is, therefore, not unduly ad hoc to say that we have an epistemic obli-
gation only in those cases in which the epistemic aim can be reached in
the third way.
Second, one might think that it is ad hoc in that it acknowledges
only subjective epistemic obligations. We saw that when it comes to non-
epistemic intellectual obligations and obligations generally, we also
have an obligation to avoid objective badness—moral, prudential, or
otherwise. It seems that we have now simply changed our view, in order
to account for our epistemic intellectual obligations. I do not think that
this is a problem. When it comes to non-epistemic obligations, we have
an obligation to avoid objective badness, because we often have a pretty
good idea of what counts as, say, a morally bad action and what does not.
And that is because as soon as we know what properties an action has,
we often only need to reflect on them in order to see whether the action
is morally good or bad—in fact, such moral insight is often direct and
does not require much reflection. However, when it comes to epistemic
obligations, things are quite different. For many of our beliefs, such as
virtually all of our perceptual beliefs, we cannot find out merely upon
reflection whether they are true or false, and it is often not clear whether
a belief is irrational or not and whether it is reliably formed or not. This
is simply a matter of how we are constituted. Given the way we are as a
matter of fact constituted, we have good reason to think that epistemic
105
106
obligations and non-epistemic obligations are in this regard different.

I do not think that this is surprising. Many philosophers believe that
what obligations we have depends on such psychological and physi-
cal factors as what people have control over. Moreover, as I explain
below, my argument leaves room for at least some objective epistemic
obligations.
The core thesis that I have defended so far in this section is that one
has an epistemic intellectual obligation to perform a belief-influencing
action A if one believes that it would be epistemically bad not to perform
A in that it leads to or maintains certain false beliefs. Now, which kinds
of scenarios should we think of when I say that someone believes that it
would be epistemically bad not to perform A in that it leads to or main-
tains certain false beliefs? I think our epistemic intellectual obligations
are as varied as our non-epistemic intellectual obligations. Someone
might come to believe that one of her doxastic mechanisms is function-
ing improperly or that she possesses a certain intellectual vice. In both
cases, believing these things usually implies believing that in certain
circumstances—circumstances in which the doxastic mechanism is
used or the vice exemplified—one is likely to acquire a false belief or
several false beliefs. What one should do, then, in such circumstances
is to respectively increase the reliability of one’s doxastic mechanism
and remove the intellectual vice. Of course, one may be excused for
not doing so, for instance, because one is unable to meet the obligation
or because one is ignorant about how one can do so. I return to these
excuses in the ensuing chapters.
These, however, are not the only epistemic intellectual obligations.
Some such obligations arise not from beliefs about our doxastic mecha-
nisms or intellectual vices, but from our beliefs about our beliefs. Since
those will be situations in which one believes that there is something
epistemically bad about one’s beliefs, I will call them cases of doxastic
discrepancy. I think there are several options here. One might believe,
for instance, the proposition r that some belief that p that one holds con-
flicts with another belief that q that one holds in the sense that q entails
~p or that q renders ~p sufficiently likely. Or one might believe the prop-
osition r that some belief that p that one holds does not fit one’s total
evidence. Or one might believe the proposition r that some belief that p
106
107
that one holds was unreliably formed. Of course, the formation of such a
meta-belief normally suffices for automatically abandoning one’s belief
that p, one’s belief that q, one’s belief that r, or several of these beliefs. In
those cases in which it does not, from the epistemic point of view there is
clearly something wrong with one’s beliefs. As we saw in chapter 2, one
cannot have an obligation to abandon one’s belief that p, one’s belief that
q, one’s belief that r, or any belief whatsoever, since we normally lack the
relevant kind of indirect control over our beliefs.
What I would like to suggest is that in cases of doxastic discrepancy,
one’s objective epistemic intellectual obligation is to do that—or one of
the actions—which will remove the doxastic discrepancy, whereas one’s
subjective epistemic intellectual obligation is to do that which from the
subject’s perspective renders it sufficiently likely that the doxastic dis-
crepancy is removed. The belief-influencing action or series of actions
could be such an activity as gathering evidence on p, q, or r, increasing
the reliability of the doxastic mechanism that produced the belief that
p, q, or r, or working on the intellectual virtues and vices that played a
role in the formation of these beliefs. Precisely which activity is required
will, of course, depend on the circumstances. Sometimes it will be criti-
cal reflection, sometimes it will be consulting experts, sometimes it will
be acquiring empirical evidence. Since any of one’s belief that p, one’s
belief that q, and one’s belief that r may be false, ridding oneself of the
doxastic discrepancy would lead to one’s abandoning one or more of
these beliefs.22
But precisely why should we think that we have such epistemic intel-
lectual obligations? Since in cases of doxastic discrepancy, we have con-
trol over performing or not performing an action and it is an action or
omission that (we believe) makes a difference to what we believe, this ques-
tion boils down to the question of what is epistemically bad about doxas-
tic discrepancy as I just spelled it out. I think that what is epistemically
bad about cases of doxastic discrepancy is that they imply either falsehood
or likelihood of falsehood. A sufficiently reflective cognitive subject will
22. For an account of epistemic intellectual obligations that is in some regards similar to
the account offered here, see Anthony R. Booth, “Doxastic Influence and Epistemic
Duty,” unpublished manuscript.
107
108
realize this. Let me explain. First, imagine that I believe that p, that q,
and that p entails ~q. Then one of these beliefs must be false. Either p or q
is false, or p does not entail ~q. This kind of doxastic discrepancy, a kind
that consists in believing that one has inconsistent beliefs, then, is epis-
temically bad in that it guarantees at least one false belief.
Second, imagine that I believe that some belief of mine does not
fit my evidence and is, therefore, irrational. In such cases, doxastic
discrepancy does not guarantee at least one false belief. For it may be
that my irrational belief happens to be true and that I nevertheless
truly and rationally believe that, given my evidence, I should not hold
that belief. What, then, is epistemically bad about such cases? I think
that what is epistemically bad about this second variety of doxastic
discrepancy is that either (i) one holds an irrational belief, a belief
that does not fit one’s evidence, or (ii) if one’s belief is not irrational,
one falsely believes that it is. Not believing in accordance with one’s
evidence, where one’s evidence includes one’s beliefs and one’s experi-
ences, makes it more likely that one holds false beliefs. This means that
(i) is also epistemically bad.
Third, imagine that I believe that some belief of mine is unreliably
formed. Such cases will count as cases of doxastic discrepancy only if
I do not also believe that my belief is nevertheless true. I may have that
additional belief, because I may have independent evidence (evidence
that does not function as the input to the relevant doxastic mechanism)
that confirms that belief. If I do not have that additional belief, then there
is clearly something bad about the situation: either (i) my belief that p
is unreliably formed, or (ii) my belief that my belief that p is unreliably
formed is false. If (ii) holds, then I have at least one false belief. If (i) is
true, then I am likely to have a false belief, because unreliable belief for-
mation implies likelihood of falsehood.
3.5 S
IX OBJECTIONS AND REPLIES
In this section, I discuss six objections that might be leveled against the
account of our epistemic intellectual obligations that I spelled out in
section 3.4.
108
109
1. Imagine that someone firmly believes that she has no false beliefs
whatsoever. My account implies that such a person has no epistemic
intellectual obligations, at least not with regard to any specific beliefs
arising from instances of doxastic discrepancy. But it seems that one
cannot escape epistemic obligations so easily and that stubbornly
believing that one has no false beliefs whatsoever does not get one off
the hook of blameworthiness.
Some reflection on a scenario like this, however, reveals that my
account can make sense of our deontological intuitions about it. First,
since most people believe that two beliefs cannot both be true if they are
inconsistent, if someone believes that two beliefs are inconsistent and
all her beliefs are true, she will automatically abandon one or both of
them, or abandon her belief that all her beliefs are true. Second, it seems
likely that if someone believes that all her beliefs are true, that belief is
itself blameworthy. It could be blameworthy, for instance, because it
issues from narrow-m indedness and one could and should have avoided
such narrow-m indedness. Finally, that a subject is able to escape certain
epistemic obligations does not imply that she can escape blameworthi-
ness. After all, it would still be the case that she would not have held
the relevant beliefs if she had met all her intellectual obligations. In this
regard, there is nothing special about epistemic blameworthiness for
beliefs. Many philosophers believe that something similar applies to
moral blameworthiness for actions and omissions. 23 If I promise you to
water your yard in northern Wales tonight, but then make myself unable
to do so—say, by flying to Haiti—then, on many accounts of obliga-
tion, I do no longer have an obligation to water your yard tonight, since,
according to many philosophers, ought implies can. But, of course, I am
still blameworthy: if I had met all my obligations, I would have been able
to water your yard and I would have done so.
2. It seems that whether or not we have an intellectual obligation to,
say, gather further evidence on some proposition depends on all sorts of
non-epistemic factors, such as how much time we have and what other,
moral or prudential, obligations we have. 24 But then are there any purely
23. E .g., Goldman (1976), Thomason (1981, 180).

24. For this objection, see Feldman (2000, 688–690).
109
110
epistemic intellectual obligations? I think this argument might be con-

vincing if there were only all-things-considered obligations. Perhaps
moral and prudential obligations sometimes or often override our epis-
temic obligations and perhaps there are, therefore, strictly speaking no
purely epistemic all-t hings-considered intellectual obligations. But that
would in fact guarantee that there are purely epistemic obligations, only
they would be merely pro tanto obligations.
3. My account of epistemic intellectual obligations seems to imply
that there are epistemic reasons for actions. After all, it implies that we
sometimes have reasons to meet an epistemic obligation by perform-
ing some action and this suggests that we have epistemic reasons for
actions, in the same way as having reasons to meet one’s moral obliga-
tion by performing an action suggests that we sometimes have moral
reasons for actions.
One might think, though, that, contrary to what my account of epis-
temic intellectual obligations implies, there can be no epistemic reasons
for actions. As philosophers such as Pamela Hieronymi have argued,
epistemic reasons constitute evidence that counts in favor or against (or
neither in favor of nor against) the truth of a proposition. Evidence can-
not count in favor of or against performing an action, for actions do not
have a truth value. It would follow that there cannot be epistemic rea-
sons for actions and, hence, that there cannot be epistemic intellectual
obligations.25
There is a lot to be said about this objection and it has been exten-
sively addressed in the literature.26 Here I will say just this: that evidence
does not only count in favor or against the truth of a proposition should
be clear, for instance, from cases of suspension of judgment. Given our
current evidence base, the rational and epistemically justified doxastic
attitude toward the proposition that the number of stars is even is sus-
pension of judgment: our evidence base counts in favor of that doxastic
attitude rather than belief or disbelief. And, given its current status in
mathematics, our evidence base with regard to Goldbach’s conjecture,
25. Here, one may be inspired by Hieronymi (2005; 2006), Moser (1989, 47–52).
26. S ee especially Booth (2006; 2009b).
110
111
which says every even integer greater than 2 can be expressed as the sum
of two primes, counts in favor of suspending judgment on it. After all,
there is some evidence for it, such as that it has been shown to hold up
through 4x1018, but the evidence seems insufficient to make belief ratio-
nal. Or, to mention one final sort of case, there are many complex crimi-
nal investigations in which there is substantial evidence to think that
a particular suspect is guilty but equally strong evidence to think that
she is not. Again, in such cases, our evidence seems to count in favor of
suspending judgment. But here is the point: suspension of judgment is
neither true nor false. It is merely a particular attitude toward a proposi-
tion that is true or false. Hence, epistemic reasons (evidence) can count
in favor of things that are neither true nor false.27
One may object that all sorts of reasons may count in favor of believ-
ing that p that are not epistemic reasons (they do not constitute evi-
dence). For example, believing that p might make me happy or make me
behave better toward my sister. Clearly, these are not epistemic reasons.
Thus, why should we think that there are epistemic reasons for or against
certain actions, namely certain belief-i nfluencing actions? In response,
let me point out that all this is perfectly compatible with what I said
above. There may be other, non-epistemic, reasons for believing that p,
such as the fact that it makes me happy. There are good ways, though,
to distinguish epistemic reasons from non-epistemic reasons. Epistemic
reasons differ from other reasons in that they count in favor or against
something—a belief or an action—in virtue of that thing’s relation to
the aim of having true rather than false beliefs. Moral, prudential, and
other non-epistemic reasons do not meet this criterion.
4. Some philosophers claim that believing that p impedes further
investigation on p. According to Charlotte Katzoff, for instance, to believe
that p is to take it that one need not further investigate whether p, and
David Owens says that belief that p is a block to open-mindedness and
considering contrary evidence concerning p.28 This means that, even if
one believes that some proposition p that one believes contradicts some
27. I thank Matthias Steup for helping me improve the argument here.

28. S ee Katzoff (2000, 94), Levy (2007, 145), Owens (2000, 142–145, 157).
111
112
other proposition q that one believes, or that one’s belief that p is irrational
or unreliably formed, one cannot have an epistemic intellectual obligation
to perform certain belief-influencing actions concerning p. For believing
that p makes it psychologically impossible for one to do so. I find this
objection unconvincing. For, as is widely acknowledged, believing that p
does not entail that one is certain that p. If a doctor believes that the result
of a test is that her patient has brain cancer but she is not quite sure, she
easily could and should check the results of that test. Surely, in general
if we believe that p, we are not inclined to further investigate whether p,
because usually when we believe that p we take ourselves to have good
reasons to think that p is true. But sometimes we do not take ourselves
to have such reasons or believe that so much hangs on whether or not p is
true that we need not just good reasons, but conclusive reasons. In those cir-
cumstances, it seems perfectly possible to further investigate whether p.
5. Next, one might think that the conditions that I distinguished
above that are supposed to give rise to certain epistemic intellectual obli-
gations are never jointly met. I argued that one has an epistemic intel-
lectual obligation, for instance, if one believes that p but also that one’s
belief that p does not fit one’s evidence. According to Charlotte Katzoff,
it is logically impossible that this antecedent holds. In her work, Katzoff
gives two different reasons to think so that, unfortunately, she fails to
distinguish. First, she says that believing that one has strong evidence
for p entails believing that p. 29 It follows from this that believing that one
has strong evidence for ~p entails believing that ~p, which, she takes it,
implies that one does not believe that p. 30 Thus, if one believes that one
has strong evidence against p, one cannot also believe that p. Second, she
claims that to believe that one has strong evidence against p entails not
believing that one believes that p. 31 Thus, her first point is that believing
29. See Katzoff (1996, 461). However, elsewhere she denies that this entails such a thing,
for there she says that believing that one has strong evidence against p does not entail
not believing that p. See Katzoff (2000, 93).
30. One might think this move of Katzoff is problematic. It seems implausible that believ-
ing that p logically excludes believing that ~p, and one might think there is some reason
to doubt that believing that p psychologically excludes believing that ~p. However, my
criticisms in what follows do not depend on these worries.
31. See Katzoff (2000, 92–93).
112
113
that one has strong evidence against p excludes the first-order belief that
p, whereas her second point is that believing that one has strong evi-
dence against p excludes the second-order belief that one believes that
p. However, both reasons have the same consequence, namely that it is
impossible for the conditions upon which the kinds of epistemic intel-
lectual obligations that I distinguished are predicated to hold.
I agree that cases in which one believes both that p and that one’s
belief that p does not fit one’s evidence are rare. But I do not think that
they are impossible, nor that they are so rare that the scope of our epis-
temic intellectual obligations is unrealistically minimized.
First, Katzoff’s arguments are directed against the claim that we
have certain intellectual obligations if we believe that p and that we have
strong evidence against p. But it follows from my view that one might
also have an intellectual obligation if one believes that p and that one’s
evidence favors neither p nor ~p, for in such a case suspension of belief
rather than belief or disbelief will be the epistemically proper attitude.
Katzoff’s argument does not count against the claim that we have such
intellectual obligations.
Second, we should note that people sometimes have high epistemic
standards. More skeptical and epistemically stringent people might
believe that they should not believe anything unless they can exclude
deceptive scenarios, and simultaneously find themselves with all sorts of
beliefs that they simply cannot resist. Such scenarios demonstrate that
believing that one lacks good evidence for p does not occlude believing
that p or believing that one believes that p. People with such inclinations
will frequently find themselves in scenarios of doxastic discrepancy. 32
Third, I think that each of us sometimes finds herself in the kind of
situation Katzoff deems impossible. 33 To give just one example: some
people believe in God but also that, epistemically speaking, they should
not believe in God, because they take themselves to have strong evi-
dence against the existence of God, such as the tremendous amounts
of evil and suffering in the world. I am not talking merely about faith in
32. Th
us also Owens (2002, 382–387).
33. Th
us also Heil (1984, 62–65), Mele (1987, 114–116).
113
114
God, faith that God exists, or subconscious beliefs, but about conscious,
occurrent beliefs: such persons think that God exists, even though they
think their evidence counts in favor of disbelieving that God exists.
6. According to Trent Dougherty, assessing belief in terms of
whether or not one has met one’s intellectual obligations that influence
which belief one holds is always pragmatic or moral rather than epis-
temic. All epistemic responsibility or epistemic blame concerns a single
time-slice of a person’s doxastic life: there is no epistemic responsibility
or epistemic blame that can only be understood diachronically or lon-
gitudinally, that is, by considering the doxastic influence a person exer-
cised. He sketches a case in which Craig, a creationist, is presented with
good reason to think that what he believes is false, which gives rise to a
duty to further inquire. Here is what he then says about the case:
There is either something at stake in the matter or there is not.

If there is nothing at stake, then, if he is anything like us and has
many other pressing concerns, there’s nothing irresponsible in
not being overly scrupulous (indeed, it would be irrational). If
there is something at stake, then it pertains either to one’s own
interests or the interests of others. If it pertains to one’s own inter-
ests, then the irresponsibility at hand is easily explained in terms
of practical irrationality. 34
I have been assuming that in the cases at hand, it pertains to one’s own
interest, namely one’s own interest in believing the truth and not believ-
ing any falsehoods. And any intellectual obligation that we have in vir-
tue of that, it seems, is an epistemic obligation. Of course, as Dougherty
rightly points out in this quote, such an obligation may often or even
usually be trumped by other, say, moral obligations. However, that still
leaves us with a pro tanto epistemic obligation.
Dougherty replies to this line of reasoning that we cannot have
such an obligation, for there are way too many truths out there. It is,
34. Dougherty (2012b, 540). For a discussion on this argument, see Axtell (2011b; 2012),
Dougherty (2011; 2012a). In what follows, I confine myself to what I consider the main
problem with Dougherty’s argument.
114
115
therefore, inevitable that our practical rather than our epistemic con-
cerns determine which intellectual obligations we have. 35 He fails to
acknowledge, though, that the epistemic aim is to believe the truth and
avoid falsehoods. I argue that the second conjunct gives rise to a fairly
limited range of epistemic intellectual obligations. One may object that
the fact that my account of epistemic intellectual obligations is cashed
out in terms of avoiding falsehood rather than acquiring true belief was
motivated by feasibility considerations and that these are practical con-
siderations. However, the feasibility considerations apply to all human
beings in virtually every circumstance and limit the very scope of obli-
gations. Compare it to moral obligations. We do not seem to have an
obligation to bring about every potential moral good, such as someone’s
happiness, anywhere on earth at any time, for that is utterly unfeasible.
But then practical considerations determine the scope of our moral
obligations, and, similarly, for prudential and other obligations. Hence,
if this is a problem for epistemic obligations, it is not a problem that is
unique to epistemic obligations, but a problem that applies to all obliga-
tions, moral and prudential obligations included.
One may object with Dougherty that an action’s having the truth
goal does not make it the case that one has an epistemic obligation to
perform that action, in the same way as an action that is done for the
sake of good cooking does not make it the case that one has a culinary
obligation to perform that action. 36 In fact, one might think, if hav-
ing the truth goal results in an epistemic obligation, then there would
be an infinite number of different kinds of obligations in addition to
moral and prudential obligations: culinary obligations, sport fishing
obligations, fiscal obligations, marital obligations, and so on. I reply
that in opposition to, say, the culinary realm, the epistemic realm,
like the moral realm, is often acknowledged as a sui generis domain
of normativity. If someone holds beliefs that have insufficient eviden-
tial support and if she avoids any evidence to the contrary, we blame
her for holding those beliefs as a result of performing those actions.
35. See Dougherty (2012b, 542). See also Dougherty (2014). For a similar position, see
Conee and Feldman (2004, 190).
36. S ee Dougherty (2011, 625–626; 2012a, 281–2 82).
115
116
We do so, even if there is nothing morally wrong with those beliefs;

we take it that people should aim at true rather than false beliefs and
that they should not avoid evidence to the contrary, since that vio-
lates an obligation with regard to the Jamesian goal. There is simply
no such thing with regard to, say, cooking: if people do not cook well,
we either do not blame them at all, or we blame them morally, since
they did not put their best effort into it while it was their responsibil-
ity to do so.
The account of epistemic intellectual obligations that I have
sketched, then, does not have the implication that our practical con-
cerns or moral obligations will always determine whether or not we have
duties to, say, further inquire. It leaves plenty of room for epistemic intel-
lectual obligations that are at least pro tanto obligations.
I conclude that our intellectual obligations are many and varied.
Some are contingent upon certain voluntary commitments, whereas
others are non-contingent. Some are moral, whereas others are pru-
dential or epistemic. Some are objective, whereas others are subjec-
tive. This is important, for if, as I argue in section 3.6, responsible
and blameworthy belief are to be cashed out in terms of whether or
not we have met our intellectual obligations, then the fact that there
is a garden variety of intellectual obligations might both explain and
justify the widespread practice of holding each other doxastically
responsible.
In this section, I have provided what I consider to be a frame-
work for thinking about which intellectual obligations we have. It
seems to me that there are many more details to be filled out. For
instance, are there different pro tanto, say, epistemic intellectual
obligations that may conf lict with each other? If so, what deter-
mines what one’s all-t hings-c onsidered epistemic intellectual obli-
gation is? And does one have an intellectual obligation to φ if ~φ-i ng
does not lead to beliefs that are objectively or subjectively bad, but
merely objectively or subjectively worse? These are difficult ques-
tions that a full-blown account of intellectual obligations that fills
in all the details needs to answer. Here, I have confined myself to
the framework that we need to develop an account of what it is to
believe responsibly.
116
117
3.6 INTELLECTUAL OBLIGATIONS

AND RESPONSIBLE BELIEF
I have argued that adult humans are subject to intellectual obligations,

obligations concerning factors that influence what we believe. I sug-
gested that our doxastic responsibility is to be explained in terms of such
influence. But precisely how is doxastic responsibility related to doxastic
influence? The aim of this section is to answer that question.
Some terminology might be helpful here. As William Alston rightly
notices, there are two ways in which one can be responsible for φ-ing. 37
First, one can be what I call originally responsible for φ-ing, namely if
one has some kind of control over φ-i ng. Second, one can be derivatively
responsible for φ-i ng if one lacks control over φ-i ng, but nevertheless has
influence on φ-i ng. These are not two kinds of responsibility, but two ways
of being responsible. Applying this distinction to my account of doxastic
influence, we can say that I have argued that we are originally responsible
for (not) performing certain belief-i nfluencing actions, such as evidence
gathering, and derivatively responsible for our beliefs.
The distinction between original and derivative responsibility is not
unique to the realm of doxastic responsibility. Let me give two examples
to illustrate why this is not the case.
First, imagine that I work in a hospital and that it is my task to fill
the oxygen bottles that are used in the ambulance. I have an obligation
to do so. Since I have control over whether or not I meet that obligation,
I am originally responsible for whether or not I do so. However, out of
laziness I fail to do so. One day, the ambulance attendants arrive upon a
scene in which someone desperately needs oxygen. Unfortunately, the
bottle is now empty, so that they cannot save him and he dies from a
lack of oxygen. In this case, I have no control over whether or not the
victim is saved, for I cannot intentionally set out to save him. I do not
37. See Alston (1989d, 137–140), Rosen (2004, 298–299). Alston is often misunderstood
on this point; he is thought to argue against the idea that doxastic responsibility can be
explained in terms of influence. See, for instance, Kim (1994, 282–2 84), Chuard and
Southwood (2009, 600). Alston’s point, however, is merely that epistemic justification
should not be construed in terms of influence.
117
118
even know that he exists or that he is a victim. It is clear, therefore, that

I have no obligation to save him. Nevertheless, it seems that I am blame-
worthy for his not being saved, because I had influence on that. For had
I filled the oxygen bottles, as I should, he would not have died. Thus, I am
not originally, but nevertheless derivatively, blameworthy for his death.
Second, in the 1950s molecular biologists Francis Crick and James
D. Watson, as well as biophysicist Rosalind Franklin, were working on
discovering the molecular structure of DNA. They did not intentionally
discover that its structure is that of a double helix, but they did inten-
tionally perform experiments that were meant to uncover its molecular
structure. Hence, they were originally responsible for performing these
experiments and derivatively responsible—in fact, given their hard
work, praiseworthy—for their discovery that DNA has the molecular
structure of a double helix. 38
The question under consideration, then, is this: precisely how does
our original responsibility for whether or not we meet our intellectual
obligations relate to our derivative responsibility for the ensuing beliefs
and, more specifically, to responsible belief? A first proposal is that they
are connected in the following straightforward way:
(3) S responsibly believes that p iff (i) S believes that p, and (ii) S
has not violated any intellectual obligations such that if S had
met those intellectual obligations, S would not have believed
that p. 39
This analysis, however, will not do. For (ii) is not a necessary condition for
responsible belief. Imagine that Julia violates an intellectual obligation
38. M any philosophers fail to take into account the distinction between original responsi-
bility, which requires intentional control, and derivative responsibility, which requires
merely influence. For instance, according to Peter Railton, there is not much left of nor-
mative epistemology unless we acknowledge that we have some kind of control over our
beliefs; see Railton (1997, 55), and Plantinga endorses the principle that you are to be
blamed for ~φ-i ng only if you have an obligation to φ; see Plantinga (1990, 52).
39. A gain, I confine myself to responsible belief, assuming that something similar can be
said about disbelief, suspension of judgment, and, maybe, deep ignorance (a variety of
ignorance that I spell out in more detail in chapter 5).
118
119
to prepare for her biology exams. Instead, she reads the Guardian, with
an article on the Mona Lisa, and thereby comes to believe that the paint-
ing was stolen in 1911. That belief of hers is clearly held responsibly,
although she would not have had it if she had met her intellectual obliga-
tions. Several philosophers have noticed this problem and proposed the
following refinement:
(3′) S responsibly believes that p iff (i) S believes that p, and

(ii) S has not violated any intellectual obligations such that
if S had met those intellectual obligations, S’s belief-forming
habits or access to relevant adverse considerations would
have changed in such a way that S would not have believed
that p.
The idea here is that, even if Julia had met her intellectual obligation by
preparing for her biology exams, upon reading the Guardian she would
have formed the belief that the Mona Lisa was stolen in 1911, because
preparing for her biology exams would not have made a difference to her
belief-forming habits and evidence concerning whether the Mona Lisa
was stolen in 1911.
I think (3′) is closer to the truth, but that it still needs revision on at
least five points. The most important point, which concerns excuses, will
be suggested in section 3.7 and defended in the ensuing chapters. Here,
I will argue for the necessity of the other four revisions. The first is minor
and primarily terminological. As I noted in chapter 2, we sometimes act
responsibly despite violating an obligation, namely if that obligation is a
pro tanto obligation rather than an all-t hings-considered obligation. It
seems that one can equally believe responsibly despite having violated
pro tanto intellectual obligations in coming to hold that belief. Thus, the
intellectual obligations mentioned in (3′) should be understood as all-
things-considered obligations.
Second, imagine that Julia has a moral intellectual obligation not
to spy on her niece. However, she violates this obligation and thereby
comes to believe that her niece is having a love affair with her boss. It
seems that in this case, Julia is morally blameworthy for believing that
her niece is having a love affair with her boss. But it is not the case that
119
120
if she had met her obligations, she would have had access to adverse
considerations, considerations that count against her niece having
an affair with her boss. Rather, she would simply have lacked any evi-
dence concerning whether her niece is having a love affair with her boss.
Also, I argued in §3.2 that doxastic mechanisms are among the belief-
influencing factors that we sometimes control. But they do not seem to
be belief-forming habits or pieces of evidence. I will, therefore, change
“S’s belief-forming habits or access to relevant adverse considerations”
into “certain belief-i nfluencing factors.”
Third, imagine that Julia violates an intellectual obligation to pre-
pare for her biology exam by reading the Guardian and thereby comes
to believe that she has violated an intellectual obligation. Clearly, if
she had met her intellectual obligations, then her evidence would have
changed in such a way that, upon reading the Guardian, she would not
have believed that she has violated an intellectual obligation. In this
case, Julia does not satisfy the analysans of (3′), but her belief that she
has violated an intellectual obligation is clearly held responsibly. I will,
therefore, add the following clause to (3′): “and S’s belief that p is not
about those intellectual obligations or belief-influencing factors.” This
move does not seem unduly ad hoc to me. Beliefs about such things as
those obligations that are relevant in determining whether someone
believes responsibly or not are of a special kind. If we want to know
whether someone believes responsibly and consider what that person
would have believed if she had met all her intellectual obligations, then
beliefs about those obligations themselves are clearly irrelevant.
Let me stress that it does not follow that one cannot be blameworthy
for one’s beliefs about such things as intellectual obligations. All that
follows from what I have said so far is that, if one is blameworthy for
a belief about an intellectual obligation, then that will be in virtue of
one’s violation of another intellectual obligation. Thus, Julia may well be
blameworthy for believing that she has violated an intellectual obliga-
tion to prepare for her exams. However, if so, that will not be because she
has violated her obligation to prepare for her exams. It will be because
she has violated some other obligation, an obligation that makes a dif-
ference to what Julia believes about whether she has met her intellectual
obligation to prepare for her biology exam.
120
121
Fourth, I think it is incorrect to use the plural “intellectual obliga-

tions” in (ii). Here is why. Imagine that I believe that p, that q, and that p
and q are incompatible. I therefore have an epistemic obligation to gather
further evidence about these propositions. I violate that obligation. If
I had met that obligation, I would have believed that p and abandoned
my belief that q. I now believe that p on the basis of wishful thinking—I
want p to be true—a nd no longer believe that q. Thus, this is a scenario
in which I do not responsibly believe that p, even though I would also
have believed that p if I had met my all-t hings-considered obligation to
gather further evidence on p, q, and the compatibility of p and q. The
reason is that there is some all-t hings-considered obligation—namely
the obligation to remove or weaken my tendency to believe things on
the basis of wishful thinking—t hat I have violated such that if I had not
violated it, I would not have believed that p. Thus, the antecedent of (ii)
should be thought of as a scenario in which I have not violated any all-
things-considered intellectual obligation that is relevantly related to my
belief. Someone who believes that p as a result of culpable wishful think-
ing clearly does not responsibly believe that p, even if she would have
believed that p if she had met all her intellectual obligations.
If I am correct about these four points, then, for now, we end up with
the following account of responsible belief:
(3″) S responsibly believes that p iff (i) S believes that p, (ii) S

has not violated any all-t hings-considered intellectual obli-
gation such that if S had met that obligation, certain belief-
influencing factors would have changed in such a way that
S would not have believed that p, and (iii) S’s belief that p
is not about that intellectual obligation or those belief-
influencing factors.
Let me close this section by discussing two other revisions that one
might think (3″) needs.
First, one might think that (3″) should say something about one’s
motivation. One can meet one’s intellectual obligations for the wrong
reason. For instance, one can meet one’s epistemic intellectual obliga-
tions not out of a desire for truth, but in order to impress other people,
121
122
or one can meet one’s moral intellectual obligations, not out of respect
for another person’s rights, but out of fear of other people finding out.
Imagine, for instance, that one meets some epistemic intellectual obliga-
tion by gathering further evidence on a proposition p that one deems
to be irrational. As a result of that, one abandons one’s belief that p and
acquires the belief that q. Imagine, however, that the reason one gath-
ered further evidence was not that one thought it is epistemically bad to
have an irrational belief, but that one knows that a philosophically ori-
ented girl one is in love with is very good at detecting irrational beliefs
and thoroughly dislikes the people holding them. So, one meets one’s
epistemic intellectual obligation for a purely pragmatic reason. There
seems clearly something wrong here, but my account tells us that in this
case one believes responsibly.
I agree that there is something wrong in such cases (and not only
with the girl and one’s being in love with the girl), but it seems to me
that that is compatible with my account. For there seems nothing blame-
worthy about this person’s belief that q. I am not sure whether there is
such a thing as purely epistemic blameworthiness, but if there is, it seems
that there is even nothing purely epistemically blameworthy about
this person’s belief that q. What is wrong in cases like these is that the
person in question lacks an appropriate desire for truth and performs a
belief-influencing action for the wrong reason. In the absence of an excuse,
it seems that he is blameworthy for these things. That is perfectly com-
patible with saying that he responsibly believes that q.
Second, one might think that the class of responsible beliefs is too
narrow on this account. For one might think that it should also include
those beliefs that, roughly, one would not have had if one had met one’s
obligations, but about which there is nothing undesirable. The idea behind
this view of responsible belief is, of course, that one is blameworthy
for φ-ing only if φ-i ng is objectively bad or undesirable.40 If I perform an
action that I believe will harm you but thereby bring about your happi-
ness, I am blameworthy for intending to harm you, but not for bring-
ing about your happiness, because there is nothing bad about bringing
40. F
or this view, see Nottelmann (2007, 47–51), Smith (1983, 556).
122
123
about your happiness. Applying the point to the doxastic realm, Nikolaj
Nottelmann says that we are epistemically blameworthy for some belief
only if that belief is epistemically undesirable in the sense that it is objec-
tively or subjectively epistemically bad.41
In response, let me first point out that on my account blamewor-
thiness and badness usually come together. Imagine, for instance, that
I believe the proposition p that the Mona Lisa was stolen in 1911, but
also that I have sufficient evidence against p. Above, I argued that in
such a case of doxastic discrepancy, I have an epistemic obligation to,
say, gather further evidence on p. Imagine that I violate that obligation.
Then, if my meta-belief is true, my first-order belief will continue to have
the undesirable property of being irrational. And if my meta-belief is
false, my second-order belief will continue to have the undesirable prop-
erty of being false. Thus, many cases in which I violate an epistemic obli-
gation and thereby continue to hold some belief will be cases in which
that belief continues to have some epistemically undesirable property.
However, I do not think this is always the case. Imagine that I believe
the proposition p that the Mona Lisa was stolen in 1911. However, I also
come to believe that that belief is based upon insufficient evidence.
Nevertheless, I find myself continuing to believe that the Mona Lisa was
stolen in 1911. Hence, I find myself in a situation of doxastic discrepancy.
Imagine also that I truly believe that my meta-belief is true, rational, war-
ranted, and so forth. Thus, there is nothing objectively or subjectively
bad or undesirable about that meta-belief. Imagine further that I am lazy
and fail to gather further evidence on p. And assume that if I had met my
epistemic obligation, I would have come to believe that my belief that p is
perfectly rational and consequently abandoned my meta-belief. Imagine,
for instance, that if I had met my obligation, I would have found evidence
for p that I have good reason to think I possessed all along but had forgot-
ten about. I think it is intuitively clear that I am blameworthy for holding
that meta-belief if I violate my obligation to gather further evidence on
p. But it is not a case in which there is anything subjectively or objec-
tively bad about my meta-belief at the time at which I should meet my
41. S ee Nottelmann (2007, 50–51).
123
124
obligation. Nevertheless, it is such that, if I do meet my obligation, then,

in retrospect, I believe that that meta-belief of mine was false. Thus, my
meta-belief would have been subjectively bad in retrospect if I had met
my obligation. Maybe it is this dissimilarity which explains why I am
blameworthy for my meta-belief if I fail to meet my intellectual obliga-
tion, but blameless for accidentally bringing about someone’s happiness
if I intend to harm him.
In the ensuing chapters, I will sometimes say that a belief that p
issues from one’s meeting or violating one’s intellectual obligation. Let
me stress that I use the verb “issue” in this context as a term of art.
A belief issued from one’s meeting an intellectual obligation if meeting
one’s intellectual obligation resulted in one’s coming to believe that p, or
if one would no longer have believed that p if one had not met one’s intel-
lectual obligation. And a belief issues from the violation of one’s intellec-
tual obligation if violating that intellectual obligation resulted in one’s
coming to believe that p, or if one would no longer have believed that p if
one had not violated one’s intellectual obligation.
3.7 D
OX ASTIC EXCUSES: FORCE,
IGNORANCE, AND LUCK
So far, every time I mentioned excuses I assumed that the reader has
some grasp of this notion. Now the time has come to be more precise
and spell out exactly what it is to be excused. I do not mean to be ask-
ing when someone is verbally excused by herself or someone else. For
it seems clear that one can be excused for something, even if no one
verbally excuses one for it. For instance, I may be completely blameless
for having accidentally killed someone, even though nobody, including
myself, is aware of that. Thus, I take excuses not to be speech acts per-
formed by a person in defense of her or someone else’s φ-ing, as some
philosophers do,42 but those states of affairs the actualization of which
renders one blameless.
42. See Austin (1979, 176), Brandt (1969, 337), Zimmerman (1988, 64–69).
124
125
The Latin expression ex causa, from which the word “excuse” is

derived, means “out of ” or “away from” an “accusation.” As the ety-
mology rightly suggests, then, we are excused for something only if we
are blameless for it. Blamelessness, however, is not sufficient for being
excused. I am blameless for drinking two cups of coffee a day, but I am
not excused for that.
What is it that we are excused for when we are excused for some-
thing?43 One might think that we are excused for the actualization of
some state of affairs 44 if and only if we are blameless for it and it is objec-
tively bad or undesirable. For instance, a doctor is excused for giving his
patient a deadly medicine if he believed that giving that medicine would
cure his patient. Here, the excuse for his objectively bad action is his
blameless ignorance of the consequences of his giving that medicine.
This idea seems to me misguided. First, I am blameless for not solv-
ing the Middle East problem and for not curing all cancer patients in my
country, but I am not excused for the actualization of these bad states of
affairs. That is because I have no obligation to prevent these bad states
of affairs from being actualized. And I do not have such an obligation
because most or all people, including me, lack the power to do so.45
Second, we can be excused for violating subjective obligations, even
if there is nothing objectively bad about performing the act in question.
Thus, if a doctor is physically forced to give his patient a medicine that he
falsely believes will kill his patient, it seems that he is excused. Objective
badness of the relevant state of affairs, then, is not necessary for being
excused for its actualization, nor is it, in conjunction with blameless-
ness, sufficient for it.
But then, why have some philosophers thought, or at least intimated,
that to be excused is to be excused for some objective badness? I think
43. A ustin (1979, 175–177) and Brandt (1969, 337) are rather ambiguous on what it is that
one is excused for if one is excused for something. The view that I defend below is simi-
lar to that of Fields (1991, 11).
44. I talk about states of affairs rather than actions, so as to include beliefs.
45. More specifically, I fail to satisfy the following necessary condition that I defended in
section 2.3 above: S at t has an obligation to φ only if (i) at t people normally have con-
trol over φ-i ng, or (ii) S normally has control over φ-i ng, or (iii) S at t has control over
φ-i ng. This is proposition (7″) that states the relation between control and obligations.
125
126
that is because what we are really excused for is the violation of obligations.
And, as we saw above, stereotypical obligations to perform some action
are obligations to avoid some kind of badness. What one is excused for
is easily confused with what usually accompanies it. More specifically,
as I argued in c hapter 2 (§2.3), we are excused for all-things-considered
rather than pro tanto obligations. For example, if I violate a pro tanto
obligation to prepare for my biology exam by meeting my moral obli-
gation to attend the funeral of my parents, who died in a car accident,
then it seems that I am justified in not preparing for my exam rather than
excused for not preparing for my exam. I agree that ordinary language
may be vague on this point, but it seems that most philosophers pre-
fer to describe such a case as one in which I am justified in violating an
obligation rather than excused for it. If this is correct, excuses should be
understood as follows:
(4) S is excused for the actualization of some state of affairs Σ

iff (i) S fails to meet an all-t hings-considered obligation that
is relevantly related to the actualization of Σ, and (ii) S is
blameless for the actualization of Σ.46
Let me say one more word about all-t hings-considered obligations.

Some philosophers argue that obligations of different sorts, such as
moral obligations on the one hand and epistemic obligations on the
other, are incommensurable. On their view, there is no such thing as
one’s obligation both-moral-and-epistemic-things-considered.47 I will not
take a stance on this controversial issue. Rather, I would like to explain
how my account of doxastic excuses can accommodate this view. We
should notice that on the incommensurability view, there could still be
conflicting intellectual obligations that are all epistemic, all moral, or
all prudential. It follows that the distinction between pro tanto and all-
things-considered would still hold. Only, it would be sort-relative. Thus,
46. I n section 3.6, I have spelled out how this relevant relation should be understood in the
case of belief.
47. Thus, for instance, Conee and Feldman (2004, 191–194). For the opposite view, see
Chisholm (1991, 127) and especially Booth (2012).
126
127
all pro tanto and all-things-considered obligations would be epistemic, or

moral, or prudential. My account can easily adopt this position by under-
standing “excused,” “all-t hings-considered obligation,” and “blameless”
in (4) as, say, “epistemically excused,” “all-epistemic-t hings-considered
obligation,” and “epistemically blameless.” For economical reasons, I will
henceforth stick with the phrasing of doxastic excuses as found in (4).
I would like to stress that particular excuses are sufficient for blame-
lessness, not necessary. For it seems that different excuses can obtain
simultaneously. Imagine that I believe that I will make you happy by
giving you a good single malt whisky for your birthday, say, a Talisker
57° North. However, unbeknownst to me, you have some illness, one of
the side effects of which is that you are seriously allergic to whisky. My
ignorance of your illness excuses me for giving you the whisky, but it is
not necessary for blamelessness. For even if I had believed that you had
that illness, but had been inculpably ignorant of the fact that that illness
has this particular side effect, I might have been blameless for offering
you this whisky. In fact, we can even imagine that someone forces me
to give you this whisky, so that three excuses obtain—t wo instances of
ignorance and an instance of force.
For the account of responsible belief that I formulate and defend
in this book, it is important to make two distinctions when it comes
to excuses. The first is the distinction between full and partial excuses.
When one is fully excused, one is not blameworthy at all. One does not
deserve any blame, at least not for the actualization of the state of affairs
for which one is excused. If I give my daughter a piece of chocolate that,
unbeknownst to me, was poisoned by a maniac who happened to choose
my house for his malicious action, and I have no indication whatsoever
to think that the chocolate is poisoned, then, it seems, I am not blame-
worthy at all for giving her that piece of chocolate. However, there are
also partial excuses. Imagine that I heard on the news that some maniac
is poisoning people’s chocolate bars in my neighborhood. I notice that
my chocolate bar is opened. But then, I know, we often leave opened
chocolate bars in the desk and finish them later. I nonetheless decide to
give it to my daughter. Imagine that it is poisoned. It seems that in that
case I am blameworthy for giving it to my daughter; I act recklessly in
doing so and violate an objective obligation not to give it to her. Still, it
127
128
seems, I am not as blameworthy as I would have been if I had known that
it was poisoned. Thus, my ignorance that it is poisoned makes me less
blameworthy than I would have been if I had not been ignorant, but I am
still blameworthy to some degree. The following two analyses capture
this difference between full and partial excuses:
(5) State of affairs Σ fully excuses S for the actualization of some

other state of affairs Σ* iff (i) S fails to meet an all-things-
considered obligation to prevent the actualization of Σ* or to
(not) do something which would have prevented the actual-
ization of Σ*, and (ii) due to the actualization of Σ, S is blame-
less for the actualization of Σ*.
(6) State of affairs Σ partially excuses S for the actualization
of some other state of affairs Σ* iff (i) S fails to meet an all-
things-considered obligation to prevent the actualization of
Σ* or to (not) do something which would have prevented the
actualization of Σ*, (ii) S is blameworthy for the actualization
of Σ*, but (iii) due to the actualization of Σ, S is less blame-
worthy for the actualization of Σ* than if Σ had not obtained.
The second distinction is the distinction between narrow and broad

excuses: whereas narrow excuses take away merely blameworthiness,
broad excuses remove all responsibility.48 For instance, imagine that
Anna, while paying a visit to the Metropolitan, as a result of a sudden
epileptic seizure hits another visitor and thereby brings about serious
suffering on his part. We excuse her for that because we will not hold
her responsible for it at all—t he action was not up to her. Compare her
situation to that of her sister, Mary. Mary has good reason to think that
an armed burglar has forcefully entered the building. When he turns
his back toward her, Mary knocks him to the ground. However, the
man is no burglar and is just playing a role in a security-check training.
48. For the same distinction, see Botterell (2009, 182–183), Strawson (1974, 8–9, 16–
17), Zimmerman (1988, 66). One or the other variety is overlooked by Austin (1979,
176), Gardner (2009, 338), Houlgate (1968, 109), Van Woudenberg (2009, 379),
Zimmerman (1996, 94; 2008, xiii).
128
129
By hitting the supposed burglar, Mary violates an objective all-t hings-

considered obligation not to hit the man. It is clear, however, that if
Mary’s belief is held responsibly, she is excused and, hence, blameless
for hitting the man. In fact, we consider her praiseworthy for acting as
she does. Whereas Mary’s excuse blocks only blameworthiness, Anna’s
excuse blocks all responsibility. I will call such excuses respectively nar-
row and broad excuses and stipulatively define them as follows:
(7) S is narrowly excused for believing that p iff (i) S has violated
an intellectual obligation that is relevantly related to S’s
believing that p and (ii) S is the proper object of positive or
neutral, but not of negative appraisal for believing that p.
(8) S is broadly excused for believing that p iff (i) S has violated an
intellectual obligation that is relevantly related to S’s believ-
ing that p and (ii) S is not the proper object of any appraisal
for believing that p.
If my aim in this book were to give an account of blameless belief, then

the distinction between narrow and broad doxastic excuses would not
matter, for both narrow and broad excuses entail blamelessness. And if,
contrary to what I argued in chapter 1, responsible belief were praisewor-
thy belief, they would have limited value, for then even being narrowly
doxastically excused would not guarantee responsible belief. However,
as I argued in c hapter 1, believing responsibly entails being the proper
object of positive or neutral appraisal. This means that if one is narrowly
excused for some belief, one still believes responsibly. Thus, if some per-
son has violated certain intellectual obligations, that person can still
responsibly hold the beliefs issuing from the violation of those obliga-
tions, namely if she is narrowly excused for those beliefs. In the ensuing
chapters, I argue that this is indeed sometimes the case.
Some authors have discussed the issue of when one is excused for
holding an epistemically undesirable (such as an unreasonable) belief or
when one is excused for holding a false belief.49 If what I have argued in
49. See, respectively, Nottelmann (2007, 208–217) and Van Woudenberg (2009).
129
130
this chapter is correct, then it follows that if one holds an irrational or

false belief, one is only sometimes excused for that in the full-blooded
sense (being blameless), because one only sometimes violated an obli-
gation in coming to hold the irrational or false belief in question. Since
the topic of this book is responsible belief, that is, being praiseworthy
or neutrally appraisable for holding a belief, in the following chapters
our focus is on being excused for holding some belief. One can still be
excused for holding a belief even if it is true, warranted, rational, and so
forth, because, one would not have held that belief if one had met one’s
intellectual obligations (but one is excused for not doing so).
There are many kinds of excuses. We appeal to mistake, aberration,
absence of mind, accident, intoxication, compulsion, insanity, or duress.
We say we did things unwittingly, impulsively, not deliberately, auto-
matically, involuntarily, inadvertently, inattentively, carelessly, or not
on purpose. We could not help, did not mean to, did not realize, or did
not intend. We were ignorant, inconsiderate, under constraint, forced,
or not in control. In the following chapters, we shall see that there is
good reason to think that (virtually) all doxastic excuses, including
luck, are reducible to one or more of the following two excuses: force
and ignorance.
3.8 C ONCLUSION
It is time to take stock. I have argued that there are three factors that
influence what we believe and that are often under our control: dox-
astic mechanisms, cognitive situatedness, and intellectual virtues and
vices. It turned out that we have certain obligations, both contingent
and non-contingent, with regard to these belief-influencing factors.
This is because to not perform them is sometimes to actualize or main-
tain a state of affairs that is morally, prudentially, or epistemically bad
from a doxastic point of view. I have argued that, roughly, one believes
responsibly if and only if in coming to hold that belief one did not
violate any intellectual obligations and one’s belief is not about those
intellectual obligations. However, we have also seen that there is some
130
131
reason to think that the account is not yet complete. For it may be that
we sometimes believe responsibly despite having violated an all-t hings-
considered intellectual obligation in coming to hold that belief. Maybe
force, ignorance, luck, or several of these phenomena sometimes provide
a narrow excuse and thereby leave room for responsible belief. In the fol-
lowing chapters I argue that this is indeed the case and that each of these
excuses teaches us something important that an analysis of responsible
belief should take into account.
131
132
[ 4]
RESPONSIBLE BELIEF ENTAILS
THE ABILITY TO BELIEVE
OTHERWISE
4.1 I NTRODUCTION
So far, I have argued that, roughly, one holds a belief responsibly if and
only if in acquiring and maintaining that belief either one has not vio-
lated any intellectual obligations or one is narrowly excused for holding
that belief. In c hapter 3, I spelled out what kinds of intellectual obliga-
tions we have and precisely how they relate to responsible belief. The
question that remains to be answered, then, is when we are narrowly
excused for some belief, that is, when we are excused in such a way
that we still believe responsibly, despite having violated certain intel-
lectual obligations. If what I argued in c hapter 3 is correct, then doxas-
tic responsibility derives from our responsibility for belief-influencing
actions and omissions.1 It is not surprising, then, that my discussion of
doxastic excuses in this and the ensuing chapters manifests significant
similarities to discussions of excuses in ethics and law, disciplines that
are both primarily concerned with actions and omissions. However,
actual analyses of excuses in ethics and law are scant and, as we shall
see, at several junctures our focus on belief requires a particular take on
the matter.
1. Henceforth, I will discuss mostly actions, assuming that mutatis mutandis the same can
be said about omissions.
132
133
R esponsible B elief E ntails the A bility to B elieve O therwise
Ever since Aristotle, it has been acknowledged that force can excuse.2
But there are few accounts of precisely when force counts as an excuse
for an action. It is even harder to find an analysis of force as a doxastic
excuse. This is not to deny that many epistemologists acknowledge that
virtually all of our beliefs are in some sense forced upon us by our evi-
dence. 3 It is widely acknowledged, though, that that kind of force as such
does not excuse. The aim of this chapter is to give a precise statement
of when force counts as a narrow doxastic excuse, that is, as an excuse
that does not remove all responsibility but leaves room for responsible
belief. And let me nail my colors to the mast by saying that I shall argue
that the crucial lesson to be learned from force as a doxastic excuse is
that, in a sense to be explained, responsible belief entails the ability to
believe otherwise, since one is not responsible for beliefs that are strictly
unavoidable.
The chapter’s structure can be outlined as follows. Since there
is a wide variety of forces, I first specify the kind of force that is rel-
evant for our purposes, that is, the kind of force that might excuse.
I argue that such force is best construed in terms of one’s ability to
form and act on certain intentions, and that, thus construed, force
comes in degrees (§4.2). Next, I show how on this view force as a
narrow doxastic excuse is to be construed. If my account of force as
a doxastic excuse is correct, then the following thesis is true: dox-
astic responsibility in general and responsible belief in particular
entails the ability to believe otherwise (§4.3). After that, I consider
in more detail whether or not force also provides a doxastic excuse
if one is blameworthy for being subject to such a force that results in
the inability to believe otherwise (§4.4). In sections 4.5–4 .7, I reply
to three objections against this thesis. 4 First, one might think that
what suffices for doxastic responsibility is control over or inf lu-
ence on certain desirable or undesirable properties of beliefs (§4.5).
Second, it may be objected that the thesis is contradicted by our
2. See Aristotle (2003, 117–129; NE III.i). See also Bennett (1980, 15), Brandt (1959,
472–473), Fischer and Ravizza (1998, 13–14), Goldman (1970, 208).
3. E .g., Clarke (1986, 40, 48), Heil (1983, 357).
4. M
y responses to these objections are based on Peels (2013b).
133
134
intuitions in doxastic analogues of Frankfurt-s tyle scenarios, that

is, circumstances in which one seems responsible for having a belief
despite being unable not to have that belief. I argue that this objec-
tion does not hold water, but suggest that one might nevertheless
reasonably remain a compatibilist about actions, omissions, and
foreseeable consequences, since doxastic responsibility differs from
responsibility for these other phenomena in that we do not foresee
which beliefs we will hold as a result of the violation of our intel-
lectual obligations (§4.6). Third, one might draw an analogy with
the asymmetry thesis in ethics by arguing that blameworthy belief
requires the ability to believe otherwise, whereas responsible belief
does not (§4.7).
4.2 W HAT IS FORCE?
There are many different kinds of force, such as the force of rhetoric, the
British naval force, the force of gravity, and, most importantly of course,
the Force that can be with you. Naturally, not all of these are relevant
for our purposes here. Relevant are those kinds of force that purportedly
excuse. The following scenarios seem to involve kinds of force that are
often considered as valid excuses in certain circumstances:
(A) I have irresistible epileptic seizures as a result of having taken

pentylenetetrazole.
(B) I am unable to go outside, since I have been locked up in a
prison.
(C) I suddenly stop talking as a result of an unexpected explosion
in the background.
In these cases, it seems that I am forced (not) to do something, because

I lack the relevant kind of control over it. And as soon as I am in the rel-
evant situation—having used pentylenetetrazole, being in prison, sud-
denly hearing an explosion—I also lack influence over whether I perform
or fail to perform the action in question. In this chapter, I will stipu-
latively use the word “force” as an umbrella term to capture all these
134
135
varieties of excuses. Given what I just said about control and influence,
it can be defined as follows:
(1) S is forced to φ iff (i) S φ-s, and (ii) what S wills cannot make
a difference as to whether or not S φ-s.
The core idea encapsulated in (1) is that one is forced to do something

if what one intends, wills, decides, or chooses cannot make a differ-
ence to what one does or fails to do. 5 Paradigm cases of being forced are
instances of compulsion, situations in which some other person makes it
impossible for one to do otherwise than one does. But I do not think that
it is unnatural to use the word “force” in a broader sense. We can prop-
erly say that our conscience forces us to ask questions when a person’s
life is at stake, that our eyes are forced to strain and focus continually,
and that old memories can be forced to the surface. In all these cases,
our decisions cannot make a difference to what happens and that seems
to explain why we use the word “force” or “forced” in these cases. It is
unsurprising, then, that philosophers especially have talked about our
beliefs being forced upon us, an idea that fits well with the thesis of dox-
astic involuntarism that I defended in chapter 2.6
As it stands, (1) makes sense of scenarios such as my having a brain
tumor, my being kidnapped, and my being put in chains. For in all these
cases my intentions cannot make any difference to certain actions or
omissions of mine. But one might think that there are other situations of
force, in which my intentions can make a difference to what I do, but in
which I may nevertheless be excused by being subject to a partial force.
Consider the following examples:
(D) A Guantánamo Bay prisoner, who believes that he should not

confess, fails to resist the force to confess exercised upon him
by waterboarding techniques.
5. F
or a similar view, see Alston (1989d, 123).
6. I n other contexts as well, philosophers have used the word “force” for anything to which
the agent’s will or intentions cannot make a difference. See, for instance, Wolf (1990,
14, 29).
135
136
(E) Jim has been lost in the desert. After two exhausting days,
he stumbles upon an oasis. He is so thirsty that he drinks the
water before verifying whether it is clean.
(F) Penelope, who is babysitting, loves burritos, but she realizes
that she has gained much weight recently and should eat no
more burritos. Before they leave, the owners of the house tell
her that she can have the burritos that are in the kitchen. She
succumbs and eats them.
These are all cases in which people are subject to some force, but the
degree to which they are seems to differ. To resist waterboarding is vir-
tually impossible, but some people can do so with extreme effort. Many
of us, it seems, could stop and check the water before drinking it, even
if we have been in the desert for two days, but to do so would require a
great effort of will. To not eat burritos when one loves them is difficult,
but with substantial effort one could do so. It seems natural to say that
in these cases, there is some force respectively to confess, drink, and eat,
but that that force is not completely irresistible. Here, one’s intentions
can make a difference: one is not subject to full force, but only to a par-
tial, even though strong, force.
I think this is worth noticing, for one might think that we are some-
times excused for violating intellectual obligations and for the ensuing
doxastic attitudes, despite the fact that we could have met the obliga-
tion as the result of an intention to do so and that if we had, we would
not have had the doxastic attitude in question. Here, we can think, for
example, of cases in which meeting one’s intellectual obligation is not
of great moral or epistemic importance and in which it is hard, but not
impossible, to meet one’s intellectual obligation. Imagine, for instance,
that I promised a friend of mine to read a paper of his, but that I fall ill
and remain in bed with a 103° F fever. Imagine that with extreme effort,
I could read the paper. Then, does not my having a fever excuse me for
violating my intellectual obligation to read the paper and for the ensuing
ignorance—false beliefs, or no true beliefs—of what my friend argues
in the paper? One might think, then, that a theory of responsible belief
should take into account not only full force, but also partial force, for
136
137
there are circumstances in which partial force seems to provide a dox-

astic excuse.
However, I think that we need not take on board partial force. For
if I can, although with extreme effort, read the paper, there are two
options. First, it is possible that the intellectual obligation is so strong
that I should put such extreme effort into it. Whether it is that strong
depends, of course, on the circumstances. Imagine, for instance, that my
friend is on death row, that he could be executed at any minute, and that
I know that his paper might contain information that I can use to get
him off death row. Then, it seems, extreme effort on my part is required
and if I fail to read the paper in that situation, then I am blameworthy for
not doing so. Hence, in situations of the first kind, partial force does not
excuse. Second, it may be that the intellectual obligation is not so strong
that I am required to put so much effort into it. But then I have merely
a pro tanto rather than an all-things-considered intellectual obligation
to read the paper, for then my intellectual obligation is outweighed or
trumped by the effort of will that it would take me to read the paper.
And, as I pointed out in chapter 2 (§2.3), we are only ever excused for
the violation of all-things-considered obligations. Hence, in scenarios of
the second kind I am blameless, but not excused, for I have not violated
an all-t hings-considered intellectual obligation. On both the first and
second interpretations of the scenario, then, it is one in which I am not
doxastically excused. And it seems that the case can easily be general-
ized: if one is subject to partial rather than full force, one is either blame-
worthy or blameless without being excused. This means that, although
there may be a natural sense of the word “force” in which force is some-
thing that comes in degrees, we do not need to revise (1). In consider-
ing force as a doxastic excuse, we can legitimately confine ourselves to
full force.
4.3 F
ORCE AS A DOX ASTIC EXCUSE
Now that we have a firmer grip on force, let us address the following
question: when is someone excused by force for holding some belief?
137
138
And, especially relevant for our purposes, when is someone excused by

force in such a way that that person still believes responsibly?
In c hapter 3, I argued that we often influence what we believe. Of
course, if what I argued in c hapter 2 is correct, whether or not we exer-
cise doxastic influence, we will end up being psychologically forced,
usually simply by the evidence we have, to believe some proposition or
other. But we can influence which proposition we end up being forced to
believe by making a difference to our doxastic mechanisms, cognitive
situatedness, or intellectual virtues and vices. What this suggests is that
force excuses one for some belief if it was impossible for one to perform
a belief-influencing action such that one had an all-things-considered
obligation to perform that action and if one had performed that action,
one would not have held that belief. Such force will count as a doxas-
tic excuse because (i) part of the badness of violating that obligation is
constituted by (the likelihood of) one’s acquiring or maintaining that
belief, and (ii) one is blameless for the violation of that obligation and for
the ensuing belief. For if there is nothing one could have done to avoid
having that belief, then it is hard to see how one could be blameworthy
for having that belief.7 It seems, then, that (virtually) all our beliefs are
synchronically forced, given the evidence we have and our psychological
constitution, but that many of them are not diachronically forced, given
our ability to influence what we believe.
I do not think that this view is exotic, for mutatis mutandis we
encounter the same phenomenon when it comes to non-doxastic conse-
quences of our actions. Most people experience an irresistible force to
breathe at least once every two minutes. There are a few exceptions: cer-
tain divers are known to have held their breath for more than nine min-
utes. But here is the point: some people could with serious effort learn
to hold their breath for longer than two minutes. Some of them could
learn to do so only with extreme effort. For others, certain people with
asthma, for instance, it is simply physically impossible to acquire that
skill. For still others, it would be very easy to learn how to do so. Those
who would succeed if they were to try sufficiently hard have influence
7. Thus also Steup (1988, 72).
138
139
on their holding their breath for at least two minutes. Something similar
can be said about many other actions, such as becoming prime minister
and making discoveries in astronomy. This is not to deny that there are
important differences between force to act and doxastic force. In trying
to resist the former, we normally intend to perform some specific action
or omission, whereas in attempting to avoid the latter we normally only
intend to change our evidence base or a belief-forming mechanism,
without knowing or having a belief about which specific doxastic atti-
tude will result from that. Such differences, however, do not affect the
truth of the point I want to make here: although all beliefs are forced at
the time at which we have them, we can nonetheless be responsible for
some of them in virtue of the fact that not all belief-i nfluencing actions
and omissions are forced.
Let us dig a bit deeper by returning to the distinction between
broad and narrow excuses that I made in c hapter 3. Remember that
broad excuses remove all responsibility (praise, blame, and neutral
appraisal), whereas narrow excuses remove only blameworthiness.
It is clear that force is sometimes a broad doxastic excuse. Imagine
that a professor assigns her students to read a paper on the French
Revolution, one she has written herself. Lindsay is one of her students.
Imagine that her professor has just finished a draft of the paper and that
in the paper she expresses certain propositions that Lindsay does not
yet believe, but would believe if and only if she were to read the paper.
However, due to a hurricane that turns Lindsay’s house upside-down
and destroys everything in it, including the paper, Lindsay is unable to
read the paper. Is she responsible for not holding certain beliefs that she
would have held if she had read the paper, as she should? It seems that
she is not: she is neither praiseworthy, nor blameworthy, nor neutrally
appraisable for that.
Is force sometimes a narrow rather than a broad doxastic excuse? That
is, is force sometimes an excuse that removes blameworthiness, but not
praiseworthiness or neutral appraisability for one’s belief? In order to see
whether it is, imagine the following slightly revised scenario. Lindsay’s
professor assigns to her students a few articles on the French Revolution.
Due to a tornado, Lindsay’s house is destroyed and she cannot read those
articles (I assume she cannot read them online or borrow them from her
139
140
friends). However, for several propositions expressed in those articles, it

is true that Lindsay has influence on whether or not she believes them,
for she could read certain books that she has on the French Revolution in
which those propositions are also expressed. Now, imagine that Lindsay
holds a particular belief that p about the French Revolution and that in
coming to hold this belief, she has not violated any obligation, except
for the obligation to read the articles, an obligation for the violation of
which, we assume, she is excused. Imagine also that she would abandon
that belief if she were to read one specific book on the French Revolution
that she has. It is clear that she is blameless for her belief. But is she still
responsible for it or is she not at all responsible for holding that belief?
I think that in such cases, our intuitions start to waver. The paradigm
case of φ-ing responsibly is acting responsibly and cases of action are cases
in which one exercises control. As I argued, all we have in the case of belief
is influence. And in c hapter 1 I argued that responsible belief is blameless
rather than praiseworthy belief. In that sense, it is a fairly weak notion; it
is quite easy to believe responsibly. If something similar applies here, we
can say that one believes responsibly if one did not violate an intellec-
tual obligation in coming to hold that belief or if one is excused for doing
so, and if one could have made a difference to one’s belief, whether one
is aware of that or not. It may be more plausible that one is responsible
for believing that p if one truly (tacitly) believes—or has a disposition to
truly believe—that one could perform certain belief-influencing actions
that are sufficiently likely to make a difference as to whether or not one
believes that p than if there are such belief-influencing actions but one
neither believes that nor has a disposition to believe that. As I said in the
Introduction to this book, there may well be paradigm cases of respon-
sible belief and boundary cases of responsible belief. It seems the cases
of responsible belief under consideration, if they are cases of responsible
belief at all, will fall into the category of boundary cases.
What if I believe that p and if neither now nor at some time in the
past there was an action or a series of actions that I could have intention-
ally performed, such that if I had done so, I would now have failed to
believe that p?
There seem to be such beliefs, for example, my belief that I exist, that
I existed yesterday, that there is a past, that 2 + 2 = 4, that there are other
140
141
minds, that modus ponens is logically valid, and that there is an exter-
nal world. How many of these kinds of beliefs we have depends on one’s
account of belief. If we also have dormant and—especially relevant for
our discussion here—tacit beliefs, as I defined them in chapter 1, then
it seems that we will have a large or infinite number of beliefs that we
cannot avoid having. For example, I believe that elephants are not iden-
tical to the number 1, nor to the number 2, and so on. Well, it seems that
if there is no way we could have avoided having them, then we are not
responsible for them. We simply have to put up with these beliefs.
One may worry that these are usually beliefs that enjoy a very
high degree of epistemic justification, some of them even the highest
degree possible, for instance, in the case of my belief that I exist. 8 So,
it follows from what I said that when we are epistemically at our best,
we form beliefs for which we are not responsible, and that might seem
counterintuitive. We should not forget, though, that the subject under
consideration is doxastic responsibility rather than whether or not we
meet certain epistemic norms or standards. As we saw in section 2.7,
we should clearly keep these two apart. That a belief was formed on a
good evidence base, that it is true, that it was formed reliably or even
infallibly, and that it meets all sorts of other epistemic desiderata is epis-
temically valuable, but it does not follow that we are responsible for it, in
the same way as the fact that we did the right thing does not mean that
we acted responsibly, since we might have been forced to perform that
action. I return to the exact relationship between epistemic justification
and responsible belief in the Appendix.
Another important class of beliefs that are unavoidable consists of
certain irrational beliefs that some people hold. Depending on the per-
son in question, one can think of such beliefs as that some conspiracy
theory is true, that one’s life has little value, that God exists, or that God
does not exist. Of course, many irrational beliefs are blameworthy, but
they seem blameworthy precisely because one could have avoided hav-
ing them. If there is literally nothing one could have done to avoid them,
then, it seems, one is not at all responsible for them. Diachronic doxastic
8. I thank Matthias Steup for raising this worry.
141
142
force, then, seems to remove all responsibility. Hence, in those cases in

which I believe that p because I violated an intellectual obligation in
coming to believe that p and meeting that obligation would have been
the only way to avoid believing that p, diachronic force counts as a broad
excuse: it removes all responsibility.
If I am right that diachronic force counts as a broad doxastic excuse,
then the following thesis is true: doxastic responsibility requires the abil-
ity to believe otherwise. By that I do not mean that we are responsible for
a belief only if we have or had control over that belief (that we could
have chosen not to hold that belief). Nor do I mean that in the very
same evidential circumstances, we could have failed to have the belief
in question. Rather, what I mean by this is that one’s belief is not dia-
chronically forced. Thus, one has the ability to believe otherwise if and
only if there was some belief-i nfluencing action or series of actions that
one could have performed such that if one had done so, one would not
have held the belief.9 Cases in which diachronic doxastic force excuses
one for having some belief invoke or activate our intuition that one is
not responsible for a belief that is completely beyond one’s influence.
Several philosophers, such as Stephen Hetherington and Mark
Leon, have claimed that responsibility for belief entails the ability to
believe otherwise.10 If what I have argued so far in this and the preced-
ing chapters is right, then that thesis should be spelled out as follows:
(2) Some cognitive subject S is responsible for believing that p

at some time t only if there is some action or series of actions
A that S did not perform but could have performed (or did
9. Th is means that to the extent that epistemic conservatism—see McCain (2008)—is a
thesis about responsible belief rather than epistemically justified belief, it needs a slight
revision to the effect that we believe responsibly only if our belief has not been defeated
and if there is something we could have done such that if we had done that, we would
no longer hold that belief. Fortunately, such a revision would not affect the argument
provided in section 1.4 that partly relies on epistemic conservatism. For the point was
that epistemic conservatism squares much better with the view that responsible belief
is blameless belief than with the view that responsible belief is praiseworthy belief. This
is still true on the revised version of epistemic conservatism.
10. See Hetherington (2002, 401), Leon (2002, 424–430). Haksar (1964, 324) has claimed
something similar for moral beliefs.
142
143
perform, but could have failed to perform) at some time

prior to t such that if S had performed (not performed) A, S
would not have believed that p at t.
Above, I have sketched a scenario in which our intuitions seem to sup-

port this thesis. Below, I defend the thesis against three objections.
Before doing so, let us return to our analysis of responsible belief.
Taking into account both the fact that force can provide a narrow dox-
astic excuse and the fact that responsible belief requires the ability to
believe otherwise, our analysis now looks as follows:
(3) S responsibly believes that p iff (i) S believes that p, (ii) there
is some belief-influencing action or series of actions A that S
could have performed such that if S had performed A, S would
not have believed that p, and (iii) S has not violated any all-
things-considered intellectual obligation such that if S had
met that obligation, certain belief-influencing factors would
have changed in such a way that S would not have believed
that p, and (iv) S’s belief that p is not about that intellectual
obligation or those belief-influencing factors, or (v) S has vio-
lated such an obligation, but is excused for that by force.11
Whereas clause (ii) guarantees that one’s belief that p is not completely
beyond one’s influence, but that one is responsible for believing that p,
clauses (iii)–(v) guarantee that one is blameless for believing that p,
either because one has not violated an intellectual obligation in believ-
ing that p or because one is excused for that.
4.4 B L AMEWORTHY FORCE
Before we consider the objections against the thesis that responsible

belief implies the ability to believe otherwise, there is one important
e analysans of (3) should be read as: (i) ∧ (ii) ∧ [[(iii) ∧ (iv)] ∨ (v)].

11. Th
143
144
issue that we need to consider, for, so far, I have confined myself to

blameless force. Blameless force is force for which one is neutrally
appraisable, praiseworthy, or not responsible at all. But what if one is
blameworthy for being subject to doxastic force? In other words: what
if one believes that p, one had certain intellectual obligations such that
if one had met those obligations, certain belief-influencing factors
would have changed in such a way that one would not have believed
that p, but one is unable to meet those obligations and is blameworthy
for that?
It seems that there are three views that one could take on this issue.
One might think that blameworthy force provides a full excuse, a partial
excuse, or no excuse at all for having the ensuing beliefs. A full excuse
removes all blameworthiness, a partial excuse reduces the degree of
one’s blameworthiness but does not remove it, and if one has no excuse,
one is fully blameworthy.
In order to decide which position is correct, imagine that Julia has an
intellectual obligation to prepare for her biology exam by reading Miller
and Levine’s Biology. Since she strongly dislikes biology, she tracks any
copy of Biology that she can find in the neighborhood and burns it. A week
later, she takes her exam and, not having read the book and being com-
pletely ignorant (believing certain false propositions and not believing
certain true propositions) on the topic, she miserably fails. Is she excused
for her ignorance on certain biological propositions whose truth value
she could and should have known—and, hence, believed—in virtue of
her blameworthy inability to meet her intellectual obligation after she
has burnt all copies of the book? It is clear that the force to which she is
subject seems to provide no excuse whatsoever. And the same seems true
for other, similar scenarios: if one intentionally violates an intellectual
obligation by making oneself unable to meet it, one is blameworthy both
for that and for the doxastic attitudes relevantly issuing from it. Hence,
at least sometimes, blameworthy force provides no doxastic excuse at all.
It seems, however, that there are other situations in which blame-
worthy force provides a partial excuse. Imagine that Meghan is raised in
a racist family. At a certain age, she becomes aware of the fact that as a
result of her family background, she is quite narrow-m inded. However,
144
145
she culpably fails to work on this intellectual vice of hers. In college, she
has to read certain books on Jewish history, but being narrow-m inded,
she literally cannot resist the impulse to put the books aside. Next, she
becomes a teacher in a remote village without any access to libraries
or the Internet. One of her first projects as a teacher is to test the stu-
dents’ knowledge of Jewish history, but, not knowing anything about
Jewish history herself, she cannot meet her professional obligation.
Consequently, she has certain false beliefs and lacks certain true beliefs
about the students’ knowledge of Jewish history. Now, is she blamewor-
thy for that ignorance? It seems that she is. But it seems that she is less
blameworthy for it than if, say, she has sufficient knowledge about Jewish
history but hates Jews so much that she fails to test the students’ knowl-
edge. What this suggests is that if some person S fails to meet an intel-
lectual obligation to φ because she is unable to do so and she is unable to
do so because she culpably violated an earlier intellectual obligation to
χ, she is blameworthy for ~φ-i ng and for the ensuing doxastic attitudes,
but less so than if she had been able to φ but culpably failed to do so.
Thus, sometimes blameworthy force provides no excuse and some-
times it provides a partial excuse. More precisely, it seems that if one
intentionally renders oneself unable to meet an intellectual obligation,
blameworthy force provides no excuse, whereas if one does not inten-
tionally render oneself unable to meet one’s intellectual obligation,
blameworthy force provides (at least) a partial excuse.12
Are there any situations in which blameworthy force provides a full
excuse? It seems to me that there are two candidates here.
First, one might think that if blameworthy force sometimes pro-
vides a partial excuse, then a long series of violations of obligations that
are due to force will result in a full excuse.13 I think this is a promising
12. Th at only blameless force excuses has been rightly pointed out by Sinnott-A rmstrong
(1984, 250), Stocker (1971, 314), Van Woudenberg (2009, 383).
13. Th is may be what Foley (2005b, 340) has in mind when he says that sloppy evidence
gathering decades earlier in one’s life does not entail that all of one’s subsequent beliefs
are irresponsible even if one would not have had those beliefs if one had met one’s intel-
lectual obligations.
145
146
idea. For it seems that we blame some person S more for violating an
intellectual obligation from blameworthy force which issued directly
from the violation of some obligation, than we blame S for violating
an intellectual obligation from, say, blameworthy force which issued
from another instance of blameworthy force which, again, issued from
another instance of blameworthy force which issued from the violation
of another obligation—at least, if the obligations in the latter scenario
do not bear more weight than the obligations in the former scenario.
In order for one to be completely blameless, however, the chain
would have to be fairly long and I doubt whether that ever happens
in our lives. More precisely, it would have to be the case that, say,
(i) S violates an obligation O, (ii) as a result of that, S cannot but
violate another obligation O*, (iii) as a result of that, S cannot but violate
another obligation O**, (iv) as a result of that, S cannot but violate another obli-
gation O***, and (v) as a result of that, S cannot but believe that p. This
situation is different from a situation in which someone violates an obli-
gation and, thereby, decades later cannot but violate an intellectual obli-
gation or cannot but believe that p. The former situation seems merely
logically possible. In real life we often have multiple occasions to meet
an obligation that we violated at an earlier time in our lives, or we violate
an intellectual obligation that has doxastic repercussions years or even
decades later. Here is an example of the latter. Imagine that a famous poli-
tician gives his diaries to a well-k nown historian and that that historian
promises the politician and even publicly announces that he will publish
that politician’s diaries after that politician’s death. Imagine also that
he immediately destroys the politician’s diaries once he receives them.
When, twenty years later, the politician dies, the historian is under an
intellectual obligation to study carefully the politician’s diaries, in order
to publish them, but he is unable to do so. He, therefore, maintains certain
false beliefs that he would have abandoned if he had studied the diaries.
Second, one might think that if, at the time at which one culpably
performed or failed to perform an action issuing in one’s being forced to
violate an intellectual obligation, one could not foresee that that would
force one to violate that (kind of) intellectual obligation at some future
time, then one is fully excused for the violation of that intellectual
146
147
obligation. Imagine, for instance, that Julia strongly dislikes biology

and, therefore, burns her biology book. Two weeks later and utterly
unforeseeably for Julia, her psychotic roommate threatens to kill all her
other roommates if Julia does not acquire ten true beliefs about photo-
synthesis within half an hour. Assuming that reading her biology book
is the only way that Julia could acquire those beliefs, she is forced to
violate this moral intellectual obligation of hers. But it seems clear that
she is morally blameworthy neither for violating that intellectual obliga-
tion nor for its doxastic and non-doxastic consequences. I agree that in
such cases, one is excused for the violation of the intellectual obligation
in question and for the ensuing doxastic attitudes. However, we should
notice that what excuses one in such cases is one’s lack of foresight, that
is, one’s ignorance of the fact that by violating the obligation O in ques-
tion, one will become unable to meet one’s future intellectual obligation
O*. For if one foresees that but still violates O, it is clear that one is not
excused for violating O* or for thereby being forced to have the ensu-
ing beliefs. In chapter 5, I explore in detail when ignorance counts as an
excuse.
If what I have argued is correct, blameworthy force provides either
a partial or no excuse for one’s beliefs or the lack thereof at all. It is
logically possible that blameworthy force provides a full excuse, but it
seems that that possibility is hardly ever, if ever at all, actualized. But
if de facto only blameless force provides a full doxastic excuse, then
should we not change clause (ii) of our analysis of responsible belief
as follows: “there is some belief-influencing action or series of actions
A that S could have performed such that if S had performed A, S would
not have believed that p or S is blameworthy for the fact that there is no
such action or series of actions”? No, for if S is blameworthy for the fact
that there is no such action, then there must have been some further
action B such that if S had performed B, S would have been able to per-
form A. But if there was such an action B, then B itself counts as an
action or part of a series of actions such that if S had performed B, S
would not have believed that p. Hence, proposition (3)—our analysis
of responsible belief—needs no revision to make sense of the fact that
only blameless force fully excuses.
147
148
4.5 C ONTROL OVER OR INFLUENCE

ON PROPERTIES OF BELIEFS
The first objection to the thesis that doxastic responsibility entails the
ability to believe otherwise is that its adherent fails to take into account
our control over and influence on certain properties of beliefs. We may,
for instance, be blameworthy and, hence, responsible for a belief if we
are responsible for the fact that that belief has certain epistemically bad
properties, even if we could not have avoided having that belief itself.
According to Nikolaj Nottelmann, for instance, we can be blameworthy
for an unreasonable or inadequately grounded belief, even if we could
not have avoided having that belief. Responsibility for a belief, then, does
not entail the ability to believe otherwise, but only control over or influ-
ence on certain properties of that belief.14
I think that this objection does not hit the target. For it seems that
control over or influence on morally or epistemically desirable or unde-
sirable properties of beliefs, although it is surely relevant to something, is
irrelevant to the issue of blameworthy or responsible belief. If S has no
control over or influence on whether or not she holds some belief that p,
but at least some influence on whether or not that belief is reasonable or
based on an adequate ground, then, it seems, she can be held responsible
for whether or not her belief that p exemplifies these properties, but not
for whether or not she believes that p. We are looking for an account of
responsible belief, that is, an account of whether one believes responsibly
or not. Doxastic responsibility is responsibility for the actualization of
the state of affairs of believing a proposition, not responsibility for one’s
belief ’s exemplifying certain properties.15 Thus, even though control
over or influence on certain properties of belief, such as its rationality, is
important, it is not relevant to our theory of responsible belief.
In order to drive the point home, consider the following scenario.
A patient is seriously ill and his death is literally unavoidable—t here is
nothing the doctors and nurses can do in order to save him. However,
14. S ee Nottelmann (2007, 53–72).

15. Nottelmann’s own analysandum is an agent’s being epistemically blameworthy for
holding some belief that p; see Nottelmann (2007, 203).
148
149
there are other things they can do that make a difference to how he
dies. They can let him suffer or give him morphine, and they can let
him die lonely or let him be surrounded by family and friends. Thus,
although they do not have any control over or influence on whether
or not the patient dies, they have influence on how he dies, that is, on
certain properties of the event of the patient’s dying. They are, clearly,
not blameworthy for the patient’s death, although they may be blame-
worthy for how he dies. The realms of action and belief, then, seem
identical in this regard: one is responsible for the actualization of a
state of affairs only if one has control over or influence on whether that
state of affairs is actualized. Whether one has control over or influence
on the exemplification of certain properties by that state of affairs is
immaterial to that.16
One might suggest that if one is blameworthy for a belief ’s exem-
plifying certain undesirable properties, one is blameworthy for the
actual belief one has, including the properties that that belief exem-
plifies.17 However, such an approach seems unpromising to me. One
is blameworthy for one’s actual belief that p if one is blameworthy for
the actualization of a specific set of states of affairs. However, believ-
ing that p cannot be a member of that set of states of affairs, for one is
not blameworthy for that. Things are similar with regard to the patient’s
death. The patient’s death seems to be the actualization of a large set
of states of affairs. Now, the doctors may be responsible for the actu-
alization of some of those states of affairs, but not for the actualization
of the state of affairs that the patient dies. For that, I have assumed, was
strictly unavoidable for them. This is even acknowledged in ordinary
language, for we would not say in the above scenario that the doctors are
blameworthy or responsible for the patient’s death. That is because what
we primarily refer to in talking about the patient’s death is the state of
affairs of the patient’s dying. Similarly, it seems to me, we should not say
that someone’s belief that p is blameworthy if that person is not blame-
worthy for the actualization of the state of affairs that she believes that p
16. Thus also Montmarquet (1993, 47–4 8).

17. N
ottelmann suggested this to me in personal conversation.
149
150
but only for the actualization of the state of affairs that her belief that p is
irrational or unreliably formed.
4.6 D OX ASTIC FRANKFURT-S T YLE

SCENARIOS
The second objection to the thesis that doxastic responsibility entails

the ability to believe otherwise is slightly more complex. The basic idea
is that the thesis is false because it rules out responsibility for beliefs in
doxastic analogues of Frankfurt-style scenarios. Let me explain what
I mean by that. Harry Frankfurt famously argued that responsibility
for actions does not require the ability to act otherwise.18 For instance,
I may be responsible for throwing a girl into a canal even if I could not
have acted otherwise, because if I had decided not to do so, an evil neu-
rosurgeon, who has been waiting in the wings to see what I would do,
would have forced me to do it via a device that he has implanted in my
brain. Frankfurt’s argument is confined to actions and omissions.19
I call such examples Frankfurt scenarios. However, something simi-
lar has been argued for the consequences of our actions and omissions,
most notably by John Fischer and Mark Ravizza. I dub such examples
Frankfurt-style scenarios.
In this section, I consider whether our intuitions in at least some
doxastic Frankfurt-style scenarios count against the thesis that respon-
sibility for belief entails the ability to believe otherwise. I call those who
believe that they do strict doxastic compatibilists. Strict doxastic com-
patibilists ought to be distinguished from doxastic compatibilists, who,
as we saw in c hapter 2 (§2.6), claim that intentional doxastic control is not
required for having doxastic obligations. The adherents of doxastic com-
patibilism as discussed in chapter 2 may very well also claim that dox-
astic responsibility entails the ability to believe otherwise and thereby
reject strict doxastic compatibilism. Also, strict doxastic compatibilists
18. See Frankfurt (1969, 829–837; 1993, 287).

19. F
or the latter, see Frankfurt (1993, 292–293).
150
151
might embrace the idea that beliefs are under our control only if we can
form them as a result of an intention to do so. Doxastic compatibilism
and strict doxastic compatibilism, therefore, ought to be clearly distin-
guished. Since I have argued against doxastic compatibilism in detail in
chapter 2, I here confine myself to strict doxastic compatibilism.
For our purposes, it is important to note that, in opposition to
Frankfurt scenarios and many non-doxastic Frankfurt-style scenarios,
doxastic Frankfurt-style scenarios come in two varieties. First, there are,
of course, scenarios in which a belief is unavoidable because it issues
from a belief-influencing action or omission for which I seem respon-
sible despite being unable to do otherwise. Imagine, for instance, that
Julia is raised in a racist family. As a result of that, she is firmly disposed
to form racist beliefs. She realizes how bad it is that she has such a dis-
position. One day, she gets a once-in-a-lifetime opportunity to freely
attend a race issues class, a training that would rid her of her racist belief-
forming dispositions. Julia, however, decides not to attend the meeting
because she is lazy and careless. It seems that she is blameworthy for
not attending the meeting and maybe also for her racist beliefs that she
maintains as a result of this omission. That seems true, even if her fam-
ily members, who heard of it, were waiting in the wings and would have
prevented her from attending the meeting if she had intended to go. This
first kind of doxastic Frankfurt-style cases, then, is relevantly analogous
to most non-doxastic Frankfurt-style cases.
The second variety of doxastic Frankfurt-style cases, however, is
quite unlike most non-doxastic Frankfurt-style cases that we find in the
literature. Adopting one of Nottelmann’s examples, imagine that Julia’s
family does not know anything about Julia’s opportunity to attend the
race issues class and that Julia decides not to attend the meeting. A year
later, she is at a conference. At that conference, a racist demagogue
presents certain crime statistics that happen to be correct. On the basis
of her racist inclinations, she believes that those statistics are correct.
However, there is another, clearly reliable and non-racist speaker, who
presented the exact same crime statistics an hour earlier. If Julia had
attended the race issues class, as she should have, her belief-forming
dispositions would have changed in such a way that she would have
believed the statistics on the basis of the reliable speaker’s testimony.
151
152
It follows that Julia could not have failed to form the belief that the sta-
tistics are correct, for whether or not she meets her intellectual obliga-
tions, she ends up believing the crime statistics. The thesis that doxastic
responsibility entails the ability to believe otherwise implies that she is
not responsible for her belief. But it seems that Julia is blameworthy for
believing that the statistics are correct if she believes them on the basis
of the unreliable speaker’s testimony.20 The nice thing about this sec-
ond variety of doxastic Frankfurt-style cases is that Julia’s situation is
one of alternate possibilities, because both meeting and failing to meet
her intellectual obligation are options that are available to her. This kind
of doxastic Frankfurt-style scenario, then, has the advantage that it can
be used to make a case against the thesis that doxastic responsibility
requires the ability to believe otherwise without appealing to counter-
factual interveners, which is something that certain philosophers find
problematic.
There is an answer that has been given to the challenge from non-dox-
astic Frankfurt-style scenarios that, it seems to me, can be transposed to
the doxastic realm. Let us first consider the non-doxastic realm. Imagine
that Sam strongly dislikes children and that upon passing by a canal in
Amsterdam he evilly decides to throw a girl who is walking next to him
into it. Unbeknownst to him, an evil surgeon, who has implanted a device
in his brain, would have forced him to do so if he had not decided to do
it. Is he responsible for the unavoidable consequence of the girl’s being
thrown into the water? I agree with Peter van Inwagen that the distinc-
tion between event-particulars and event-universals or, more specifically,
consequence-particulars and consequence-universals, provides the right
conceptual tools to analyze such scenarios.21 It is, of course, a matter of
great controversy how events are to be individuated. But let us assume
with Van Inwagen that they are to be individuated by their causes. Thus,
20. See Nottelmann (2007, 162). A few epistemologists have discussed doxastic Frankfurt-
style scenarios or similar kinds of scenarios. Among them are Corlett (2008, 191–192),
Jäger (2004, 218–225), Leon (2002, 424–430), Zagzebski (2001, 148–151).
21. See Van Inwagen (1983, 166–170). This distinction, but not the conclusions drawn
by Van Inwagen on the basis of it, is also accepted by many compatibilists. See, for
instance, Fischer and Ravizza (1993b, 326–327).
152
153
the same consequence-u niversal, say, the girl’s being thrown into the
water, can be brought about in different ways and may therefore be
instantiated by different consequence-particulars, consequences that
are more finely individuated than consequence-u niversals. If Sam had
been forced to throw the girl into the water, the same consequence-u ni-
versal would have occurred, for the girl would have been thrown into
the water. But it is not the case that the same consequence-particular
would have occurred, for it would have been the evil surgeon’s device
rather than Sam who would have caused the girl’s being thrown into
the water. Thus, in certain Frankfurt-style scenarios people are respon-
sible for the relevant consequence-particular, but not for the relevant
consequence-u niversal.22
What I would like to suggest is that this strategy can be applied
mutatis mutandis in the doxastic realm. Propositional belief can have
different causes. I believe that there are at least two people in the yard
when I see two people walking in the yard, when I hear my friend’s tes-
timony that there are two people in the yard, or when I hear three dif-
ferent voices in the yard. Thus, the same belief-u niversal—belief in the
same proposition—may have many different causes. But then there are
many different belief-particulars.
I think this helps to make sense of doxastic Frankfurt-style sce-
narios. Let us first consider the second variety of Frankfurt-style sce-
narios. If, contrary to fact, Julia had met her obligations, she would have
believed that p, but she would not have believed that p on the basis of
the demagogue’s testimony. Rather, she would have believed that p on the
basis of the reliable speaker’s testimony. There is, as such, nothing wrong
with her believing the crime statistics. There is something wrong with
22. Van Inwagen’s proposal has been criticized by Fischer and Ravizza; see Fischer (1994,
140–147), Fischer and Ravizza (1998, 99–101). According to Fischer and Ravizza, in
Frankfurt-style scenarios, the agent lacks control over what happens in the alterna-
tive scenario. They conclude that responsibility requires at most that something else
could have happened, not that one had the ability to do otherwise. However, their criti-
cism would apply only in the first and not in the second kind of doxastic Frankfurt-style
cases and I will grant that in the first kind of doxastic Frankfurt-style cases, the agent is
not responsible for the belief in question. Hence, their argument does not undermine
my response to the objection from our intuitions in doxastic Frankfurt-style scenarios.
153
154
her believing them merely on the basis of the testimony of a person who
is racist and unreliable. For Julia would not have done that if she had
met her intellectual obligations. And that is why she is blameworthy for
that. More generally, in doxastic Frankfurt-style scenarios of the sec-
ond variety, one is blameworthy for the consequence-particular of one’s
belief, a belief that has its causes (its sources) essentially, but not for
the consequence-u niversal of one’s belief. For the former was avoidable
and would have been avoided if one had met one’s intellectual obliga-
tions, whereas the latter was unavoidable. Thus, what one is responsible
and, more specifically, blameworthy for in such cases is holding a belief
on a bad or inadequate ground, not holding the belief as such.
This is not to deny that one may feel some inclination to blame Julia
for her belief itself in the second doxastic Frankfurt-style scenario that
I sketched above, but my account provides a double explanation of that
inclination. First, we easily mentally confuse this situation with one in
which Julia could have failed to believe the crime statistics upon hearing
the demagogue’s testimony, because in that alternative situation there is
no reliable testimony in support of those crime statistics. Second, there
is something very similar to Julia’s believing the crime statistics with
which it is easily confused, namely Julia’s belief-particular in the crime
statistics, where that belief would not be that belief if it had been based
on different testimony. Thus, we can make good sense of our intuitions
about Julia’s belief without holding her responsible or blaming her for
having that belief.
The account of responsible belief that I have defended so far in
this book does not specify whether I have belief-universals or belief-
particulars in mind. I submit that it can be interpreted in both ways, as
long as the word “belief” is understood univocally in the analysis: respon-
sibility for belief-universals entails that one could have had a different
belief-universal and responsibility for belief-particulars entails that one
could have had a different belief-particular. The same applies to proposi-
tion (2) above, the thesis that doxastic responsibility entails the ability
to believe otherwise: it can be understood as a principle about belief-
universals and it can be understood as a principle about belief-particulars.
This implies that, even though Julia is blameworthy for her belief—
or, more precisely, her belief-particular—in the second kind of doxastic
154
155
Frankfurt-style case, she is not blameworthy for her belief in the first kind
of doxastic Frankfurt-style scenario—neither for her belief-universal nor
for her belief-particular. In the first scenario, Julia freely decides not to
attend the racial issues class, but would have been forced not to attend it
by her family if she had intended to attend class. In the actual scenario,
the causes of her racist belief—say, the testimony of her parents and her
psychological constitution—are the same as the causes of her belief in
the counterfactual scenario in which her family forces her not to attend
the meeting. Thus, even her belief-particular is unavoidable in this sce-
nario. It seems to me, however, that this squares well with our intuitions.
If we know that her family was watching her and would have prevented
her from attending class if she had intended to attend class, we can only
properly blame her for having decided not to attend class and for the
consequence-particular of freely not having attended class, but not for
maintaining her racist beliefs (assuming, at least, that attending class has
been the only opportunity to get rid of her racist beliefs). For we know
that even if she had acted as she should have acted, she would have main-
tained that belief.
Now, does it follow from what I have argued that compatibilism
in general is false? If it did, then, given the large number of compati-
bilists nowadays, many philosophers would perhaps be inclined to jet-
tison the thesis that doxastic responsibility entails the ability to believe
otherwise. Fortunately, no such thing follows. First, in what I said
about beliefs and belief-influencing actions, I have confined myself to
consequence-u niversals and consequence-particulars. I have not said any-
thing about actions and omissions. It is not at all uncommon among phi-
losophers writing on this topic to argue for an asymmetry between the
extent to which responsibility for respectively an action, an omission,
and a consequence requires the ability to avoid that action, omission,
or consequence. 23 Strict compatibilism may still be true for actions and
omissions—my account simply says nothing about that.
23. Van Inwagen, for instance, agrees that compatibilism is true for actions, but argues
that it is false for omissions and consequences; see Van Inwagen (1983, 161–170). And
Fischer argues that compatibilism is true for actions and consequences, but not for
omissions; see Fischer (1985–1986).
155
156
Moreover, as we saw in c hapter 2, doxastic attitudes are different from

many other consequences of our actions and omissions for which we are
responsible, in that in performing belief-influencing actions or omis-
sions, we hardly ever foresee which particular doxastic attitude we will
acquire as a result of that. When, in a typical Frankfurt-style scenario,
I shoot the mayor or walk on without attempting to rescue a drowning
girl, I foresee which consequence-particular will occur as a result of that,
namely respectively the mayor’s being shot and the girl’s drowning.24
In a doxastic Frankfurt-style scenario, however, I do not foresee which
belief I will form or maintain as the result of the violation of an intellec-
tual obligation. Usually, we merely foresee which kinds of belief we will
maintain or acquire. Thus, by not reading a book on herons, I know that
I will remain ignorant of certain true propositions about herons, but I do
not know which things I would come to believe or which beliefs I would
abandon if I were to read the book.
Precisely how is this relevant? Well, if I intend to shoot the mayor, then
I foresee that if I am successful, I will bring it about that the consequence-
universal of the mayor’s being shot occurs. One might think, therefore,
that I bear responsibility for the mayor’s being shot, even when a counter-
factual intervener was waiting in the wings, so that the occurrence of the
consequence-universal was strictly unavoidable. After all, I took it to be
avoidable and I took it that whether or not it would occur was up to me.
When one violates an intellectual obligation, however, one has no such
foresight. It hardly ever happens, if ever at all, that some belief is unavoid-
able but that one takes it to be avoidable, that is, that one takes it that there
is some belief-influencing action such that if one performs it, one will not
have that specific belief. It is much more plausible that one is responsible
for something unavoidable but foreseen and taken to be avoidable than
for something unavoidable but neither foreseen nor taken to be avoidable.
Hence, even if one is not a compatibilist about beliefs and other
non-foreseeable consequences, one might still be a compatibilist about
foreseeable consequences. Of course, these are not more than a few very
broad brushstrokes. A substantial defense of this position would require
more detail and much more by way of argument. Nonetheless, I think
24. See, for instance, Fischer and Ravizza (1992, 376–379; 1998, 93–95), Haji (2000a, 264).
156
157
it shows at least that there is some room to argue that doxastic incom-
patibilists may remain compatibilists about foreseeable consequences.
In conjunction with the fact that my account says nothing about actions
and omissions, I do not think that the fact that many ethicists are com-
patibilists poses a problem for the thesis that doxastic responsibility
requires the ability to believe otherwise.
4.7 T HE DOX ASTIC ASYMMETRY

OBJECTION
The final objection to the thesis that doxastic responsibility entails the
ability to believe otherwise runs as follows. I argued in chapter 1 (§1.4)
that responsible belief is blameless belief for which one is responsible.
Thus, responsible belief is neutrally appraisable or praiseworthy belief.
However, all the examples in favor of the idea that doxastic responsi-
bility requires the ability to believe otherwise that I gave above are
cases of blameworthy belief. According to some philosophers, though,
blameworthiness requires the ability to do otherwise, whereas praise-
worthiness does not. The doxastic asymmetry thesis says nothing about
neutral appraisability. Maybe this is because this normative attitude is
hard to distinguish from certain merely evaluative attitudes. In what fol-
lows, I focus on praiseworthy belief, taking it that if both blameworthy
and praiseworthy belief entail the ability to believe otherwise, neutral
doxastic appraisability does so as well. Since on this objection respon-
sible belief and blameworthy belief are asymmetrical as to whether they
require the ability to believe otherwise, let me call it the doxastic asym-
metry objection. 25
25. Th is thesis, then, is analogous to the asymmetry thesis in ethics according to which
blameworthiness for actions requires the ability to act otherwise, whereas praiseworthi-
ness for actions does not. It is not analogous to two other theses in ethics that have also
been referred to as the “asymmetry thesis,” namely the thesis that responsibility for
actions entails the ability to act otherwise whereas responsibility for omissions does not
(see Fischer (1985–1986), Fischer and Ravizza (1991, 261)), and the thesis that right-
ness and wrongness entail the ability to act otherwise, whereas praiseworthiness and
blameworthiness do not (see Haji (2000b)).
157
158
The only philosopher I know of who has put forward something

like the doxastic asymmetry objection is Weatherson.26 As we saw in
chapter 1, however, his argument in support of this thesis fails. Either
the praise that Weatherson’s cricket captain deserves is merely evalu-
ative rather than normative, or the captain is the proper object of the
normative attitude of praise in virtue of his control over or influence on
his imaginative field placing. Let us, therefore, consider the main argu-
ments that have been provided for asymmetry in ethics and see whether
they can be transposed to the doxastic realm.
First, according to Susan Wolf, if one’s actions are not psychologi-
cally determined, they seem rather whimsical and it is hard to see how
one could ever be praiseworthy for them.27 By “psychological determi-
nation,” she means the inability to do otherwise, given one’s beliefs and
other evidence, one’s desires, one’s character, and so forth. Transposing
this to the doxastic realm, the idea would be that belief-influencing actions
should be psychologically determined in order for them not to be a mat-
ter of whim—in chapter 2, I already agreed that virtually all of our
beliefs are psychologically determined.28
I do not think that this argument is sound. Imagine that I culpably
rather than praiseworthily believe that p. On the doxastic asymmetry
objection, this means that I could have performed a belief-i nfluencing
action such that if I had performed it, I would not have believed that
p. That means that such an action would have been neutrally apprais-
able or even praiseworthy and the ensuing beliefs would have been
neutrally appraisable or even praiseworthy. But if that is true, then
at least some praiseworthy belief-influencing actions are not psycho-
logically determined, namely those that I could have performed when
I did in fact perform a blameworthy belief-influencing action. Thus,
one can have praiseworthy beliefs and one can perform praisewor-
thy belief-influencing actions, even if the latter were not psychologi-
cally determined and if, therefore, the former were avoidable. Hence,
26. S ee Weatherson (2008, 551–552).

27. S ee Wolf (1980, 153–154, 162).
28. Paradoxically, she also says that praise is compatible with its not being determined what
one does and that if one’s action is not determined, then that fact does not render one’s
action whimsical; see Wolf (1980, 165; 1990, 81).
158
159
psychological underdetermination does not entail whimsicality.

Therefore, psychological underdetermination does not count against
the thesis that doxastic responsibility entails the ability to believe
otherwise.
Second, according to Wolf, we praise people without asking our-
selves whether they could have done otherwise. 29 Thus, we praise peo-
ple for not telling a lie, not hurting a fly, buying a present for a friend,
speaking out against injustice, helping a family member who is in trou-
ble, and saving a drowning child, without asking ourselves whether
they could have done otherwise. In fact, it seems that we continue to
praise them if we find out that they could not have done otherwise. All
that we require for praiseworthiness is that people do something for
the right reason. I think it is not easy to transpose this argument to the
doxastic realm, for it seems much harder to think of belief-i nfluencing
actions that one could not fail to perform. It seems that it is normally
up to the individual whether or not one gathers further evidence on
a topic or works on a virtue or vice. Let us, nonetheless, assume that
there are such actions. Perhaps certain people are so intellectually con-
scientious that they cannot fail to gather further evidence on beliefs
that they consider incompatible with each other and which they take
themselves to hold.
I think all these examples provide no good reason to think that
praiseworthy belief does not entail the ability to believe otherwise.
There are at least two good reasons why we often praise people with-
out asking ourselves whether they could have done otherwise. First, as
I argued in chapter 1, not all kinds of praise express normative attitudes.
In other words, in praising people for φ-i ng, we do not always hold them
responsible for φ-ing. I can praise Miranda for her beauty in the same
way as I can praise my Ford for its speed. Many instances of praise seem
to express admiration or gratitude or some such evaluative rather than
normative attitude. Second, in those cases in which we do express a nor-
mative attitude in praising someone for φ-ing, we often assume—t hat
is, we often dormantly or tacitly believe—that they had control over
29. S ee Wolf (1980, 155–159; 1990, 59, 67–93). Thus also Smith (2005, 251, 257).
159
160
or influence on φ-ing. Maybe we praise Fred for being unable to tell a

lie, but it seems that that is because we assume that in the course of his
life he has done so much character building that he now cannot but tell
the truth. If science made such enormous progress that we could create
adult human beings out of non-a nimate matter and we could in this way
create human beings that are unable to tell a lie, it seems incorrect to
take the normative attitude of praise toward such creatures and thereby
hold them responsible for telling the truth.
A third objection runs as follows. It seems clear that when it comes
to blameworthy belief we value the ability to believe otherwise. For to be
able to believe otherwise in such a case is to be able to believe responsi-
bly and we value believing responsibly. But when it comes to responsible
belief, what value would there be in the ability to believe otherwise, that
is, to believe blameworthily?30 It seems that there is nothing desirable
about such an ability.
In response, I would like to stress that my claim that responsible
belief entails the ability to believe otherwise is not based on the idea
that that ability is somehow valuable; it may well have no particular
value. The ground for my assertion is merely ontological. We should not
hold someone responsible for some belief if she cannot believe other-
wise, because that would be incorrect: such a person would not be the
proper object of one of the normative attitudes for holding that belief.
Nonetheless, I think there is a certain value to being able to believe oth-
erwise in cases of responsible belief. After all, if what I have argued is
correct, then, if one cannot believe otherwise, one is not responsible
for one’s belief. But if one is not responsible for one’s belief, one is not
the proper object of neutral or positive appraisal. Thus, all of our beliefs
would be such that we are not responsible for them or we are blame-
worthy for them. A situation—t he actual situation, it seems—i n which
we are neutrally appraisable and sometimes even praiseworthy for our
beliefs seems clearly preferable to that, because it enlarges the scope of
our freedom (the scope of our control and influence).
30. Here, I again adapt an argument that has been given by Susan Wolf; see Wolf (1990,
55–58).
160
161
Finally, one might object that if God exists, he is praiseworthy if

anyone is. But God could not fail to be praiseworthy or fail to perform
praiseworthy actions. Hence, there is asymmetry. Again, I think it is
quite difficult to transpose this particular argument to the doxastic
realm. It seems that those who praise God, praise God for who he is
and for certain of his actions, but not for his beliefs. Nor do they praise
God for performing certain belief-influencing actions, that is, actions
that make a difference to what he believes. For, at least on the classical
Anselmian conception of God that the adherents of most monotheis-
tic religions embrace, God is necessarily omniscient. That means that
there are no belief-influencing actions that God could perform: God
cannot do any evidence gathering, let alone improve his doxastic mech-
anisms or work on his intellectual virtues. The only belief-i nfluencing
action that God can perform is to take a decision from which beliefs
about that decision and its consequences issue.
Nevertheless, people of religious faith do seem to praise God for
certain things. But I do not think that this counts in favor of the asym-
metry thesis in general or the more specific doxastic asymmetry thesis
or the objection based on that. First, religious believers praise God for
being gracious, generous, loving, almighty, and so forth. And, presum-
ably, God could not have lacked these attributes. But I do not think that
in praising God in this way, they express a normative attitude, that is,
an attitude by which they deem God responsible for having these attri-
butes. In fact, that whole idea seems bizarre. Rather, believers stand in
awe and express that feeling or conviction by way of praise. Second,
there are other instances of praising God that do express normative atti-
tudes. Thus, Christians praise God for creating the world, sending his
Son, giving his Holy Spirit, and so forth. However, these are all actions
that, according to mainstream Christian theology, God could have
decided not to perform. I conclude that the fact that religious believ-
ers praise God, a being who could not fail to be praiseworthy, does not
count in favor of the doxastic asymmetry thesis.
Since all four objections fail, we have not been given a good reason
to think that there is doxastic asymmetry. But virtually all philosophers
at least initially have the intuition or hold the belief that responsibility
161
162
for φ-ing requires the ability to ~φ—t hat is why Frankfurt’s argument
came as a surprise to so many philosophers. Moreover, if both blame-
worthiness and praiseworthiness entail responsibility and doxastic
blameworthiness entails the ability to believe otherwise, then, unless
we have good reason to think otherwise, it is hard to see why doxastic
praiseworthiness would not entail the ability to believe otherwise. On
balance, then, we have good reason to reject the doxastic asymmetry
thesis and to accept the thesis that responsible belief entails the ability
to believe otherwise.
4.8 C ONCLUSION
In sum, force is an important doxastic excuse. In fact, a large number

of excuses seem to be varieties of excusing force. When we say that a
person violated an intellectual obligation because of intoxication, com-
pulsion, insanity, or duress, or that she violated an intellectual obliga-
tion automatically, involuntarily, or impulsively, or that in violating an
intellectual obligation she was under constraint or not in control, what
we seem to have in mind is that her will or intentions could not make a
difference to what she did or failed to do and that she is excused for the
ensuing belief in virtue of that. We can make a difference to many of our
beliefs by exercising control over certain belief-i nfluencing factors. If in
coming to hold or in maintaining some belief, we could not but violate
the intellectual obligations the meeting of which would have made a dif-
ference to that belief, then we are excused for doing so and for the ensu-
ing belief. If we had another opportunity to influence that belief, the
excuse is narrow: we are still responsible for having that belief and can
still believe responsibly. If there was no other opportunity to do so, then
the excuse is broad: we are not at all responsible for having that belief.
An important insight to be gleaned from this account of force as
a doxastic excuse is that doxastic responsibility entails the ability to
believe otherwise. This thesis is undermined neither by the fact that we
sometimes control or influence which properties our beliefs exemplify,
nor by our intuitions in doxastic analogues of Frankfurt-style scenarios,
162
163
nor by arguments in favor of a presumed asymmetry between blame-

worthy belief and responsible belief. This means that, contrary to what
philosophers such as Nottelmann claim, we can perfectly well construe
an account of responsible belief in counterfactual terms, as conditions
(ii) and (iii) of my analysis of responsible belief do (see proposition
(3)). 31
It also has an important implication for the relation between epis-
temic justification and responsible belief. In c hapter 2, I pointed out that
Alston introduced his argument from doxastic involuntarism as an argu-
ment against the deontological conception of epistemic justification.
However, it is clear that knowing and, hence, being epistemically justi-
fied in believing that p does not require the ability to believe otherwise.
I cannot but believe that I exist and that 4 is larger than 1, even though
these beliefs are perfectly epistemically justified. This means that epis-
temically justified belief cannot be identical to responsible belief. In the
Appendix, I will spell out the relation between epistemically justified
belief and responsible belief in more detail.
31. See Nottelmann (2007, 161–163).
163
164
[5 ]
RESPONSIBLE BELIEF
IS RADICALLY SUBJECTIVE
5.1 I NTRODUCTION
The second doxastic excuse—t hat is, an excuse for holding a particular
belief—t hat I would like to consider is ignorance. That ignorance some-
times excuses has been widely acknowledged ever since Aristotle.1 For
instance, it seems that I am blameless for offering a friend chocolate
pudding if, unbeknownst to me, it was poisoned by one of his jealous
colleagues—u nless, perhaps, I should not have been ignorant. Fewer
philosophers, however, have noticed that ignorance can also excuse one
for having certain beliefs and hardly anyone has spelled out how it can do
so. The aim of this chapter is to fill this lacuna.2
The chapter is organized as follows. First, I address the question what
ignorance is and defend the thesis that there are at least three different
kinds of ignorance (§5.2). Next, I argue that there are four kinds of cases
in which ignorance provides an excuse for holding a belief that issues
from the violation of an intellectual obligation: roughly, ignorance of
1. See Aristotle (2003, 123–129, 145–147, 299–305; NE III.i.13–27; v.7–12; V.viii.3–12).

See also Brandt (1969, 349), Fischer and Ravizza (1998, 12–13), Goldman (1970, 208),
Rosen (2003, 61–62), Smith (1983, 543–571), Zimmerman (2008, 169–2 05).
2. The topic has recently received some attention, especially in two closely related con-
texts: the tracing problem for ignorance, that is, when ignorance counts as an excuse for
ignorance (see several of the essays in Peels (2016b)) and in the context of the debate
on the epistemic condition for responsibility (see several of the papers in Wieland and
Robichaud (2016)).
164
165
R esponsible B elief I s R adically S ubjective
one’s having that obligation, ignorance of one’s ability to meet it, igno-
rance of how to meet it, and lack of foresight regarding that obligation
(§5.3). Subsequently, in order to tighten our grasp upon the notion of
ignorance as a narrow doxastic excuse (an excuse that only lowers the
degree of one’s doxastic blameworthiness), I distinguish it from igno-
rance as a broad doxastic excuse (that is, an excuse that blocks all respon-
sibility for one’s belief) (§5.4). Finally, I argue that the different kinds of
ignorance that I distinguished count as narrow doxastic excuses only if
one is blameless for such ignorance. In doing so, I respond to the worry
that the thesis that only blameless ignorance provides a doxastic excuse
leads to an infinite regress (§5.5).
Throughout the entire chapter, I assume that, unless indicated oth-
erwise, the ignorance under consideration is blameless. If what I argue is
correct, then at least two things follow. First, ignorance sometimes pro-
vides a narrow doxastic excuse—t hat is, sometimes one believes respon-
sibly despite having violated an intellectual obligation in coming to hold
that belief, in virtue of the fact that one is in a particular way excused for
it by ignorance. Second, whether or not one believes responsibly is, in a
sense that I explain below, radically subjective.
5.2 W HAT IS IGNORANCE?
The word “ignorance” derives from the Latin in (not) and gnarus (knowl-
edgeable, acquainted with). This suggests that to be ignorant is to lack
knowledge. Let us call this view the Standard View. It is embraced by,
among others, Julia Driver, Peter Unger, and Michael Zimmerman. 3
Thus, if Marion is ignorant that Julia will pick her up at St. George
Street, then on the Standard View, she does not know this. Etymology,
however, can be misleading. For example, the English verb “to believe”
derives from the German “belieben” (to love), but nobody thinks that,
3. The Standard View that ignorance is lack of knowledge is explicitly accepted by Fields
(1994, 403), Haack (2001, 25), Zimmerman (1988, 75; 2008, ix). It is implicitly accepted
by Anscombe (1963b, 400), Driver (1989, 373–376), Flanagan (1990, 422), Houlgate
(1968, 109), Unger (1975, 93).
165
166
therefore, propositional belief entails some affection or other. On an

alternative view, which has been called the New View and which I have
defended elsewhere, ignorance is lack of true belief.4
In order to see how these two views differ from each other, let us
consider the ways in which one can fail to know that p and then consider
whether they count as being ignorant that p:
(i) p is false;
(ii) S disbelieves the true proposition p;
(iii) S suspends belief on the true proposition p;
(iv) p is true and S neither believes that p, nor disbelieves that p,
nor suspends belief on p;
(v) S believes the true proposition p, but S’s belief that p lacks warrant,
where warrant is that which turns true belief into knowledge.
It seems to me that instances of (i)–(iv) all count as cases of ignorance. As

to (ii) and (iii), if someone disbelieves that Isaac Newton published his
Hypothesis of Light in 1675, then she is ignorant that he did, and if some-
one suspends belief on whether Bangkok is the capital of Thailand, then
he is clearly ignorant of this truth. Instances of (iv) are also instances of
ignorance. Most people on earth have never considered the proposition
p that the nucleus of a helium atom contains two protons and two neu-
trons. Nor do they seem to tacitly believe or tacitly disbelieve this propo-
sition. It follows from our analysis in chapter 1 (§1.3) that they do not
even (tacitly) suspend judgment on such a proposition. They lack certain
concepts that one needs to have in order to grasp the proposition. The
right way to think of a case like this, then, seems to acknowledge that
such people are simply deeply ignorant of the truth of this proposition.
Things are slightly more complicated when it comes to (i). For it seems
that someone who believes a false proposition p and, thereby, lacks knowl-
edge of p, is not ignorant of p. I cannot be ignorant of Naples’ being the
capital of Italy, because it is not. I can only be ignorant of truth. I propose,
then, that we understand the Standard View as saying that to be ignorant
4. See Peels (2010b; 2011a; 2012). For a reply to my papers and a defense of the Standard
View, see Le Morvan (2011; 2012; 2013).
166
167
of p is to not know whether p is true or false. So far, then, we have not found
a counterexample to the view that ignorance is lack of knowledge.
However, the Standard and New Views diverge when it comes to (v).
On the Standard View, cases in which one holds the true belief that p
without warrant are cases of ignorance, whereas on the New View, such
cases do not count as ignorance. One way to illustrate the difference is to
consider a Gettier example. Imagine that I enter my living room and look
at the clock. The clock tells me that it is 3 p.m., so I come to believe that it
is 3 p.m. I know that the clock normally works fine. However, the clock
stopped working twenty-four hours ago. Am I ignorant that it is 3 p.m.?
On the Standard View I am, whereas on the New View, I am not. On the
New View, there are other propositions in the neighborhood of whose
truth I am ignorant, such as that the clock stopped working twenty-four
hours ago and that the clock is unreliable on this particular occasion. Of
the proposition that it is 3 p.m. itself, however, I am not ignorant.
Moreover, on the New View, even cases of mere true belief are
not cases of ignorance. Consider Jim, a truck driver from Miami, who
believes contrary to all the evidence that he is going to be the next presi-
dent of the United States. He thus comes to believe the proposition q
that the next president of the United States currently lives in Miami.
As it turns out, the next president is Ms. Jones, a female congress mem-
ber living in Miami whom Jim has never heard of. In this case Jim truly
believes, but does not know that q. Is he ignorant of q? On the New View,
he is not. There are all sorts of truths in the neighborhood that, on the
New View, he is ignorant of and it is hard mentally to isolate q from all
those other truths, truths such as Ms. Jones is going to be the next presi-
dent, Ms. Jones lives in Miami, and The next president is currently a congress
member. Adherents of the New View sometimes suggest that we may be
inclined to think that Jim is ignorant of q because we know that he is
ignorant of all these other propositions. If we focus on q, however, the
idea is that it seems that Jim is not ignorant of q.
Elsewhere, I have defended the New View in detail. 5 Here, I will not
do so. This is because true belief does not provide an excuse, not even
5. See Peels (2010c; 2011a; 2012; 2014b), Le Morvan and Peels (2016). Of course, one
could also defend views on the nature of ignorance that are in some sense in-between the
167
168
a partial one.6 Thus, even if true belief that lacks sufficient warrant for
knowledge does count as ignorance, it is not relevant to the issue at hand.
Imagine, for instance, that it is my task to set fire to an old barn.
Then, whether I know that there is someone in there or whether I merely
truly believe that there is someone in there does not seem to make a dif-
ference to the degree of my blameworthiness: in both cases I am blame-
worthy to an equally high degree. For whether I know or rationally
believe or merely believe that there is someone in the barn does not
make an important difference to my phenomenology. In all these cases,
I sincerely think that there is someone in the barn; that is how reality
appears to me.
If there are degrees of belief and if they are to be spelled out in terms
of conviction, then maybe I am more blameworthy if I am certain that
there is someone in the barn than if I am merely fairly convinced that
there is someone in it. But notice that such varieties in degree of belief
are not necessarily correlated with whether one knows, believes on good
evidence, or believes without any evidence. One could in principle, quite
irrationally, be one hundred percent sure without having any reasons
or evidence.
Let me stress that I do not contend that a true belief along these
lines renders one blameworthy to the highest degree possible. Maybe
someone who burns the barn in order to do wrong for wrong’s sake is
even more blameworthy.7 In such cases, however, it seems that one’s
evil intention adds something to the degree of one’s blameworthiness.
Whether one believes or knows that there is someone in the barn makes
no difference to the degree of one’s blameworthiness: in both cases, one
is not excused at all. One might think that we should add some ceteris
paribus clause here. For if I had this true belief but also believed that
that belief was irrational or otherwise epistemically defective, then
I may very well be excused for putting fire to the barn. In such cases,
Standard and New Views, such as the view that ignorance is the lack of reliably formed
true belief. For further explorations of the nature and varieties of ignorance, see Peels
and Blaauw (2016).
6. Th
e same claim is made, as a conjecture, by Rosen (2008, 597).
7. Thus, for instance, Beardsley (1979, 577).
168
169
however, I will be excused in virtue of those other beliefs, not in virtue

of my true belief that there is someone in the barn. Thus, true belief that
falls short of knowledge, even if it counts as ignorance, does not provide
an excuse.
Let us, for the purpose of clarity in what follows, define the three
varieties of ignorance that both the Standard and New Views acknowl-
edge as follows:
S is disbelievingly ignorant that p iff (i) it is true that p, and

(1)
(ii) S disbelieves that p.
S is suspendingly ignorant that p iff (i) it is true that p, and (ii)
(2)
S suspends belief on p.
S is deeply ignorant that p iff (i) it is true that p, and (ii) S
(3)
neither believes that p, nor disbelieves that p, nor suspends
belief on p.
Below, we shall see that these varieties of ignorance differ in the extent
to which they excuse one for having some belief. The Standard View dis-
tinguishes a fourth kind of ignorance:
S is unwarrantedly ignorant that p iff (i) it is true that p, (ii) S

(4)
believes that p, and (iii) S does not know that p.
Since I argued above that true belief that lacks warrant, even if it counts
as ignorance, does not excuse, in what follows I focus on disbelieving,
suspending, and deep ignorance.
5.3 IGNORANCE AS
A DOX ASTIC EXCUSE
In c hapter 3, I argued that one can be excused for some belief only if it
issues from the violation of an all-things-considered intellectual obli-
gation. If one has not violated such an obligation, then there is noth-
ing to be excused for. A natural place, then, to look for ignorance that
excuses one for having some belief is ignorance with respect to one’s
169
170
all-things-considered intellectual obligations. I think there are four

kinds of such excusing ignorance.8
1. Put quite crudely, one can be excused for some belief when one
was ignorant that one had the intellectual obligation from the violation
of which that belief issued. More precisely, what is relevant here is that
it was bad from a doxastic point of view: as I argued in chapter 3 (§3.3),
the badness should at least partly be constituted by the moral, pruden-
tial, or epistemic badness of the beliefs issuing from the violation of
such an obligation. Why does such ignorance provide a doxastic excuse?
Well, imagine that Eleonore enrolls as a history graduate student and
that, as a student, it is her intellectual obligation to read the entire two-
hundred-page student manual. However, she has no idea that there is
a manual that she ought to read, since nobody told her so. Hence, she
remains ignorant on certain issues that have to do with her studies. Is
she blameworthy for that? I think it is rather clear that she is not. If one
is blamelessly ignorant of one’s obligation, then it is hard to see how one
can properly be blamed for violating one’s obligation or for the conse-
quences of that.9
However, one can know that not performing some belief-i nfluencing
action is bad and that one has an obligation to perform it and yet be
ignorant that one has an all-things-considered obligation to perform that
action. For instance, I may perfectly well realize that I have an epistemic
obligation to gather evidence on p, but falsely believe that that obliga-
tion is trumped by a moral obligation not do so, so that I am ignorant
that gathering evidence on p is my all-t hings-considered obligation. In
such a scenario, my ignorance may excuse me for violating my all-t hings-
considered intellectual obligation, despite my knowing that I have an
obligation to gather evidence on p.
8. I also distinguish these varieties of excusing ignorance—even though not ignorance as a

doxastic excuse in particular—i n Peels (2014b).
9. For the same intuition, see Plantinga (1990, 52). According to Fields (1994), having
acted from a false moral belief is not an excuse. In his defense of this thesis, however,
Fields conflates blame and disapproval. Also, he provides a pragmatic justification of the
practice of blaming without considering the issue of whether someone deserves blame. As
we saw in chapter 1, what matters for responsibility is whether one is the proper object
of normative attitudes.
170
171
Hence, one kind of ignorance that excuses is ignorance of one’s all-

things-considered obligation rather than ignorance of one’s pro tanto obli-
gation. It follows that one can act from ignorance and be excused by that
ignorance, even if one is not ignorant of every wrong-making feature of
one’s act; ignorance of all-t hings-considered wrongness suffices. Given
what I have said so far, the first kind of ignorance that provides a doxastic
excuse can be defined as follows:
S is excused for believing that p if S was ignorant that she had

(5)
an all-things-considered intellectual obligation O (not) to
actualize Σ such that if she had met O, she would not have
believed that p.
Before we move on to the second kind of excusing ignorance, I would

like to return to my distinction between disbelieving, suspending, and
deep ignorance. I think it is clear that if I am disbelievingly or deeply igno-
rant that not performing some belief-influencing action would be all-
things-considered bad, then I am excused for not performing it. But how
should we think of suspending ignorance? For example, what if Eleonore
suspends judgment on whether it is her intellectual obligation, as a his-
tory graduate student, to read the entire manual? Or what if I suspend
belief on whether my belief that the Mona Lisa was stolen in 1911 fits the
evidence I have? I think it should be clear that suspending ignorance,
even if it is blameless, does not provide a full excuse. If Eleonore sus-
pends belief on whether she should read the entire manual, say, because
she only vaguely remembers someone saying something to that effect,
she should either err on the side of caution by reading the entire manual
or find out whether or not she should read the full manual. This seems to
apply generally, to all intellectual obligations.
Let me stress that what I mean is that suspending ignorance does not
provide a full excuse. It seems that it provides a partial excuse: it seems
that someone who fails to read a manual because she suspends belief
on whether she should do so should not be blamed as much as someone
who is aware that she should do so but fails to do so, say, out of laziness.
2. The second kind of ignorance that provides a doxastic excuse is
ignorance of one’s ability to meet the intellectual obligation in question.
171
172
This kind of excusing ignorance can be analyzed more precisely as

follows:
S is excused for believing that p if S was ignorant that she had

(6)
the ability to meet her all-things-considered obligation O
(not) to actualize Σ such that if she had met O, she would not
have believed that p.
By the “ability to actualize Σ” I mean something rather simple here,

namely the physical or mental feasibility of actualizing Σ. It does not
require that one knows or has true beliefs about how to actualize Σ. It is
because of this rather restricted meaning of the phrase “ability to actual-
ize Σ” that we can distinguish the kind of ignorance discussed here from
the third kind of excusing ignorance, which I discuss below.
Here is an example that illustrates how ignorance that one is able
to meet one’s intellectual obligation can provide a doxastic excuse.
Imagine that a teacher assigns to two of her students, Stephanie and
Rachel, the task of giving a class presentation on Operation Valkyrie, a
well-k nown attempt to assassinate Hitler. Stephanie knows that if she is
to do that, she needs to borrow certain books from the university library
that are not available online. Reading those books is the only way to
acquire the relevant information. However, right after class, large parts
of the city are flooded because of heavy rainfall, so that she cannot reach
the library. By accident, Rachel (not Stephanie) finds out that another
professor of theirs has a large World War II collection, among which are
all the relevant books on Operation Valkyrie, and that he is happy to
lend them to any of his students. Since he would be happy to lend them
to Stephanie as well, but Stephanie is completely unaware of the fact that
he has these books, her ignorance that she is able to meet her obliga-
tion seems to provide a full excuse for not preparing a good presentation
on Operation Valkyrie and for her ensuing doxastic attitudes that she
would not have had if she had met her intellectual obligation.
Does deep ignorance of one’s ability to meet one’s intellectual obli-
gation provide a doxastic excuse? Remember that if one is deeply igno-
rant of p, then one neither believes that p, nor disbelieves that p, nor
suspends belief on p. In normal circumstances, this entails that one has
172
173
never considered whether p is true. It is much harder to imagine that such

ignorance excuses, but that is only because it is much harder to imagine
that one is blameless for not even considering whether p. It is clear, for
instance, that if Stephanie does not go to the library because she has not
thought about it, then that does not excuse her. For if Stephanie has to
give a presentation, she should have thought about going to the library.
However, I submit that if one’s deep ignorance of one’s ability to meet
one’s obligation is blameless, then we cannot blame the agent for violat-
ing an intellectual obligation and for the ensuing beliefs. If Stephanie,
for instance, does not believe that she can go to the library, not even
dormantly or tacitly, then she cannot intentionally go to the library. But
then it seems improper to blame her for not doing so.
Finally, what about suspending ignorance of one’s ability to meet
one’s intellectual obligation? As with suspending ignorance of one’s
all-things-considered obligation, it seems to me that such ignorance
only partially excuses. If Stephanie has no idea whether she can go the
library, then she should find out whether she can do so, or, if it is not too
dangerous, simply try to go to the library. If she neither tries to reach the
library nor investigates the matter further, it seems that her suspend-
ing ignorance does not provide a full excuse. It might even provide no
excuse whatsoever—I will not discuss that here. If she investigates the
issue, but finds no answer, she can simply go and have a look, if it is not
too dangerous. If it then turns out to be impossible, then that (or her
blameless belief that that is the case) counts as her excuse, and if it turns
out to be too dangerous, then she is excused by that (or by her blameless
belief that that is the case).
3. The third kind of proposition that is such that ignorance of it seems
to excuse one in certain circumstances is that there are certain means
that are available to one to meet one’s obligation.10 Or, slightly more
precisely:
S is excused for believing that p if S is ignorant that X 1, X 2 , … ,

(7)
or X n is a sufficiently good means that is available to her to
10. M any philosophers, such as Rosen (2008), overlook this third kind of excusing
ignorance.
173
174
meet her all-things-considered intellectual obligation O

(not) to actualize Σ such that if she had met O, she would
not have believed that p.
This kind of ignorance differs from the one distinguished in section 5.2.
Imagine that Rachel accidentally finds out that another professor has
a large World War II book collection, among which are all the relevant
books on Operation Valkyrie, and tells Stephanie that she (Stephanie)
is able to prepare her presentation, despite the fact that the library can-
not be reached because of recent floods in the city. Imagine also that
Stephanie knows that Rachel is highly reliable, but that they are so intel-
lectually competitive that Rachel refuses to share with Stephanie how
she can get access to the relevant books. Then Stephanie will know that
she can meet her obligation, but she will be ignorant as to how to meet
that obligation. Of course, if one knows how to meet one’s obligation,
then one knows that one can meet one’s obligation. But, as this example
shows, the reverse does not hold. This means that, for the sake of com-
pleteness, we should distinguish this as a separate variety of ignorance,
for even if one knows that one has an intellectual obligation and that one
is able to meet it, one may still be excused by ignorance, namely if one is
ignorant (again, inculpably so) as to how to meet that obligation.
Ignorance of this kind of proposition fully excuses both when it is
disbelieving ignorance and when it is deep ignorance. When Stephanie
believes on the basis of strong evidence that that professor has no
books on Operation Valkyrie or when she has not even considered it
and is blameless for that, it seems that her ignorance of how to meet
her obligation fully excuses her for not preparing a good presentation
on Operation Valkyrie and for the ensuing beliefs and other doxastic
attitudes on the topic she would not have had if she had met her intel-
lectual obligation.
One might think that, mutatis mutandis, the same applies to a
case in which Stephanie knows that she is able to meet her obligation
but suspends judgment about how she can meet her obligation. If she
knows that she can collect the relevant books on Operation Valkyrie,
but suspends judgment on whether professor x has those books and is
willing to lend them, she can send him an email or try to call him. But
174
175
here is the problem: virtually any professor might happen to have a col-

lection of books on World War II among which are the relevant books on
Operation Valkyrie. True, this is unlikely, but Stephanie has no reason
to completely exclude this for those professors she does not know well.
But then if she considers all those professors and suspends judgment in
each case, it would follow that she is not fully excused by her ignorance
of how to meet her obligation, and that seems clearly false.
There are at least two ways to meet this worry, though. First, one
might think that, given the low probabilities involved, it is rational for
Stephanie to disbelieve that rather than suspend judgment on whether a
particular professor x whom she does not know has the relevant books
on Operation Valkyrie, and that if the probabilities were significantly
higher, she would have an obligation to investigate and suspending
ignorance would not fully excuse. Second and more importantly, even
if it is rational to suspend judgment for each professor as to whether he
or she possesses the relevant books on Operation Valkyrie and whether
he or she is willing to lend them, that suspending ignorance (on a large
series of propositions) all by itself does not provide a full excuse. It
is only if we add that it is impossible or out of proportion to find out
whether one of them has the relevant copies of the books that Stephanie
is fully excused.
4. Finally, remember that we saw in chapter 4 that one can be excused
by ignorance if one is culpably forced to violate an intellectual obliga-
tion, namely if one could not foresee that violating a previous obliga-
tion would issue in one’s being forced to violate the relevant intellectual
obligation. We saw that in such cases, one is excused by ignorance rather
than by force. Similarly, it seems that one can be excused by ignorance if
one violates an intellectual obligation from culpable ignorance, namely
if one could not foresee that violating a previous obligation would issue
in one’s being ignorant of (one’s ability to meet) the relevant intellec-
tual obligation. Thus, ignorance sometimes excuses if it issues in force
or ignorance of (one’s ability to meet) one’s all-t hings-considered intel-
lectual obligation. Let me, quite stipulatively, dub this kind of ignorance
lack of foresight.
Imagine, for instance, that I have an obligation to prepare for my
biology exam and that I know that I am able to meet that obligation by
175
176
reading Miller and Levine’s Biology. Imagine also that I culpably fail to
meet this obligation. However, I am inculpably ignorant that the knowl-
edge that I would have acquired by reading the book would have enabled
me to save someone’s life in an unforeseeable situation a week after the
exam. (Let us stipulate that I have an obligation to save that person’s
life; we assume, for instance, that most people would be able to do so.) It
seems that in such a case, I am blameless for not saving that person’s life.
Since by not saving that person’s life I violate an obligation, this means
that I am excused for not saving her life. It is clearly possible to construe
similar examples for intellectual obligations—rather than the obligation
to save someone’s life—a nd doxastic excuses, but since the example of
saving someone’s life is much easier and lack of foresight is already com-
plex enough, I will stick with this example.
In virtue of what am I excused in such a situation? First, one might
think that what excuses me in such a scenario is my ignorance of how to
save that person’s life. This suggestion, however, is unconvincing. If I do
not know how to save that person’s life and if that ignorance is blame-
worthy, it is not yet clear whether I am to be blamed for not saving that
person’s life. For whether or not I am to be blamed for that depends on
whether, at the time at which I violated my obligation to prepare for my
biology exam, I could foresee that not doing so would result in my igno-
rance of propositions that I would have to know in order to save some-
one’s life on that future occasion. Hence, a second, more convincing,
proposal is that what excuses me in such a case is my blameless lack of
foresight regarding my future obligation to save that person’s life and
my inability to meet that obligation in virtue of not preparing for my
biology exam.
This is not to say that every case in which I violate my obligation to
prepare for my biology exam, and in which I later need the relevant bio-
logical knowledge to save someone’s life, and in which my lack of fore-
sight at the time at which I violate my obligation is a scenario in which
lack of foresight excuses me. Imagine that after my biology exam, but
before encountering the victim, I learn that there is life-saving informa-
tion in the book that I need at some point in the nearby future, but that
I can no longer acquire that information. In that case, it is not my lack of
foresight that excuses me for violating my obligation to save the victim’s
176
177
life, for at the time at which I encounter the victim, I no longer lack such
foresight. Rather, it is the inability to gather that relevant information
after my biology exam. All I have claimed here is that in at least some
cases lack of foresight provides a full excuse for violating an obligation.
It is a tricky issue precisely how lack of foresight ought to be cashed
out. Rather than trying to be exhaustive, let me indicate two characteris-
tics of such foresight. First, it seems that foresight does not require occur-
rent beliefs about future intellectual obligations, future force, or future
ignorance. If it did, then we would hardly ever be excused by lack of
foresight, for it is impossible to occurrently foresee all the consequences
of the violation of one’s obligations that are foreseeable. However, in
chapter 1, we saw that we also have dormant and tacit beliefs, and it
seems that such beliefs will suffice for foresight. Second, one need not
believe that one will actually have the relevant intellectual obligation,
or that one will actually be forced to violate that intellectual obligation,
or that one will be ignorant of (one’s ability to meet) that intellectual
obligation. It suffices that one believes that it is sufficiently likely that one
will have that intellectual obligation at some time in the future, and that
by violating one’s present obligation, one sufficiently raises the chances
of being forced to violate that future intellectual obligation or of being
ignorant regarding that obligation.
Let us return again to my distinction between disbelieving, deep,
and suspending ignorance. It seems that, again, only disbelieving and
deep ignorance provide a full excuse when it comes to lack of foresight.
Imagine that I suspend belief on the true proposition that by violating
my obligation, I sufficiently raise the chances of violating certain future
obligations. Imagine also that I nonetheless violate that obligation. It
seems that that will not get me off the hook. True, in at least some cases,
I will not be as blameworthy as I would have been if I had known this
true proposition. But it seems that I am not completely blameless either.
Hence, suspending lack of foresight provides at most a partial excuse.
One may wonder precisely why, in opposition to disbelieving and
deep ignorance, suspending ignorance, as we saw, provides at most a
partial excuse in all these four cases—ignorance of one’s obligation,
ignorance of one’s ability to meet one’s obligation, ignorance of how to
meet one’s obligation, and lack of foresight regarding one’s obligation.
177
178
Here, I have focused on arguing that there is this difference rather than
explaining why there is this difference. What I said provides some sug-
gestions, though: it seems that suspending ignorance, in opposition
to disbelieving and deep ignorance, gives rise to further obligations,
namely an obligation to investigate or find something out if the stakes
are sufficiently high. This issue clearly deserves further consideration,
but given the purposes of this chapter, I will leave it for another occasion.
So far, I have argued that there are four kinds of ignorance that pro-
vide a doxastic excuse: ignorance of one’s all-t hings-considered intellec-
tual obligations to which one’s belief is relevantly related, ignorance of
one’s ability to meet those obligations, ignorance of how to meet those
obligations, and lack of foresight regarding those obligations. I argued
that in such cases disbelieving and deep ignorance provide full doxastic
excuses. I think there are exceptional circumstances, though, in which
disbelieving ignorance does not provide a full doxastic excuse. Imagine
that I believe the proposition q that I cannot gather any further evidence
on a proposition p that I believe. However, imagine also that I hold a
further belief r that my belief that q is irrational or unreliably formed.
Here, I am disbelievingly ignorant that I can gather further evidence
on p, but since I am aware (or, at least, believe) that my disbelief that
q is irrational, I may very well be blameworthy if I fail to gather further
evidence on p.
A correct analysis of ignorance as a doxastic excuse, then, should
include a clause about higher-order beliefs about the epistemic propriety
of one’s lower-order beliefs. If one believes that one cannot meet one’s
obligation, but also that that belief is irrational in that one should believe
that one can meet one’s obligation, then such ignorance does not excuse
(if it can still count as ignorance at all; I will not take a stance on that).
What if one does not believe that, but is disbelievingly, suspendingly,
or deeply ignorant of the fact that one’s lower-order doxastic attitude
is unreliably formed or irrational? Well, if such ignorance is blameless,
then that seems to provide a full excuse, but not if it is blameworthy (I
return to blameworthy ignorance below). It seems that the following
clause captures our intuitions in scenarios of the kind just sketched: “It
is not the case that S believes or is blameworthily ignorant that S’s belief
that p is epistemically defective.” This means that, confining ourselves
178
179
to ignorance that fully excuses, we can summarize the view that I have
defended as follows:
S’s ignorance of some proposition p excuses S for holding

(8)
some belief B iff (i) B issues from the violation of one or
more of S’s all-t hings-considered intellectual obligations O,11
(ii) S’s ignorance of p is disbelieving or deep ignorance, (iii)
p is the proposition that S has O, that S is able to meet O, that
X1, X 2, … , or X n is a sufficiently good means available to S
to meet O, or that by (not) performing an action or series
of actions A, S sufficiently raises the chances of violating O,
and (iv) it is not the case that S believes or is blameworthily
ignorant that S’s attitude toward p is epistemically defective
in that S should believe that p.
Before moving on, I would like to address two worries that one might
have concerning this account of ignorance as a doxastic excuse.
First, one might worry that the idea that mere disbelieving igno-
rance can excuse is too strong in that it contradicts the principle, advo-
cated by a number of philosophers, that one should act from p (treat p as
a reason to act) only if one knows that p.12 I do not think that this objec-
tion holds up under scrutiny, though. On the one hand, if the principle
does not allow for excusing circumstances, it is clearly false. Someone
who believes on the basis of good evidence that p and has no reason to
mistrust her belief that p clearly blamelessly acts on p, even if it turns
out that p is false. On the other hand, if the principle allows for excusing
circumstances, as it should, then blameless disbelieving ignorance looks
like one of the best candidates. If one believes that something is (not)
11. Here, I mean that B issues from the violation of S’s intellectual obligations in the way
specified in section 3.6.
12. For a defense of this principle, see, for instance, Hawthorne and Stanley (2008, 577–
578). It is important to note that this principle is not meant as a merely evaluative princi-
ple, but as a principle that has to do with responsibility, as is evidenced by Hawthorne’s
and Stanley’s use of the terms “blame” (p. 572, p. 587) and “excuse” (p. 573, p. 582).
Hawthorne and Stanley explicitly allow that one can be excused for violating the prin-
ciple that one should act only from knowledge.
179
180
the case and has no reason to mistrust that belief, then it seems that one
is off the hook for acting on that belief, even if it is false or otherwise fails
to be an instance of knowledge.
Second, one might think that S is doxastically excused by igno-
rance only if S violates the relevant intellectual obligations from igno-
rance or because she is ignorant. Imagine, again, that Stephanie has an
intellectual obligation to read books on Operation Valkyrie, but that
she falsely believes that she cannot get to the library because of recent
floods in her city. However imagine that, this time, she is an intellec-
tually lazy person and that even if she had not been ignorant that she
could reach the library, she would have watched a movie rather than
working on her presentation. One might think that in a case like this,
Stephanie’s ignorance does not excuse her, and that we should, there-
fore, add the following counterfactual clause to (8): “and if S had not
been ignorant of p or excused in some other way, then S would have
met O.”13
We should tread carefully here, though. As I said, we should assume
that the case under consideration is one in which Stephanie is blame-
lessly disbelievingly or deeply ignorant that she can go to the library.
Let us consider deep ignorance first. Well, if Stephanie is blameless for
not even considering whether she can go to the library, then it is hard
to see how she is to be blamed for not doing so. For it seems that going
to the library is an action that she can perform only if at some point she
considers going to the library. Blameless deep ignorance of her ability
to go to the library, then, provides an excuse for not doing so and for
her ensuing ignorance on Operation Valkyrie. And something similar
seems to apply to blameless deep ignorance of one’s having an obliga-
tion and blameless deep lack of foresight regarding an obligation. If
one is blameless for not even considering whether one has an obliga-
tion or whether an action or series of actions sufficiently increases the
likelihood of violating a future obligation due to force or ignorance,
then it is hard to see how one could properly be blamed for violating
that obligation.
13. Here one may be inspired by Donagan (1977, 128–130), Guerrero (2007, 63–6 4),
Rivera-L ópez (2006, 135), Zimmerman (1997, 424).
180
181
What about blameless disbelieving ignorance? Normally, such

disbelief would play a role in one’s decision-m aking process. For
instance, if Stephanie considers whether she should go to the library
and finds herself believing that she cannot, then that belief of hers
will normally be a sufficient reason for her to decide not to go to the
library. But maybe there are situations in which disbelieving igno-
rance plays no role. Maybe Stephanie is so lazy that immediately
upon considering whether to go to the library, she decides not to do
it because of all the work that she associates with going to the library.
Because of her laziness, she immediately takes a decision on the issue
without activating her dormant belief that she cannot even go to the
library. Similarly, it seems possible that one is disbelievingly ignorant
of one’s obligation or disbelievingly lacks foresight, but that such dis-
believing ignorance plays no role in one’s decision-m aking process.
In such cases, one would have acted in the exact same way if one had
not been ignorant.
I think that this is an attractive view. Nonetheless, it seems to me that
the objection is unconvincing, for the intuition that we cannot properly
blame someone for violating an obligation if she is ignorant seems as
strong as ever. If Stephanie is ignorant that she can go to the library,
then how could we properly blame her for not going to the library? And
if one blamelessly falsely disbelieves that one has an obligation to φ,
then how can one properly be blamed for failing to φ? True, it seems
that Stephanie is blameworthy if she would still fail to go to the library
if she truly believed that she could. But we can explain this intuition
by acknowledging that there are others things in the neighborhood, but
nonetheless different from not going to the library and different from
the ensuing ignorance, that she is blameworthy for in such a case. Here,
we can think of her being lazy, her being such that even if she had not
been ignorant she would not have gone to the library, and her not con-
sidering whether she can go to the library.14 And mutatis mutandis the
same thing could be said about the other kinds of blameless disbelieving
14. For a somewhat similar approach, see Houlgate (1968, 112–113), Rosen (2008, 598n),
Zimmerman (1988, 79, 139n).
181
182
ignorance. I conclude that our analysis of ignorance as a doxastic excuse

need not include a clause about S’s motivation.
We can put the results of this and the previous sections into the fol-
lowing table, where “F” stands for “full excuse,” “P” for “partial excuse,”
and “N” for “no excuse”—where, again, I assume that the ignorance under
consideration is blameless, that is, that one is blameless for being ignorant.
Table 5.1 ignorance as a doxastic excuse
Which Kind of Ignorance?
Disbelieving Deep Suspending Unwarranted

Ignorance Ignorance Ignorance Ignorance
Ignorance of one’s F F P/N N

obligation
Ignorance of one’s F F P/N N
ability to meet one’s
obligation
Ignorance of What?
Ignorance of how F F P/N N

to meet one’s
obligation
Ignorance regard- F F P/N N
ing one’s future
obligation
Ignorance of any N N N N
other kind of
proposition
5.4 IGNORANCE AS A NARROW

AND A BROAD DOX ASTIC EXCUSE
So far, I have argued that there are four kinds of disbelieving ignorance
and deep ignorance that provide a full doxastic excuse, as long as one is
182
183
blameless for violating the relevant intellectual obligation(s). Thus, they

remove blameworthiness for one’s belief despite the fact that that belief
issues from the violation of an intellectual obligation. And they count
as a doxastic excuse, because in the absence of such ignorance and any
other excusing condition, one would be blameworthy for the ensuing
beliefs. But I have not yet distinguished between ignorance as a narrow
and a broad excuse, that is, as an excuse that blocks respectively blame-
worthiness and all responsibility. I argued in c hapter 4 that one can
sometimes believe responsibly despite being excused for the violation of
the intellectual obligations from which that belief issued, for instance,
if one was forced to violate the intellectual obligations in question, but
one nonetheless had the opportunity to influence one’s belief on other
occasions, that is, if one could have believed otherwise.
Let us start with ignorance as a broad doxastic excuse, that is, as an
excuse that removes all responsibility for one’s belief. Certain humans
are so profoundly ignorant about normative propositions and often also
about the circumstances they are in that in virtue of their ignorance, we
hold them responsible neither for their actions and omissions nor for the
consequences of their actions and omissions, such as their beliefs. Here
we can think, for instance, of little children and Alzheimer’s patients.
These people are so ignorant about life that we do not consider them
the proper object of any normative appraisal for their beliefs, whether it
be blame, praise, or neutral appraisal, or, in fact, for anything else. They
have no idea that they can influence what they believe and that on cer-
tain occasions humans ought to do so.
Now, there are at least two different ways to think of these people.
On the one hand, one might think that these people have intellectual
obligations, but that they are broadly excused for violating them all the
time. Thus, they have obligations to influence their beliefs, but they are
never the proper object of normative appraisal for meeting or violating
those obligations. On the other hand, one might think that they do not
even have such intellectual obligations: given that they are profoundly
ignorant, even about what intellectual obligations are, they never have
them. I prefer the second view. As I said before, to be excused should be
an exceptional rather than the normal situation. If the first view were
correct, little children and Alzheimer’s patients would be excused all
183
184
the time for their beliefs, for violating intellectual obligations, and for
violating all sorts of other obligations. And that seems simply false.
Applying this to the doxastic realm, it seems that little children are not
excused, but rather not even the proper object of normative appraisal for
their beliefs. They do not believe responsibly or irresponsibly. And that
seems to be the case because they do not have any obligations, espe-
cially not any intellectual obligations. We can evaluate their beliefs in
all sorts of ways—as true or false, as fitting or not fitting the evidence,
and so forth—but, as we saw in chapter 1, we should clearly distinguish
such evaluative from normative judgments on people’s doxastic attitudes.
We can take this observation into account by adding to our analysis of
responsible belief that the person in question is subject to intellectual
obligations.
Does ignorance ever count as a broad excuse, then, for the violation
of an intellectual obligation? It is hard to think of such cases. For when
I am ignorant of something I am usually still responsible for what I do,
given how I act on the beliefs that I have. Only if I am completely and pro-
foundly ignorant about my circumstances and about normative issues
can I not be held responsible for whether or not I violate an intellectual
obligation. But then it seems that either I do not have any intellectual
obligations at all (as we saw above) or that I am forced to act as I do. This
means that to the extent that ignorance provides a doxastic excuse—
and, above, we have seen several cases of it—it virtually always counts
as a narrow excuse, that is, as an excuse that removes only blameworthi-
ness, so that one still believes responsibly. This means that ignorance as a
doxastic excuse differs from force as a doxastic excuse in that ignorance
virtually always counts as a narrow excuse, whereas force, as we saw
in chapter 4, sometimes counts as a narrow excuse and sometimes as a
broad excuse.
Given what I have argued so far, our full-blown analysis of respon-
sible belief now looks as follows:
S responsibly believes that p iff (i) S believes that p, (ii) S has

(9)
certain intellectual obligations, (iii) there is some belief-
influencing action or series of actions A that S could have
performed such that if S had performed A, S would not have
184
185
believed that p, (iv) S has not violated any all-things-considered

intellectual obligation such that if S had met that obligation,
certain belief-influencing factors would have changed in such
a way that S would not have believed that p, and (v) S’s belief
that p is not about that intellectual obligation or those belief-
influencing factors, or (vi) S has violated such an obligation,
but is narrowly excused for that by force or ignorance.15
Whereas conditions (i)–(iii) guarantee that one is responsible for

believing that p, conditions (iv)–(vi) guarantee that one’s belief that p
is blameless.
5.5 B
L AMEWORTHY BELIEF
AND THE REGRESS PROBLEM
So far, I have assumed that the ignorance under consideration is blame-

less. As we saw, such ignorance, as long as it is ignorance of the right
proposition, excuses one for the violation of an intellectual obligation
and for the ensuing doxastic attitudes. But, of course, ignorance can be
blameworthy. Sometimes one should have known better. Or, more pre-
cisely, sometimes one would have had a different doxastic attitude if one
had met one’s intellectual obligations. Now, can blameworthy ignorance
also provide a doxastic excuse?16 It seems that there are at least three
views that one might take here: one might think that blameworthy igno-
rance provides a full doxastic excuse,17 a partial doxastic excuse,18 or no
doxastic excuse at all.19
15. A gain, the analysans of (9) should be read as: (i) ∧ (ii) ∧ (iii) ∧ [[(iv) ∧ (v)] ∨ (vi)].
16. A ristotle already distinguished between ignorance for which one is responsible and
ignorance for which one is not. See Aristotle (2003, 145–147; NE III.v.7–8).
17. See, for instance, Ross (1939, 163–164), Smith (2016). Smith (1983, 548–551) also dis-
tinguishes between these three views.
18. For this view, see Beardsley (1979, 578), Burrington (1999, 516–517), Joyce (1914, 404).
19. According to Rivera-L ópez (2006, 134), this is more or less the standard position. For
this position, see Aristotle (2003, 145–147; NE III.v.7–9), Kornblith (1983, 35–36),
Murray (1914, 104), Wolterstorff (2010a, 106), Zimmerman (2008, 175).
185
186
In order to see which view is correct, consider the following exam-

ple. As a student Eleonore applies to the history department and gets
accepted. Eleonore knows that among her obligations is the obliga-
tion to read carefully the department manual, which she can pick up
from the secretary. However, she is lazy and fails to read the manual.
The manual contains the information that each student has to write, in
addition to the term papers, another paper on any topic she likes and
hand it in before the end of the year. At the end of the final term, then,
Eleonore falsely believes that she has met all her requirements and goes
on vacation.
Imagine that when she returns from vacation, she has a meeting with
her supervisor. He asks her why she did not hand in the paper on time.
Eleonore pleads ignorance. Her advisor then asks her why she did not
read the manual, in which it is clearly stated that she has such an obli-
gation. Eleonore apologizes and acknowledges that she failed to read
the manual because she was not sure whether it might contain relevant
information for her studies and was too lazy to check. Now Eleonore’s
advisor is fully informed about why Eleonore did not hand in the paper
on time. Is it appropriate for him to blame her for falsely believing that
she had met all her obligations? It seems clear that it is. For if Eleonore
had met her intellectual obligation to read the student manual, she
would not have been ignorant of her obligation to write the paper and she
would not have falsely believed that she had met her obligations. It thus
seems that such blameworthy ignorance, like blameworthy force, does
not provide a full excuse. However, it does seem that it provides a partial
excuse. After all, Eleonore did not intentionally violate her obligation
to hand in the paper on time. If she had done so intentionally, then, it
seems she would have been even more blameworthy. Hence, blamewor-
thy ignorance sometimes provides a partial rather than a full excuse.
Now, let us consider a slightly different scenario. In this scenario,
Eleonore knows that the manual might contain information about term
papers and so strongly dislikes term papers that she purposely avoids
reading the manual. Then, it seems, her ignorance of her obligation
to write her term paper does not excuse her at all, not even partially.
Someone who intentionally remains ignorant about her obligations or
her ability to meet her obligations seems in no way excused. Intentional
186
187
ignorance provides not even a partial excuse. Hence, sometimes blame-

worthy ignorance provides no excuse and sometimes it provides a par-
tial excuse.
The most interesting question is, of course, whether blameworthy
ignorance ever provides a full excuse. I can think of only one candi-
date here, namely scenarios in which one’s ignorance is that of lacking
foresight. Thus, one can at some time t violate an obligation O due to
ignorance of, say, one’s ability to meet O, where that ignorance is due
to the violation of another obligation O* at some earlier time t*, where
at t* one could not foresee that by violating O* at t* one would come to
be ignorant of one’s ability to meet one’s obligation O at t. Let me give
an example to illustrate this point. Imagine that I have an obligation to
prepare for my biology exam. However, I am lazy and violate that obliga-
tion. If I had met my obligation, I would have come to know how photo-
synthesis works. A week later, I happen to be in a bar where a psychopath
threatens to shoot someone if nobody can explain to him how photosyn-
thesis works. I cannot meet my obligation to tell him how photosynthe-
sis works, because I am ignorant. And my ignorance is blameworthy, for
if I had met my obligation to prepare for my biology exam, I would not
have been ignorant. Since I could not at all foresee that I would need to
know how photosynthesis works in order to save someone’s life, it seems
that my ignorance of photosynthesis excuses me, even if that ignorance
is blameworthy. And, mutatis mutandis, the same would seem to apply
to the violation of intellectual obligations (rather than the obligation to
save someone’s life) and the ensuing beliefs.
However, in virtue of precisely what am I excused in such cases?
First, it is not my ignorance at time t at which I violate my obligation to
save a life. For remember that some state of affairs excuses one only if it is
a sufficient condition for being blameless. But being culpably ignorant at t
is not a sufficient condition for being excused for violating my obligation
at t. Whether or not I am excused at that time at least partially depends
on whether I could foresee that that ignorance would matter in this situ-
ation and that this situation was sufficiently likely to arise. Second, it
seems that my blameless lack of foresight (an instance of ignorance) at
t* at which I violated my obligation to prepare for my biology exam is a
sufficient condition. For, if I lack such foresight at t*, then it is hard to
187
188
see how I could be blamed for violating my obligation at t. That I am

ignorant at t* entails that I am ignorant of my obligation at t, my ability
to meet that obligation at t, or some such thing, for, as we saw, one can
only be ignorant that p if p is true. Thus, I am fully excused by my lack
of foresight at t* and I am not fully excused by my ignorance at t. The
ignorance constituted by my lack of foresight at t*, though, was blame-
less rather than blameworthy. This suggests that only blameless igno-
rance provides a full excuse. It follows that the first view on the extent to
which blameworthy ignorance excuses is false: blameworthy ignorance
does not provide a full excuse. However, strictly speaking, the second
and third views are also false, for blameworthy ignorance sometimes
provides a partial excuse and sometimes it provides no excuse at all.
Proposition (9), my analysis of ignorance as a doxastic excuse, should be
understood accordingly.
According to some philosophers, though, the thesis that ignorance
is blameworthy only if one could and should have avoided it by meeting
one’s intellectual obligations implies that we are on the road to an infinite
regress. This argument, if sound, constitutes a major threat to my account
of ignorance as a doxastic excuse, for my account relies on the truth of
that thesis, so I propose we consider the argument in more detail.20
In order to see how the argument runs, let us return to one of the
examples that I sketched in the Introduction to this book. Sir Coningham,
an officer in the British Air Force, orders the attack on three German ships
that, unbeknownst to him, are filled with some ten thousand camp survi-
vors, many of whom die as a result of that. Whether we can at least partially
blame Coningham for this seems to depend on whether he should have
known better, that is, on whether or not his ignorance was blameworthy. If
it was not, then it seems unfair to blame him, but if he should have known
better, then it seems that he is at least partly blameworthy for the trag-
edy. When, then, would his ignorance be blameworthy? If what I argued
is correct, then his ignorance is blameworthy only if at some earlier time
he could and should have found out that there were prisoners on board.
However, here is the problem. Imagine that Coningham indeed failed
to investigate the matter sufficiently carefully. Then, we may assume that
20. My assessment of the argument is based on what I argue in Peels (2011b).
188
189
at that earlier time he (falsely) believed that he need not gather any addi-
tional evidence. But if he falsely believed that, he was ignorant. If he was
blamelessly ignorant, it seems unfair to blame him for acting as he did. If
he was blameworthily ignorant, then presumably he performed a blame-
worthy action in the past that resulted in that blameworthy ignorance.
But we can say the exact same thing about that prior blameworthy action,
and about a blameworthy action prior to that, and so on. Hence, a vicious
regress threatens. Clearly, the same trouble can be raised for much more
mundane situations in which someone acts from blameworthy ignorance.
There is some reason to think that the point even extends to certain
cases in which someone clearly displays evil intentions. Joseph Stalin,
presumably, thought that it was a good thing to order the massacre of
thousands of Polish officers in the Katyn forest. Since that belief was
false, he was ignorant of its wrongness. If he should have known better,
then there must have been some prior blameworthy act from which his
ignorance—consisting of his false belief—issued. And so on. The prob-
lem seems to arise for any situation in which a person acts in ignorance,
but in which she nevertheless seems blameworthy for what she does. In
all those cases (or, at least, in too many of them), it seems that, ultimately,
we cannot explain why that person’s ignorance is blameworthy. The only
exceptions are cases of akrasia, scenarios in which a person does some-
thing or fails to do something despite occurrently believing that doing
so (not doing so) is wrong.21 Cases of clear-eyed akrasia, however, are
rare: it seems that what we believe and what we do hardly ever come
apart in this way. It follows that everybody is hardly ever blameworthy
for what he does or fails to do and for what he believes or fails to believe.
This argument has been presented and discussed by James
Montmarquet, Michael Zimmerman, and William FitzPatrick.22 Before
21. Th is is how Zimmerman understands akrasia and how it is usually understood; see, for
instance, Audi (1979, 177), Mele (1987, 109). Below, I return to how we should under-
stand akrasia.
22. See FitzPatrick (2008, 589–6 01), Montmarquet (1993, vii–v iii, 7–9, 45–4 8; 1995, 41–
43; 1999, 845), Zimmerman (1997, 410–421; 2008, 173–193). Rosen (2004, 295–311)
offers an argument similar to the argument discussed here, but the conclusion of his
argument is that we can hardly ever know when someone has performed a culpable
action, since we cannot ascribe akrasia with any confidence.
189
190
considering how it can be met, let me try to be a bit more precise about
its structure. I think it can be rendered slightly more formally as follows:
(10) If S is blameworthy for some action A, then she does A either

from akrasia or from ignorance. [prem.]
(11) Cases of akrasia are highly exceptional. [prem.]
(12) If S is blameworthy for some action A, then in nearly all cases,
S does A from ignorance. [from (10) and (11)]
(13) S is blameworthy for some action A done from ignorance
only if S is blameworthy for that ignorance. [prem.]
she performed A from blameworthy ignorance. [from (12)
and (13)]
(15) S is blameworthy for her ignorance only if she is blamewor-
thy for some past action B from which that ignorance issued.
[prem.]
she is blameworthy for some other blameworthy action B
that preceded A. [from (14) and (15)]
(17) If S is blameworthy for some action A, then in nearly all cases
she is blameworthy for an infinitely long chain of blamewor-
thy actions each of which precedes the other. [from (16)]
(18) There are no infinitely long chains of blameworthy actions
each of which precedes the other. [prem.]
(19) For any person S and action A, S is hardly ever blameworthy
for performing A. [from (17) and (18)]
However, if we are hardly ever blameworthy for an action, then we are

hardly ever blameworthy for belief-influencing actions and, therefore,
hardly ever blameworthy for our beliefs. These conclusions are deeply
disturbing. It is, therefore, important to find a satisfactory way of deal-
ing with the argument. I know of two responses to this argument.
According to Montmarquet and FitzPatrick, we should reject prem-
ise (15): not every instance of blameworthy ignorance is to be explained
by a prior blameworthy action from which that ignorance issues. There
is such a thing as fundamental and underived culpability for ignorance,
190
191
namely if it issues from insufficient regard of truth and evidence.

Insufficiently regarding truth and evidence is not to be considered as
an action, but as a modality of the belief-forming process, as a way of
believing.23
Perhaps Montmarquet and FitzPatrick are right that insufficient
regard of truth and evidence does not count as an action, but it seems to
me that their suggestion will not do. This is because people can clearly
be blameless for insufficient regard of truth and evidence. They may be
brainwashed or raised in a society that cares very little about truth and
evidence. Insufficient regard of truth and evidence cannot be intrin-
sically blameworthy. FitzPatrick seems to be aware of this worry, for
he adds that one is blameworthy for exercising some intellectual vice
only if one could reasonably have been expected to do better. 24 This,
however, does not solve the problem. For we have not been given an
answer to the question of why or in virtue of what one could reason-
ably have been expected to do better. Even more importantly, assuming
that insufficient regard of truth and evidence is always wrong, it still
seems that if people act with insufficient regard of truth and evidence,
they act either from akrasia, that is, against their better judgment, or
from ignorance that acting with insufficient regard of truth and evi-
dence is wrong. Hence, this approach relocates the problem rather than
solving it.
Zimmerman suggests a different approach to the regress problem.
On his view, the argument surprisingly shows that we are far less often
blameworthy for our beliefs and actions than we might initially think.
Thus, we should blame each other only in cases in which an act is done
either from akrasia or from ignorance that directly or indirectly issued
from an act done from akrasia. 25 I take it that this option is clearly unat-
tractive: we should be suspicious of philosophical arguments that call
for a substantial revision of one of our socially most important and
deeply entrenched practices. We should take this route only if each of
the premises of the argument is at least as convincing as the thesis that
23. See Montmarquet (1993, 41–43, 45–47; 1999, 843–8 45).

24. See FitzPatrick (2008, 606–610).
25. See Zimmerman (1997, 421–426; 2016).
191
192
people are frequently blameworthy for their ignorance and for actions
performed from ignorance.
Given that Montmarquet’s and FitzPatrick’s response to the argu-
ment fails, is there a way to avoid the disturbing conclusion of the argu-
ment and, thus, to avoid Zimmerman’s strongly deflationary proposal?
As to (13)—the premise that one is blameworthy for some action A
done from ignorance only if one is blameworthy for that ignorance—
ever since Aristotle philosophers have defended this premise and
I have defended it above. 26 In c hapters 2 and 3, I have argued for (15),
the premise that one is blameworthy for one’s ignorance only if one is
blameworthy for some past action from which that ignorance issued.
And it is implausible to deny (18), the premise that there are no infi-
nitely long chains of blameworthy actions. Even if we could live forever,
none of us has lived forever and, hence, there are no such things as infi-
nitely long chains of blameworthy actions. Hence, (13), (15), and (18)
are unproblematic. It seems to me, though, that there are at least three
other problems with the argument: (10) is false, (11) is false, and (16)
does not entail (17).
Starting with the latter, let us assume that (16) is true, that is, that
if S is blameworthy for A, then in nearly all cases she is blameworthy
for another blameworthy act B that preceded A. Imagine, for instance,
that if S is blameworthy for A, then the probability that A is preceded by
another blameworthy act B is .95. The probability that B is preceded by
another blameworthy act C is, of course, also .95, so that the probability
that A is preceded by two blameworthy acts B and C is .95 x. 95 = .9025.
Clearly, the longer the series of actions, the more likely it is that an
action performed from akrasia occurs at some point in the series. It may
be unlikely that a particular act is done from akrasia, but it is not unlikely
that some act or other in a longer chain of actions is done from akrasia.
But if (17) is false, at most a significantly weaker conclusion than (19)
can be established.
Second, (11) says that actions done from akrasia are rare. This may
be true for actions in general (even though I doubt even that). It does not
26. See Aristotle (2003, 145–147; NE III.v.7–8), Rosen (2003, 64–6 6), Zimmerman
(1997, 412).
192
193
follow that acting from akrasia is rare for all particular kinds of actions.
More specifically, it does not seem that exceptional when it comes to
belief-influencing actions, actions such as gathering evidence and work-
ing on our intellectual virtues and vices. Students frequently violate an
obligation to prepare for an exam, policemen often violate an obligation
to gather all the relevant evidence, and many narrow-m inded people fail
to become more open-m inded in the course of their lives. They do so,
knowing perfectly well what their intellectual obligation is. I think there
is a particular reason for this. For most morally important actions, such as
sentencing a criminal or buying an expensive car, we have a pretty clear
idea of what the main consequences are. We easily foresee that the crim-
inal ends up in jail and that the amount of money in my bank account is
significantly reduced.
With belief-influencing actions, however, things are different. We
usually do not foresee which particular beliefs will issue from them and
which actions we will perform on the basis of those beliefs. That makes it
much easier to violate such obligations, even if we believe that we should
not do so (which makes them cases of akrasia). We know that we run a
certain risk, but since we do not foresee the precise consequences, we
are more likely to succumb to the temptation of violating the obligation
in question, despite our belief that we should not. Zimmerman leaves
some room for the idea that what kind of consequences is involved
makes a difference to how easy or difficult it is to act from akrasia. 27 He
fails to notice the important point, though, that we fairly easily act from
akrasia when it comes to intellectual obligations, since in such cases we
do not foresee which particular doxastic attitudes we will come to hold
as a result of violating or meeting them. This point is crucial, for culpable
ignorance will be due to the violation of such obligations.
Finally and most importantly, (10) is false: it is not the case that if
S is blameworthy for some action A, then S does A either from akrasia
or from ignorance. For, as I argued in c hapter 1, we not only have occur-
rent beliefs, but also dormant and tacit beliefs. And it seems that some-
one with dormant or tacit beliefs about her circumstances or about
the normative status of an action can be blamed for not activating
27. S ee Zimmerman (2008, 190).
193
194
those beliefs. 28 Imagine that I am babysitting and that, before leav-

ing, my friend tells me that her daughter Sarah is seriously allergic to
milk. If, an hour later, I completely forget about that and give Sarah a
glass of milk, I am clearly blameworthy for that, simply because I had
the dormant, but unactivated (non-occurrent) belief that I should
not give her any milk. I could and should have thought about it or
remembered it.
I clearly did not act from akrasia: it is not that I was fully aware of the
danger but succumbed to the temptation by weakness of will. Nor was
I ignorant that Sarah is allergic to milk or ignorant that I should not give
Sarah any milk. Surely, I knew (and, thus, truly believed) that Sarah is
allergic to milk. That dormant belief, perhaps in combination with the
absence of excusing circumstances, is sufficient to properly blame me
for giving Sarah milk. Mutatis mutandis, the same applies to many other
situations, such as violating one’s promise by forgetting to take some
groceries along on the way home or forgetting one’s friend’s birthday. In
the former case, one fails to activate one’s dormant belief that one should
bring some groceries along, and in the latter case, one fails to activate
one’s tacit belief that today is one’s friend’s birthday.29 Not all situations
in which one performs a blameworthy action are situations of clear-eyed
akrasia or ignorance. The argument overlooks dormant and tacit beliefs
and thereby fails to acknowledge the normative weight that such beliefs
have: we can be blameworthy in virtue of not doing what we dormantly
or tacitly believe we should do.
What this suggests is that we are sometimes blameworthy for violat-
ing an intellectual obligation in virtue of our dormant and tacit beliefs.
This is not to deny that we may be excused for the violation of such an
obligation. Only, such an excuse will be different from ignorance; it will
be such a thing as compulsion or acting under stressful circumstances.
28. The false premise implicit in (10), then, is that only occurrent beliefs can count as
reasons for which one acts or should act. For this premise, see Rosen (2004, 307),
Zimmerman (1997, 421–422; 2008, 190–191). For other counterexamples to this prin-
ciple, see Frankish (2009, 86), Wolf (1990, 88). Smith (2011a, 116–117) gives more
examples, although she does not consider all cases that she mentions as ones in which
one is blameworthy.
29. Th
us also Smith (2005, 236), Weatherson (2008, 552), Wolf (1980, 164).
194
195
The above examples show how natural and intuitive this position
is: unless special conditions hold, we would all blame me for giving milk
to Sarah and we think we would properly do so because of my dormant,
unactivated belief that Sarah is allergic to milk.
One could, of course, bite the bullet and stress that only occur-
rent beliefs give rise to obligations and, unless certain excuses hold, to
blameworthiness. Since this claim clearly contradicts our intuitions in
the examples that I just gave, it seems that we should not accept it with-
out an argument of some kind. Michael Zimmerman provides such an
argument. According to Zimmerman, only occurrent beliefs can render
one blameworthy for the violation of an obligation because only occur-
rent beliefs play a role in the reasons for which one performs an action.
And unless something plays a role in the reasons for which one performs
an action one cannot in virtue of that thing be blamed for performing
the action. In one respect Zimmerman qualifies his position, though.
He grants that one can be blameworthy in virtue of one’s dormant and
tacit beliefs in cases of routine or habitual actions, cases in which one
performs actions for reasons that one does not consider at the time at
which one performs them. 30
However, there is good reason to reject Zimmerman’s claim.
Imagine that I am teaching a class on evolutionary theory and that, in
the course of my lecture, I tell the students that Darwin’s On the Origin
of Species was published in 1859. This is something I tell them because
I believe it, because I take it to be something that the students ought
to know, because I believe that there are students in the room, because
I believe that I can transfer knowledge to my students by telling them
something, and so forth. I tell them this because I believe these things in
the sense that I would not tell them this if I lacked any of these beliefs.
But, clearly, I need not consciously consider all the relevant propositions
in order for it to be true that my telling the students that Darwin’s On
the Origin of Species was published in 1859 is based on those beliefs. In
fact, it seems I could not give the lecture if I were occurrently to consider
each of these many propositions simultaneously. Thus, in a case like this,
30. S ee Zimmerman (2008, 190–191).
195
196
my dormant and tacit beliefs play a crucial role in my reasons for per-
forming this particular action. 31
Now, I do not think that there is any natural sense in which the
action of my saying that On the Origin of Species was published in 1859
is a routine or habitual action. It may be the very first time in my life
that I utter this sentence, but, as the example shows, I may still do it for
reasons that at that time I do not consciously consider. If Zimmerman
were to insist that this action is habitual, then it seems that we would
have to say that many of our actions, if not the majority, are habitual.
Since Zimmerman allows that in the case of habitual actions we can be
blameworthy in virtue of our dormant and tacit beliefs, it would follow
that we could still be blameworthy for many of our actions.
In summary, we have traced two ways of acquiring or maintaining
blameworthy ignorance. First, one can perform the belief-influencing
action from which one’s ignorance issued from clear-eyed akrasia. As
we saw, such akrasia is not as rare in the case of evidence gathering and
working on our intellectual virtues and vices as it is when it comes to
other actions. Second, one can perform an action resulting in ignorance
with unactivated dormant or tacit beliefs about one’s circumstances
or the normative status of that action that one should have activated.
It seems highly likely that for a substantial number of our belief-
influencing actions, one or both of these possibilities are realized at that
moment or at some earlier moment to which those actions are relevantly
related. I conclude that the argument fails to establish that we are hardly
ever blameworthy for our belief-i nfluencing actions and for the ensuing
beliefs.
It is worth noting that the following major insight follows from what
I have argued in this section and sections 5.3 and 5.4: responsible belief
is, in a specific sense of the word, radically subjective. For whether one
believes responsibly crucially depends on how one has acted given one’s
beliefs. Of course, one’s beliefs may be blameworthy, but I argued that if
that is the case, then that is so in virtue of an earlier occasion on which
one acted blameworthily given one’s beliefs.
31. A udi (1986, 515, 521) and Stump (1993, 254) also advocate the view that beliefs need
not be occurrent in order to guide action.
196
197
Let me be clear that I do not mean that whether we believe responsi-
bly is subjective in the sense that it entirely depends on people’s perspec-
tive on things. If what I argued in c hapter 4 is correct, then whether or
not we believe responsibly depends at least partly on whether or not the
world provides us with an opportunity to make a difference to what we
believe. What I have argued in this chapter, however, is that given that we
are responsible for a belief, whether we believe responsibly or not crucially
depends on how we act, not given how the world in fact is, but given how
we believe the world to be and how we believe the world ought to be.
Of course, we are sometimes blameworthy for believing the world to
be a particular way and, at least sometimes, we are blameworthy for act-
ing on those blameworthy beliefs. However, if what I have argued is cor-
rect, then one will be blameworthy in such a case only if there was some
earlier point in time at which one could and should have performed a
belief-influencing action or series of actions such that if one had per-
formed it or them, one would not have held those beliefs. And one will
be blameworthy only if at some time one violated an intellectual obliga-
tion despite having the—occurrent, dormant, or tacit—belief that one
had that all-t hings-considered intellectual obligation.
This is not to deny that in blaming someone people hardly ever
explicitly ask themselves whether that person at some point violated a
subjective obligation in coming to hold or in maintaining that belief.
But there are good explanations for that. First, many people lack the
cognitive apparatus to think through these issues, even in a particular
case. Second, our normative attitudes are often quite irrational: we are
strongly inclined to blame someone for an objectively bad act, even if we
have not asked ourselves whether that person was truly blameworthy
for it. Third, we often assume that someone at some point violated cer-
tain subjective obligations in coming to hold or maintain some belief.
In blaming the neo-Nazi, I assume that his racist beliefs are not or have
not always been completely beyond his influence. I assume that there is
something that he could and should have done such that if he had done
it, he would not have held such heinous beliefs. 32
32. Th
us also Plantinga (1983, 35–37).
197
198
5.6 C ONCLUSION
I have argued that ignorance sometimes provides a doxastic excuse. It

seems that many doxastic excuses are varieties of excusing ignorance. If
we say that some person violated an intellectual obligation from mistake
or aberration, or if she did not do so on purpose, did not mean to do it, or
did not intend to do it, what we seem to have in mind is that that person
was in some way ignorant and is excused for the ensuing belief in virtue
of that ignorance.
Let me summarize how ignorance bears on doxastic responsibility
on the account that I have defended in this chapter. Some philosophers
have argued that ignorance is absence of true belief. Even if that view,
which I call the New View, is false, that is no problem, for true belief
that falls short of knowledge and that would count as ignorance on the
Standard View, on which ignorance is lack of knowledge, never provides
a doxastic excuse. Moreover, I argued that only disbelieving and deep
ignorance provide a full doxastic excuse. Suspending ignorance pro-
vides either only a partial doxastic excuse or no doxastic excuse at all.
Ignorance provides an excuse for some belief if it is ignorance of
one’s having the all-t hings-considered intellectual obligations to which
that belief is relevantly related, of one’s ability to meet those obligations,
of how to meet those obligations, or lack of foresight regarding those
obligations. Only blameless ignorance provides a full doxastic excuse.
No infinite regress follows from these claims, for it is relatively easy to
act from akrasia when it comes to violating our intellectual obligations
and we can be blameworthy for violating our intellectual obligations in
virtue of our dormant and tacit beliefs. It follows from this account of
ignorance as a doxastic excuse that whether or not one believes respon-
sibly is a radically subjective matter in the sense that when one is respon-
sible for a belief, whether or not one believes responsibly depends on
whether one has acted in accordance or contrary to one’s beliefs about
one’s intellectual obligations.
198
199
[6]
RESPONSIBLE BELIEF IS
COMPATIBLE WITH DOXASTIC LUCK
6.1 I NTRODUCTION
As some philosophers, such as Jonathan Bennett, acknowledge, luck,

like force and ignorance, sometimes provides an excuse.1 Now, one
might think that luck provides an excuse in addition to force and igno-
rance, for it seems that luck has to do with chance, coincidence, or improb-
ability rather than with being forced or being ignorant.
Luck has been extensively discussed in both ethics and epistemol-
ogy. In ethics, however, philosophers have focused on the way that luck
affects actions and their non-doxastic consequences, such as someone’s
death, rather than beliefs. In epistemology, philosophers have zeroed in
on the way that luck affects whether or not someone has knowledge. The
purpose of this chapter is to find out whether luck provides an excuse
for some of our beliefs and, if so, whether it is a separate excuse from
force and ignorance.2 In doing so, I shall refer to both the debate about
moral luck and that about the relation between knowledge and luck.
Since moral luck has to do with responsibility and the issue of knowledge
and luck has to do with belief, both debates are relevant to my account of
responsible belief.
The plan of this chapter is as follows. First, I provide an analysis
of luck by arguing, roughly, that some event is lucky for some person
1. See Bennett (1980, 16).

2. One of the very few to touch the topic is Nottelmann (2007, 199).
199
200
if and only if that person lacks control over it, it is significant to that
person, and that event could easily have failed to obtain (§6.2). This
analysis will turn out to be helpful both in seeing whether or not luck
can provide a doxastic excuse in addition to force and ignorance, and
in solving what I call the “problem of doxastic luck,” a problem that
I discuss at the end of this chapter. Subsequently, I distinguish four
varieties of doxastic luck and argue that to the extent that they excuse,
they are reducible to force or ignorance (§6.3). Nonetheless, one of
these varieties of doxastic luck—t he one that does not excuse—shows
that my account of responsible belief is defective. I propose a solution
to this problem by arguing that one is blameworthy for the beliefs that
issue from the violation of an intellectual obligation only if, in a sense
to be explained, those beliefs are non-accidentally related to that intel-
lectual obligation (§6.4). Finally, I discuss and propose a solution to
what I call the problem of doxastic luck. Briefly, the problem is that it
seems that luck is incompatible with blameworthiness, that what we
believe is inevitably a matter of luck, and that sometimes we do not
believe responsibly because we believe blameworthily. I defend the
thesis that luck is compatible with responsible belief, and blamewor-
thy belief in particular, by arguing that whether a person is doxasti-
cally blameworthy or not depends on what she does in the actual world
and in nearby possible worlds, but not in possible worlds that are fur-
ther away, since they are immaterial to luck and responsibility in the
actual world (§6.5).
6.2 W HAT IS LUCK?
Before we consider whether luck can be a doxastic excuse, let us ask what
luck is. I take luck to include both good and bad luck. This may sound
somewhat strange. How can good luck provide an excuse? After all, if
I had good luck, something beneficial happened to me, whereas if I am
excused, I did something wrong—otherwise, I argued in c hapter 3,
there would be nothing to be excused for. I answer that if the actualiza-
tion of a state of affairs is overall beneficial for someone, then it might be
a case of good luck for that person. That does not exclude the possibility
200
201
R esponsible B elief I s C ompatible with D oxastic L uck
that as a consequence of that overall good state of affairs, an all-t hings-

considered obligation is violated.
The following example illustrates this. I have to take a biology exam
and the only way to prepare for it is by studying my biology book. Someone
loves biology so much that he steals a biology book. It happens to be my
biology book. However, he realizes that in stealing a biology book he does
something wrong, so he makes up for it by replacing my biology book with
a lump of gold that has the same weight as my (heavy) biology book. It
seems that in this case, good luck—things turn out very well for me—
excuses me for violating my intellectual obligation to prepare for my biol-
ogy exam by reading my biology book and for the ensuing false beliefs and
ignorance. Thus, both good and back luck can count as a doxastic excuse.
One gets easily confused on words and expressions like “lucky,” “lucky
for S,” “unlucky,” and “non-lucky,” so let me be explicit and more or less stipu-
latively define these words and expressions. In doing so, I will use the phrase
“the actualization of a state of affairs Σ” rather than “events” so as to cover a
wide variety of phenomena, including actions and beliefs. When I say that a
state of affairs Σ is actualized, I do not mean that someone actualizes it. The
state of affairs that there is a thunderstorm may be actualized at some time,
but it is not actualized by someone. Now, here are the definitions:
(1) The actualization of a state of affairs Σ is lucky iff the actual-

ization of Σ is due to luck.
(2) The actualization of Σ is lucky for S iff the actualization of Σ is
due to luck and advantageous for S.
(3) The actualization of Σ is unlucky for S iff the actualization of
Σ is due to luck and disadvantageous for S.
(4) The actualization of Σ is non-lucky iff the occurrence of the
actualization of Σ is not due to luck.
Now, let me present what I consider to be a correct account of luck.

I do not claim that this account is original; it is at least similar to other
accounts in the literature, such as that of Duncan Pritchard. 3 The account
3. See Pritchard (2005, 128–133). As we will see, however, it also differs from Pritchard’s
account on certain points.
201
202
is confined to events that are lucky or unlucky for some specific person,
since we are concerned with the doxastic responsibility of individuals.
(5) The actualization of a state of affairs Σ is lucky or unlucky for

S at t iff (i) S lacks intentional control over the actualization
of Σ at t, (ii) the actualization of Σ is significant to S at t, and
(iii) Σ is actualized at t in the actual world, but not in a wide
class of nearby possible worlds in which the relevant initial
conditions for the actualization of Σ are the same as those in
the actual world.4
To take a stock example, imagine that I have won the lottery. I have influ-
ence on winning the lottery, for I could fail to win the lottery by not even
buying a ticket. But I lack control over winning the lottery, for whether
I win it is not up to me—at least, not if the lottery is a fair one. 5 And,
given the large sum of money involved, winning the lottery is significant
to me. Finally, if it is a fair lottery, I could easily have failed to win the
lottery: in a significant number of nearby possible worlds, I do not win
the lottery.
I think it is helpful to contrast this account with rival accounts of
luck. Maybe the philosophers in question intended merely to give a
rough characterization of luck, not a substantial account or analysis.
Nevertheless, treating these characterizations as accounts is helpful for
our purposes, for it shows us what an accurate analysis of luck needs to
take on board in order to be satisfactory. In doing so, we shall see that
the conditions mentioned in the analysans of (5) are individually neces-
sary and jointly sufficient for luck.
4. Does it follow that there is no luck in a deterministic world? That depends on whether or
not one takes it that other possible worlds are close to that world, that is, whether the
laws of nature could easily have been different. If they could, then there is probably luck
in that world, for then things could easily have been otherwise. If not, it follows indeed
that there is no luck in that world. But would we not speak of luck even in a deterministic
world? Maybe we would, but it seems that such talk could easily be explained by our
deep-seated intuition that in many situations things could easily have been otherwise.
5. F or the distinction between control and influence, see c hapter 2, section 2.5.
202
203
Many philosophers, such as Claudia Card and Zimmerman, say or

assume that the actualization of Σ is lucky for S just in case it is beyond
S’s control whether or not Σ obtains.6
The first thing to notice about this account is that it is ambiguous in
that it does not say what kind of control is relevant: is it causal or inten-
tional control? Remember that, as I defined these two kinds of control
in chapter 2 (§2.3), causal control requires merely that it is physically
possible that one brings about some state of affairs, whereas intentional
control also requires that one knows or holds a true belief about how to
bring about the relevant state of affairs. Imagine, for instance, that I have
no idea which button I should press in order to open a box in front of me
and that I arbitrarily choose button 7, which happens to be the right but-
ton. Then, I exercised mere causal control in doing so, but the event of
my opening the box still seems lucky. If, however, I exercise intentional
control and purposely press button 7, knowing that that would open the
box, then the box’s opening is not a matter of luck. Hence, what is rel-
evant for luck is lack of intentional control.
The second thing to notice about these accounts of luck is that the
absence of intentional control is not sufficient for the actualization of Σ’s
being lucky for one. The differential rotation in the upper atmosphere
of Jupiter is beyond my intentional control, but clearly not lucky for me
or anyone else. What this suggests is that the occurrence of Σ is lucky
for some person only if it is of some significance to that person.7 And the
occurrence of a state of affairs Σ is significant to a person if it harms her,
pleases her, irritates her, is valued by her, or some such thing.
According to other philosophers, such as Carolyn Morillo, an event
is lucky just in case it is accidental. 8 What could they have in mind? There
seem to be at least three options. First, it could be that by “accidental”
they mean logically or metaphysically contingent. This, however, is
6. See, for example, Athanassoulis (2005, 20), Card (1990, 199), Greco (1995, 83), Moore
(1990, 301–305), Nussbaum (1993, 76), Richards (1993, 167–169), Zagzebski (1994,
397–413), Zimmerman (1993, 219, 231; 2002, 559n; 2006, 585–590).
7. Coyne (1985, 322) defines luck as the absence of control in conjunction with signifi-
cance. As we shall see below, even this is not sufficient.
8. See Harper (1996, 276), Morillo (1984, 109, 125).
203
204
implausible, for then the vast majority of our actions, such as my drink-
ing a cup of coffee now, would be lucky. Second, they could mean that
an event is lucky just in case its occurrence is due to chance. Some phi-
losophers even explicitly use that word in this context.9 What they mean
by that is, presumably, that the event in question is improbable. But that
seems equally flawed. The probability of my eating a piece of bread today
that contains the exact same number of atoms as the piece of bread that
I ate yesterday is low, extremely low indeed, but not lucky or unlucky
for me. What we should add, of course, is that the event in question has
some significance for me. But even then low probability will not do. If
one is involved in a serious train accident, where one out of two pas-
sengers die, one is lucky to have survived, even though it was not more
probable than not that one would survive. What seems required at most
is that it is at least sufficiently improbable that the event occurs, where
what counts as sufficient is context-dependent.
However, precisely what do we mean by “improbable”? I think that
this notion is best cashed out in terms of what happens in nearby rather
than far away or all possible worlds, where a possible world is closer to
the actual world, roughly, if it is more like the actual world. Something is
sufficiently improbable if it does not happen in a sufficiently large num-
ber of nearby possible worlds. What happens in far away possible worlds
is irrelevant to luck, whether or not it happens in the majority or only a
small percentage of those far away possible worlds.
Let me illustrate this by way of an example. Imagine that one day,
after work, I find my car where I left it in the morning. Am I lucky that it
is still there? Unless special conditions hold, it is clear that I am not: cars
are only rarely stolen. I would be rather unlucky if it were stolen. Maybe
in the vast majority of possible worlds there are no cars, because the
physical universe or the development of life in those worlds looks radi-
cally different. But all that is irrelevant: what matters to whether or not
an event is lucky is what happens in nearby possible worlds in which the
initial conditions at t are identical to or very much the same as those in
the actual world at t. If ten different people intend to steal my car at t, but
9. See Browne (1992, 345–3 46), Harper (1996, 274).
204
205
all of them fail to do so because of accidental events—e.g., getting ill,

the train’s breaking down, etc.—then I would be lucky that my car is not
stolen, for in many nearby possible worlds my car is stolen. Of course,
the notion of “closeness” is vague. But that is actually an advantage of my
view, for there will certainly be boundary cases in which we hesitate to
say whether or not an event is lucky or non-lucky.10
I have argued that absence of control is not sufficient for luck: it is
also necessary that the event does not occur in a substantial number
of nearby possible worlds and that the event is of significance. Is the
absence of control nevertheless necessary for luck? Some philosophers
think that it is not,11 but it seems to me that it is. Imagine that Sam is
strongly inclined to ask Julia to marry him, so that it is highly proba-
ble that he does, that is, that he does so in most of the nearby possible
worlds. Nevertheless, he decides not to propose to her. Would we say
that his not proposing to her was a lucky event for him? I do not mean
to ask whether his not marrying Julia or any of the other consequences
of not proposing to Julia is a lucky event. Those may very well be. But
it seems that Sam’s not proposing to Julia itself is not a matter of luck.
Improbability conjoined with significance may be necessary, but it is not
sufficient for an event’s being due to luck. We also need the absence of
control, or, more specifically, the absence of intentional control.
However, the claim that absence of control is necessary for luck has
been contested. Jennifer Lackey has provided what she considers to be
a counterexample to this claim. Imagine that Ramona, a demolition
worker, is about to press a button that will blow up an old warehouse.
Unbeknownst to anyone, a mouse chewed through the relevant wires.
However, before Ramona tries to blow up the warehouse, Ramona’s
coworker hangs his jacket on a nail in the precise location of the severed
wires. The hanger is made of metal, so that the wires are connected and
so that Ramona can blow up the old warehouse after all. Ramona has
control over blowing up the warehouse, but if she blows up the ware-
house, that is a lucky event.12
10. Th at luck is a matter of what happens in nearby possible worlds is often noted in the
epistemological literature; e.g., Greco (2003, 353–354).
11. S ee Pritchard (2005, 128–133; 2006, 4), Rescher (1995, 211–212).
12. S ee Lackey (2008, 258–259).
205
206
It seems to me that Lackey’s analysis of this scenario is unconvinc-

ing. Maybe Ramona suffered from bad luck when the mouse chewed
through the relevant wires. And, presumably, Ramona was lucky that her
coworker happened to hang his jacket in the precise location of the sev-
ered wires. Both of these states of affairs may well satisfy the three con-
ditions for luck that I distinguished. But it should be clear that once the
jacket has been hung up in the precise location of the severed wires, it is
not a matter of luck that Ramona blows up the warehouse, for, given the
initial conditions of the jacket’s hanging there, she blows up the ware-
house in nearly all or most nearby possible worlds. Her blowing up the
warehouse is under her control, but it is not a case of luck.13 The under-
lying point here is that not everything that is made possible by a lucky
event is itself a lucky event. The fact that humans exist might be due to
lucky events during certain stages of our evolutionary history. It does
not follow that everything we do is a matter of luck. Similarly, if Ramona
is lucky that she is in a situation in which she can blow up the warehouse,
it does not follow that her blowing up the warehouse is a lucky event.
Notice that my analysis of luck is in a sense radically objective:
whether or not an event is lucky for one depends on whether it is beyond
one’s control, of significance to one, and sufficiently likely not to occur,
whether one believes that or not. I think it would be misguided to make
the notion of luck subjective, as some philosophers, such as Andrew Latus,
do.14 Imagine that Evelyn has taken part in a game of Russian roulette.
Unbeknownst to her, the revolver contains no bullets, so that there is
no way she could have killed herself. Of course, from Evelyn’s perspec-
tive, she is lucky to have survived the roulette. But would we, who know
that the revolver contains no bullets, say that she was lucky? Of course
not. We know that she was non-lucky, although Evelyn may have had
good reasons to believe that she was lucky. People may even mistakenly
believe that something has or lacks value for them.15 But if one person or
a group of persons can be ignorant of certain facts about a situation that
13. F or a somewhat similar diagnosis of this scenario, see Levy (2009, 491–493).
14. See Latus (2003, 467–470). This also seems to be the view of Rescher (1993, 146–147).
15. Of course, it may be a subjective matter whether something is significant to one in the
sense that, for instance, stamps can be significant to one because one collects them.
What I deny is that luck should be understood subjectively in the sense that one is lucky
206
207
are relevant for determining whether or not an event is a matter of luck,

it is possible that all of us are ignorant of certain facts about certain situ-
ations, facts that are relevant for determining whether or not an event is
lucky. Thus, even though it is a subjective matter what we consider to be
cases of luck, it is not a subjective matter what are cases of luck. Luck is not
in the eye of the beholder.
6.3 F
OUR VARIETIES OF DOX ASTIC
LUCK AND LUCK AS
A DOX ASTIC EXCUSE
One might think that it obviously follows from (5)—my analysis of

luck—that each case of doxastic luck must be an instance of force or
ignorance or both, and that luck, therefore, does not provide a doxastic
excuse in addition to force and ignorance. Here is the argument. One
lacks intentional control over the occurrence of Σ only if (i) one cannot
actualize Σ, or (ii) one does not know and one holds no true belief about
how to actualize Σ (the disjunction here is inclusive). (i) is an instance of
force and (ii) an instance of ignorance. But that means that every case of
luck is also a case of force or ignorance.
However, the question of whether luck provides a doxastic excuse
in addition to force and ignorance cannot be answered that easily. This
is because if the argument that I just sketched were sound, we would be
excused by force or ignorance for virtually all of our beliefs that issue
from the violation of an intellectual obligation. For I argued in c hapter 2
that virtually all of our beliefs are beyond our intentional control, many
of our beliefs are at least of some significance to us (if only epistemic sig-
nificance), and it seems that many of our beliefs are such that we could
easily have failed to have them, e.g., if we had looked the other way or if
we had not stumbled upon that particular book.
What goes wrong in the argument is that it assumes that we are
only ever originally responsible for our beliefs and never derivatively
only if one realizes or believes that the event in question is significant to one; one may
fail to see that at some point.
207
208
responsible. It is true that for most of our beliefs, they are either forced
upon us or we did not foresee that we would have those beliefs. However,
if there is derivative responsibility, we can be responsible for our beliefs,
even if they are forced upon us at some time and even if we did not fore-
see which specific beliefs we would hold, because we can exercise influ-
ence on our beliefs. If responsibility only had a synchronic dimension,
then luck, to the extent that it excuses, would obviously be reducible
to force or ignorance as an excuse, given that luck entails the absence
of intentional control. But, since, as I argued in chapter 3 we have dia-
chronic influence on our beliefs, it remains to be seen whether luck pro-
vides a doxastic excuse in addition to force and ignorance.
Let us distinguish four major ways in which one can have doxastic
luck. By “doxastic luck” I mean the following: one believes that p, believ-
ing that p is beyond one’s intentional control, one could easily have failed
to believe that p, and believing that p is of some significance to one, as
many of our beliefs are. Making the distinction between these four vari-
eties of luck is important for three reasons. First, doing so helps us to see
that to the extent that they excuse—as three of them sometimes do—
they are reducible to the doxastic excuse of force or ignorance. Second,
although it does not excuse, the fourth kind of doxastic luck shows us
that our account of responsible belief needs an important revision on
one point. Third, distinguishing these four varieties of luck suggests
that virtually all of our beliefs are subject to luck. The latter is one of the
three propositions that jointly constitute the problem of doxastic luck,
to which I return in section 6.5.16
Before doing so, however, I would like to clearly distinguish doxastic
luck from a closely related phenomenon, namely that of doxastic property
luck. By the latter, I mean its being a matter of luck that one’s belief exem-
plifies certain properties that are epistemically, morally, prudentially, or
16. I s it not a matter of luck that I exist? And if it is, then are not all of our beliefs subject to
luck? I think there is some truth to this suggestion. At the time I was conceived, it could
easily have happened that I would not have come into existence. And when I barely sur-
vive a heart operation, it might be a matter of luck that I still exist. Nonetheless, dur-
ing large parts of our lives it does not seem a matter of luck that we exist. For it seems
that, once one exists, worlds in which one ceases to exist are usually not close to the
actual world.
208
209
otherwise desirable or undesirable. The most well known of these prop-

erties is, of course, truth.17 We could say, for instance, that a belief is
alethically lucky if it is a matter of luck that one’s belief is true. Especially
since the publication of Edmund Gettier’s famous paper “Is Justified
True Belief Knowledge?” in 1963, the concept of alethic luck has played
an important role in the debate about the analysis of knowledge. If I look
at the clock and it tells me that it is 3 p.m., while the clock stopped work-
ing twenty-four hours ago, I am the victim of alethic luck, a kind of luck
that is widely considered to be incompatible with knowledge. To the
extent that luck has been discussed in the epistemological literature, the
focus has been almost exclusively on alethic luck. It may be that it can
also be a matter of luck that a belief is, for instance, adequately based or
harmful. Whether one suffers from doxastic property luck or not makes
a difference to whether or not one’s beliefs have certain properties, but not
to whether or not one holds those beliefs. However, the aim of this chapter
is to find out if luck ever provides a doxastic excuse, that is, an excuse
for having some belief. Thus, even though doxastic property luck may be
important in all sorts of ways, it is not relevant for our purposes.
The first way that one can be doxastically lucky is by having mech-
anism luck. Mechanism luck comes in at least two varieties. First, the
range of people’s doxastic mechanisms differs and which doxastic mech-
anisms one has seems to be a matter of luck. Some people are blind,
whereas other people can see. The former cannot form beliefs on the
basis of visual perception, whereas the latter can. It seems that this is
often a matter of luck; an accident or a little genetic change might make a
difference as to whether or not one can see. And, clearly, whether or not
one can see will make a difference to some of one’s beliefs. Second, the
functioning of people’s doxastic mechanisms differs and how well they
function often seems a matter of luck.18 Some people’s eyes work bet-
ter than mine, other people’s eyes worse. Of course, one can sometimes
work on the functioning of one’s doxastic mechanisms, for instance, by
17. Th is kind of luck has also been distinguished by Engel (1992, 66–70), Harper (1996,
279), Pritchard (2005, 146), who calls it “veritic epistemic luck,” Unger (1968, 159).
18. Th is second variety of mechanism luck is also distinguished by Harper (1996, 279),
Pritchard (2005, 134).
209
210
wearing glasses. However, one initially finds oneself with certain doxas-
tic mechanisms with a certain degree of proper functioning and in some
cases there is nothing one can do to improve the functioning of one’s
doxastic mechanisms or to acquire a doxastic mechanism that one lacks.
This second variety of mechanism luck will also carry along doxastic
luck. For example, people whose eyes are working better than mine will,
upon looking at a traffic board from a certain distance, acquire beliefs
that I will not acquire in those very same circumstances.
It seems, though, that this kind of luck, to the extent that it provides
a doxastic excuse, is a particular case of force or ignorance. If there is
nothing I could do to acquire or lose a particular doxastic mechanism
and if there is nothing I could do to improve its functioning, then it
is hard to see how I can be blamed for the ensuing beliefs. If there is
something I could have done to that effect without being aware of it, it
seems that I am excused by my ignorance rather than force. However,
if I have intentional control over which mechanisms I have and how
well they function and I am aware of that, then it seems that if I have an
obligation to, say, improve their functioning, but fail to do so, then I am
blameworthy both for violating that obligation and for the beliefs that
are relevantly related to it. I conclude that to the extent that it excuses,
mechanism luck is reducible to the doxastic excuse of force or ignorance.
Second, there is evidential luck: it is often a matter of luck what evi-
dence one is presented with.19 Imagine that a military commander plans
an invasion and, therefore, orders his spies to gather as much informa-
tion on the situation at the enemy’s coast as they can. Mr. Nose is one of
them. Given his task, he is under a professional or maybe even a moral
obligation to gather evidence on the situation at the enemy’s coast. When
he is about to gather that evidence by consulting the local resistance,
enemy soldiers arrest him for a minor traffic violation and throw him
into prison. It was bad luck that he was arrested: in most nearby possible
worlds, he is not arrested and gathers all the relevant evidence. However,
now that he is in prison, he is unable to do so. He is doxastically excused
19. Th at there is evidential luck has been rightly noticed by Harper (1996, 279), Pritchard
(2005, 136). Evidential luck is the doxastic analogue of what in ethics has been called
“circumstantial luck”; see Rescher (1993, 153–154), Zimmerman (2002, 563).
210
211
for his ignorance—false beliefs and deep ignorance—about the situa-

tion at the enemy’s coast due to bad evidential luck.
However, it seems that his bad evidential luck that counts as a doxas-
tic excuse is reducible to the excuse of force or ignorance. Mr. Nose could
have met his intellectual obligation, for instance, by not taking a car that
particular night or consulting the resistance fighters the next morning.
But he was ignorant that he could meet his intellectual obligation in that
way. And, mutatis mutandis, the same applies to many other situations
in which evidential luck provides a doxastic excuse: we violate our intel-
lectual obligations but are excused for the ensuing doxastic attitudes,
because we did not know, say, how to meet our intellectual obligation to
acquire the relevant evidence.
Is evidential luck ever reducible to force when evidential luck counts
as an excuse? It is much harder to come up with such cases. That is
because if one is excused by force, one could not meet one’s intellectual
obligation to gather or avoid certain evidence. But if gathering or avoid-
ing evidence was impossible for one, then it is hard to see how one could
still be in a situation of evidential luck. Luck, after all, requires that the
state of affairs that obtains does not obtain in a significant number of
nearby possible worlds. And that seems to imply that it was possible for
one to acquire the relevant evidence and, thereby, meet one’s intellec-
tual obligation.
Third, one can be doxastically lucky by being aretaically lucky. 20 One
is aretaically lucky if it is a matter of luck that one has the intellectual
virtues (good luck) and vices (bad luck) that one has. When one grows
up, one finds oneself at a certain point with particular virtues and vices.
Given a slight change in one’s genetic material or the circumstances in
which one was raised, one might have lacked some virtue that one has,
one might have had a virtue that one lacks, one might not have had some
vice that one has, and so forth. Naturally, from a certain age onward,
one can improve one’s intellectual virtues or deteriorate one’s intel-
lectual vices. Nobody starts from scratch, though: one finds oneself
20. Ethicists talk about constitutive luck, that is, luck in being the kind of person one is. See
Latus (2000, 155), Nagel (1993, 60).
211
212
with certain intellectual virtues and vices that arise from one’s physi-
cal makeup (one’s brain, one’s genes) and one’s environment (education,
culture). To the extent that one has not been able to work on one’s intel-
lectual virtues and vices or to the extent that doing so would not make
a difference to whether or not one has certain intellectual virtues and
vices, it is, at least for some virtues and vices, a matter of luck that one has
them or lacks them. They are beyond one’s intentional control, impor-
tant to one and others, and they could easily have been absent or present.
Now, to the extent that aretaic luck makes a difference to one’s
beliefs, aretaic luck seems to imply doxastic luck. I think there are quite a
few such cases. For instance, whether or not I am open-m inded or intel-
lectually thorough is something that will often make a crucial difference
to what I believe. If my intellectual virtues or vices render it impossible
for me to believe that p in particular evidential circumstances, and if it is
a matter of luck that I have the intellectual virtues and vices that I have,
then it seems that aretaic luck at least sometimes leads to doxastic luck.
Again, however, I think that to the extent that it excuses, aretaic luck
is a special case of force or ignorance. Imagine that I have an intellec-
tual obligation to work on my open-m indedness. If I suffer from aretaic
luck, then I lack intentional control over my open-m indedness. But my
lacking intentional control over that means that there are three options.
Either my intentions do not make a difference to whether or not I am
(more) open-m inded, in which case I am forced, or I am ignorant about
how I can become more open-m inded, or both. Clearly, if I do have such
intentional control and still fail to work on my open-m indedness, I am
blameworthy for that and for the ensuing beliefs. Thus, to the extent that
aretaic luck excuses, it is a specific variety of force or ignorance.
Of course, mechanism luck, evidential luck, and aretaic luck mir-
ror the three belief-i nfluencing factors that I distinguished in c hapter 3
(§3.2). In fact, that these three kinds of doxastic luck are reducible to the
excuses of force and ignorance naturally follows from the fact that they
mirror these belief-i nfluencing factors. For I argued that whether we hold
a belief responsibly depends on whether we have met our intellectual
obligations concerning certain belief-influencing factors from which
that belief issues. But if it is a matter of luck that some belief-i nfluencing
factor obtains, it follows that its obtaining is beyond our control. If it is
212
213
beyond our control, then we cannot be blameworthy for the fact that
that belief-i nfluencing factor obtains. But if we are not blameworthy for
that, then it is hard to see how we could be blameworthy for states of
affairs—i n this case, beliefs—t hat issue from that belief-i nfluencing fac-
tor.21 I have nonetheless discussed each of them in some detail, to show
that there are important differences among them—e.g., the fact that, to
the extent that it excuses, evidential luck is virtually always reducible
to ignorance and virtually never to force—a nd because the notions of
mechanism luck, evidential luck, and aretaic luck play a crucial role in
the problem of doxastic luck that I discuss in section 6.5 below.
The fourth way in which we can be doxastically lucky is crucially
different from the first three kinds of doxastic luck. I dub it consequen-
tial luck.22 What I mean by this is that it is often a matter of luck which
beliefs issue from the violation of a particular intellectual obligation.
Imagine, for instance, that Julia has an intellectual obligation to attend
a racial issues class and that she culpably fails to meet that obligation,
say, because she is lazy. Unbeknownst to anyone else except her teacher,
during the lessons her teacher not only addresses racial issues, but also
shares stories about her many travels to Africa, merely for the fun of it. If
Julia were to attend class, she would abandon certain false beliefs about
Africa that, I stipulate, she now blamelessly holds. Thus, she would aban-
don her belief that the crocodile is the deadliest African animal (a piece
of false information that she read in a recent newspaper article) and
her belief that Sudan is the largest African country (something she was
taught at school before Southern Sudan split from Northern Sudan).
If she were to attend class, she would, on the basis of her teacher’s
clearly reliable testimony, come to believe that the hippo is the deadli-
est African animal and that Algeria is the largest African country. It is
clear that if Julia violates her intellectual obligation to attend class, she
is blameless for having two false beliefs about Africa. But it is a matter
21. Some philosophers have rightly noticed that luck is closely bound up with force and
ignorance; see, for instance, Rescher (1995, 26). I have made the slightly stronger claim
that to the extent that it excuses, it is reducible to force or ignorance.
22. A gain, there is an analogue in ethics, which concerns luck for consequences generally;
see Rescher (1993, 152–153), Zimmerman (2002, 559).
213
214
of luck that she would have abandoned those beliefs if she had attended
class. For her teacher could have decided not to share that knowledge
with her students and Julia could easily have had a different teacher who
lacked such knowledge.
In fact, it seems that virtually any belief is subject to consequential
luck. After all, I argued that even if we have intentional control over
whether or not we meet an intellectual obligation, we hardly ever have
control over the ensuing beliefs. And it seems that for most of our intel-
lectual obligations, such as those of gathering evidence and working on
our intellectual virtues and vices, whether or not we meet them does
not guarantee that one has a particular belief. This is because precisely
which beliefs issue from whether or not one meets one’s intellectual
obligations will often depend on factors that are beyond one’s control
and that could easily have been different.
Consequential luck, however, is not reducible to the excuse of force
or ignorance. And that is because it does not excuse at all. On the one
hand, if Julia does not attend a racial issues class and thereby fails to
abandon certain beliefs about the African fauna and African geogra-
phy, because her teacher happens also to share her knowledge on that
topic, she is not excused for continuing to hold those beliefs. For, even if
it were not a matter of luck that her teacher shares her knowledge about
Africa in class, Julia would not be blameworthy for continuing to hold
certain false beliefs about Africa. This is because her intellectual obliga-
tion arises from the fact that she should do something about her racist
beliefs, not about any other beliefs. On the other hand, the fact that it is
a matter of luck precisely which racial beliefs of hers change as a result
of attending the racial issues class clearly does not count as an excuse for
Julia’s continuing to hold certain racist beliefs if she fails to attend class,
because she can still be blameworthy for certain racist beliefs if she fails
to attend class.
Of course, we still need an answer to the crucially important ques-
tion of how we can distinguish cases of consequential luck of the first
kind from cases of consequential luck of the second kind. For, as it stands,
my account of responsible belief has the false implication that Julia is
blameworthy for certain false beliefs of hers about the African fauna and
geography, because those are beliefs she would have abandoned if she
214
215
had met her intellectual obligation to attend a racial issues class. Section
6.4 is devoted to solving this problem.
6.4 C
ONSEQUENTIAL LUCK
AND ACCIDENTALIT Y
Reverting once again to Julia’s case, imagine that Julia violates an intel-
lectual obligation to attend a racial issues class and, unbeknownst to
her, the teacher not only teaches racial issues, but also, for the fun of
it, shares her knowledge about African fauna and geography that she
acquired during her travels to Africa. Then Julia is not blameworthy
for certain false beliefs that she holds about, say, African animals and
African countries that she would have abandoned if she had attended
class—again, assuming that when she has the opportunity to attend
the racial issues class, she holds those beliefs blamelessly. However, as
it stands, my account implies that she is blameworthy for those beliefs,
for it happens to be the case that she would have abandoned them if she
had met her obligation to attend the racial issues class. How can we solve
this problem?
First, one may suggest that Julia foresees that she will maintain or is
more likely to maintain certain racist beliefs as a result of not attending
class, whereas she does not foresee that she will maintain certain beliefs
about the African fauna and geography as a result of that. However, we
can easily amend the case in such a way that Julia knows that the teacher
shares her knowledge about African fauna and geography with her stu-
dents. Even then, though, it seems rather clear that she is not blamewor-
thy for those beliefs about Africa that she would have abandoned if she
had attended class. The fact that she would have abandoned her false
beliefs about Africa if she had attended the racial issues class is clearly
too accidental for that.
Second, one might say that it is a matter of luck that those beliefs
issue from the violation of one’s intellectual obligation to attend that
racial issues class. In at least some nearby possible worlds, Julia does not
change some of her beliefs about Africa as a result of attending the racial
issues class, for instance, because in those worlds, her teacher decides
215
216
not to share her knowledge about Africa or because the class is taught
by a different teacher. I think that this suggestion fails, though. As
I argued in section 6.3, virtually all of our beliefs are to some extent or
other subject to consequential luck, so that this response to the problem
would remove responsibility for virtually all of our beliefs. Imagine, for
instance, that the topic of the race issues class is the persecution of Jews,
especially during World War II. If Julia had attended class, she would
have changed some of her beliefs on Jews. But in a nearby possible world,
the teacher, who hesitates about the topic, chooses to treat the history of
Afro-A mericans. If she had attended the race issues class in that possible
world, Julia might have changed some of her beliefs on black people, but
not some of her beliefs on Jews. However, it seems rather clear that if
Julia fails to attend class, she is blameworthy for maintaining certain
racist beliefs about Jews, even if it is a matter of luck (consequential luck)
that she would have abandoned those beliefs if she had met her intellec-
tual obligation to attend class.
I would like to suggest a different approach: Julia’s beliefs about the
African fauna and geography are accidentally related to her intellectual
obligation in a way that her racist beliefs are not. In order to see how this
solves the problem, let me start with some stage setting. First, I think it
is important to focus on one’s original intellectual obligation. By that
I mean the more general obligation from which one’s specific obligation
is derived. Thus, Julia has an intellectual obligation to attend a racial
issues class at a particular time and place because she has noted that she
holds racist beliefs, that that is a terrible thing, and that she should try
to do something about it. In this case, then, Julia’s original obligation
is to perform a belief-influencing action that will rid her of her racist
beliefs. If she gets a unique opportunity to do so by attending a racial
issues class, then she has a derivative obligation to attend that class, a
specific obligation that derives from her original obligation to perform a
belief-i nfluencing action that rids her of her racist beliefs. Let us call her
original obligation O and her derivative obligation O*.
Let us assume that Julia holds certain racist beliefs r1, r2 , and r3 and
certain beliefs about African fauna and geography a1, a2 , and a3 that are
all false and that are all such that if she were to meet O*, she would aban-
don them, because her teacher not only teaches racial issues but also
216
217
shares, for the fun of it, her knowledge about the African countries that
she visited. Now, what I would like to suggest is that r1, r 2 , and r3 are non-
accidentally related to O, whereas a1, a2 , and a3 are accidentally related
to O, in the sense that Julia has O partly in virtue of or because of the fact
that there is something subjectively and in this case also objectively bad
about r1, r 2 , and r3. Of course, there is also something objectively bad
about a1, a2 , and a3, for these beliefs are false. But Julia has an intellectual
obligation to try to remove her racist beliefs because she realizes that her
racist beliefs are false and morally bad, not because a1, a2 , and a3 are false.
Julia has an obligation O* to attend the racial issues class because she has
an obligation O to try to remove certain racist beliefs that she holds. But
she has an obligation O to try to remove those racist beliefs because she
acknowledges that they are false and morally bad.
If what I have argued in this section is correct, then our new analysis
of responsible belief looks as follows:
S responsibly believes that p iff (i) S believes that p, (ii) S

(6)
is the proper subject of intellectual obligations, (iii) there
is some belief-influencing action or series of actions A that
S could have performed such that if S had performed A, S
would not have believed that p, and (iv) S has not violated any
original all-t hings-considered intellectual obligation to which
S’s belief that p is non-accidentally related such that if S had met
that obligation, certain belief-i nfluencing factors would have
changed in such a way that S would not have believed that p,
or (v) S has violated such an obligation, but is excused for that
by force or ignorance.23
Notice that I have removed the clause “S’s belief that p is not about those
obligations or belief-i nfluencing factors,” which I included in earlier ver-
sions of my account of responsible belief. This is because the clause “S
has not violated any original intellectual obligations to which S’s belief
that p is non-accidentally related” also takes care of the cases that the
23. A gain, the analysans of (6) should be read as: (i) ∧ (ii) ∧ (iii) ∧ [(iv) ∨ (v)].
217
218
earlier clause was meant to take care of. Beliefs about one’s intellec-
tual obligations and about belief-influencing factors that would have
changed as a result of meeting those intellectual obligations are acci-
dentally related to those intellectual obligations in the specific sense in
which I have used the word “accidentally,” because it is not the case that
one has those intellectual obligations because it would be objectively or
subjectively bad to have those beliefs.
6.5 T HE PROBLEM OF DOX ASTIC LUCK
If what I argued in this chapter is correct, then it turns out that we face
a problem. The problem arises from the fact that each of the following
three theses seems true, while it also seems that they cannot all be true:
(7) We are blameworthy for at least some of our beliefs.

(8) Virtually all of our beliefs are subject to some kind of luck
or other.
(9) Luck is incompatible with blameworthiness.
I call an account that convincingly shows that and how these three the-
ses are compatible a solution to the problem of doxastic luck and attempt to
provide such an account in this section. 24
Let me briefly indicate why each of these propositions is at least
prima facie plausible. As to (7), throughout this book I have given sev-
eral examples in which people are blameworthy for some belief or other.
And if my argument in chapter 3 is correct, then we must be at least
sometimes blameworthy for our beliefs, for it is clear that people every
now and then violate their intellectual obligations.
As to (8), I argued above that it is often a matter of consequential
luck precisely which beliefs issue from the violation of an intellectual
obligation. But it seems that virtually all of our beliefs also suffer from
mechanism luck, evidential luck, or aretaic luck. For instance, for any
24. W hat I argue in this section is based on Peels (2015a).
218
219
belief that I have, if I had lacked certain doxastic mechanisms, I would

not have had that belief. Here, we should think especially of those vari-
eties of luck that are reducible to force. For, as I argued in chapter 4,
whether we are responsible (and, hence, possibly blameworthy) for a
belief depends on whether or not the world provides us with an oppor-
tunity to make a difference to whether or not we hold that belief. But
there seem to be many possible worlds in which I lack certain intellec-
tual virtues or vices, certain doxastic mechanisms, or certain evidence
that allow me to hold a certain belief. Also, I argued in chapter 4 that in
certain doxastic Frankfurt-style scenarios, one is blameworthy only for
a belief-particular, not for a belief-u niversal. But then again, what one
is blameworthy for is a matter of luck, for it can be a matter of luck that
one is (not) in a doxastic Frankfurt-style scenario. 25 Hence, we have dif-
ferent reasons to think that luck inevitably affects which beliefs we hold
and that (8) is, therefore, true.
Finally, as to (9), it seems that if the actualization of some state of
affairs Σ is due to luck, then one is not blameworthy for the occurrence
of Σ. The driving intuition behind this proposition is that we are not
blameworthy for what is not within our control. I am not blameworthy
for the deaths of many people in the Middle East these days or for the
large-scale fraud in one of the major companies in my country when
I was still a child, because these are events that are beyond my control.
The same seems to apply in the doxastic realm. Imagine, for instance,
that Julia and Melanie both have an intellectual obligation to attend a
racial issues class, since they find themselves with the same racist beliefs
that they acknowledge to be bad. So far, this is the only opportunity they
have had to rid themselves of their racist beliefs. Both Julia and Melanie
culpably violate their obligation to attend class. The teacher of Julia’s
class is particularly good and attending that class would have made a
significant difference to Julia’s racist beliefs. The teacher of Melanie’s
class happens to be particularly bad and attending that class would not
have made a difference to Melanie’s racist beliefs.
25. According to Nottelmann (2007, 162), this in fact constitutes a problem for a counter-
factual account of responsible or blameworthy belief, since, according to him, whether
or not we are responsible for a belief should not depend on luck.
219
220
Now, on my analysis, Julia is blameworthy for certain racist beliefs,

whereas Melanie is not, because Julia could have made a difference to
them, whereas Melanie could not. However, it is purely a matter of luck
that this is the case. It is something that is beyond Julia’s and Melanie’s
control. And it tells us nothing about Julia or Melanie. How could their
blameworthiness depend on such purely external factors—factors that
are completely beyond their control—as which teacher they happen to
have and how good that teacher is? That would seem both unfair and
incorrect.
I just gave an example of consequential doxastic luck that seems
incompatible with doxastic blameworthiness. But something similar can
be said for mechanism doxastic luck, evidential doxastic luck, and aretaic
doxastic luck. Imagine that I realize that my eyes are not functioning too
well and that I need to buy glasses. In this way I can improve the func-
tioning of my doxastic mechanism of belief-formation on the basis of
visual perception. My friend might not have that obligation because he
is blind. Am I blameworthy for the beliefs that I would have had if I had
met my intellectual obligation to improve the functioning of that dox-
astic mechanism? More precisely, am I more to blame than my friend,
just because he happens not to have such an intellectual obligation and,
therefore, cannot be blameworthy for (not) having certain beliefs? Or
imagine that I am intellectually sloppy and fail to gather evidence on p,
so that I continue to disbelieve that p. Am I more blameworthy than my
friend who happens to be raised in an intellectually conscientious fam-
ily, so that he happens to be intellectually thorough, gathers evidence on
p, and does not disbelieve that p? Again, it is hard to see how luck could
make a difference to whether or not one is blameworthy for a belief.
Now, (9) is about blameworthy belief. One might think that some-
thing similar applies to responsible belief: whether or not one believes
responsibly cannot be a matter of luck.26 For two reasons, I focus on
blameworthy belief here. First, the topic of this chapter is luck as a doxas-
tic excuse: a state of affairs that removes blameworthiness. Since we saw
26. I n fact, some philosophers have claimed that responsibility or moral worth in general
is incompatible with luck; see, respectively, Corlett (2008, 190) and Statman (1993, 1).
220
221
that virtually all of our beliefs are subject to luck, the problem naturally
arises that it seems to follow that we are blameworthy for none of our
beliefs. Second, it seems that if blameworthy belief is compatible with
luck, then so is responsible belief. After all, we saw in chapter 1 that if one
is blameworthy for one’s belief, one is responsible for it, and responsible
belief is belief for which one is responsible but not blameworthy. Thus,
whether or not one believes responsibly is a matter of whether or not, if
one is responsible for one’s belief, one believes blameworthily.
As far as I know, nobody has formulated, let alone discussed, a solu-
tion to the problem of doxastic luck. Let us, therefore, consider the
responses that have been given to the closely related problem of moral
luck in ethics. The problem of moral luck in ethics is that we blame peo-
ple for certain non-doxastic consequences of their actions, even though
whether or not those consequences would obtain as a result of perform-
ing those actions is sometimes a matter of luck and it seems unfair to
blame people for consequences that are due to bad luck. Imagine, for
instance, that two truck drivers, Sam and John, get drunk. In fact, they
get equally drunk. Both of them, being drunk, drive home. By bad luck,
a pedestrian crosses Sam’s path and, being unable to react adequately,
Sam runs him over. By good luck, nobody crosses John’s path. Being
drunk, he would not have been able to avoid running a pedestrian over
if a pedestrian had crossed his path. It seems that we blame Sam more
than John. But if the fact that somebody crosses Sam’s path is merely
due to bad luck, then how can we properly blame Sam more than John?27
Let us transpose the solutions that have been offered to the problem
of moral luck and see whether they hold when it comes to the problem of
doxastic luck. It is not my aim here to decisively refute these responses,
but rather to point to some major problems that they face that warrant
the quest for an alternative solution.
First, some philosophers, such as Jonathan Adler and Margaret
Coyne, simply deny that blameworthiness is incompatible with luck:
one can be more blameworthy (blameworthy to a higher degree)
27. Th at luck in general is unavoidable is rightly argued by Feinberg (1962, 348–350),
Rescher (1995, 19).
221
222
for a belief than someone else because one happens to suffer from bad
luck, whereas that other person happens not to suffer from bad luck. 28
We should accept and live with the fact that life is unfair. I find this
response unsatisfying. True, life can be lucky or unlucky for a particular
person. It does not follow that that person’s degree of blameworthiness
is something that depends partly on external factors. Of course, one can
still stick to the conviction that one’s degree of blameworthiness can be
partly a matter of factors that are due to luck. Remember, however, that
for most philosophers (7)–(9) jointly constitute a problem, because (9)
has, at least initially, great plausibility.29 But we have not been given a
reason to think that the initial plausibility of (9) is deceiving apart from
the fact that, together with (7) and (8), it leads to a contradiction. 30 If no
further justification can be provided, one could just as well deny (7) or
(8). The proposal that I will make below does provide a justification for
abandoning (9) that is independent of the fact that it leads to a problem.
Second, other philosophers, such as Brynmor Browne, have sug-
gested that we should abandon such normative attitudes as praise and
blame and replace them with more objective attitudes. It would be unfair
to blame someone for something that is due to bad luck, but it would not
be unfair to take a more objective attitude, such as disapproval, toward
someone for having a particular belief that is due to luck. 31 I find this
response problematic as well. As I argued in chapter 1, there seem to
be situations in which normative attitudes like praise and blame are
perfectly legitimate. In some situations it would even be improper not
to have such normative attitudes. It seems improper not to blame a
Neo-Nazi for his heinous racist beliefs, if we know that he acquired or
maintained them culpably. Moreover, if what I argued in chapter 1 is
right, then to give up entirely the normative attitudes (or beliefs that the
28. For this position, see Adams (1985, 14), Adler (1987, 248), Andre (1993, 125), Coyne
(1985, 322), Walker (1993, 243–2 47), Weatherson (2008, 567), Williams (1993, 37–
38), Williamson (2007a, 116–119).
29. Th
is is something which Nagel (1993, 58, 69), for instance, acknowledges.
30. A n exception is Nagel (1993, 66), who claims that our will (intentions, decisions) is
the product of antecedent circumstances. Such a view of the will, however, is highly
controversial.
31. See Browne (1992, 349–356).
222
223
normative attitudes are proper in some situation) is to give up responsi-

bility. But we do hold each other at least sometimes responsible for our
beliefs and sometimes blame each other for holding certain beliefs, and
we have not been given any good reason, apart from the problem of dox-
astic luck itself, to think that this practice is mistaken.
On a third response that one might give to the problem, 32 we should
say the following about the scenario involving Julia and Melanie. The
question whether Melanie is more blameworthy than Julia is nonsensical,
for Julia and Melanie are blameworthy for different sets of states of affairs
and one can compare people’s blameworthiness only if they are blame-
worthy for the same states of affairs. Julia is blameworthy for not attending
the racial issues class and for her racist beliefs that she would have aban-
doned if she had attended class, whereas Melanie is only blameworthy
for not attending class, since, given the bad organization of her teacher,
she would not have abandoned any racist beliefs if she had attended class.
We cannot compare Julia’s blameworthiness for violating an intellectual
obligation and for holding certain beliefs on the one hand with Melanie’s
blameworthiness for violating an intellectual obligation on the other.
I do not find this response to the problem of moral luck convincing
either. One can say truly that one is more interested in Hume’s A Treatise of
Human Nature than in his Enquiries Concerning Human Understanding,
or that one was happier about the birth of one’s first child than about the
acceptance of one’s first article, or that one is more satisfied with one stu-
dent’s work than with that of another. We can even make such compara-
tive judgments for different states of affairs in the realm of responsibility.
Imagine that Sam notices that an old man is being robbed and seriously
beaten by a group of armed men. He decides to call an ambulance, but
deems it too dangerous to intervene. Fred witnesses the same events,
but intervenes, scares off the attackers, and then calls an ambulance.
It surely seems that Fred is more praiseworthy than Sam. And if we are
asked what he is more praiseworthy for, it seems right to say that Fred is
more praiseworthy for intervening and then calling an ambulance than
32. A s far as I know, nobody has actually defended this response in the literature: This idea
was suggested to me by René van Woudenberg.
223
224
Sam is for merely calling an ambulance. But if this is true for praisewor-
thiness, then why would it not be true for blameworthiness?
Again, I do not claim that this is all there is to be said about these
three responses to the problem of moral luck. But it seems to me that the
problems that I have formulated for them warrant looking for an alterna-
tive solution.
On a fourth approach that is advocated by, among others, Latus
and Zimmerman, and that I will defend myself, we should distinguish
between scope and degree of blameworthiness, where the scope of one’s
blameworthiness concerns how large the set of states of affairs is for
which one is responsible, whereas the degree of one’s blameworthiness
concerns how blameworthy one is for the actualization of those states
of affairs. 33
The following case of consequential doxastic luck illustrates the point.
Assume that both Julia and Melanie blamelessly hold racist beliefs r1, r 2 ,
and r 3. They have good reasons to think that the only opportunity to
do something about this is to attend a racial issues class. However, they
are both lazy and culpably violate an intellectual obligation to attend
class. Julia would abandon her racist beliefs r1, r 2 , and r3 if she were to
attend class, but Melanie would have abandoned none of those beliefs,
for her teacher is so badly organized that her teacher does not even come
to a substantial treatment of the issue. It follows that it is a matter of evi-
dential luck and doxastic luck that Melanie would have maintained her
beliefs r1, r 2 , and r3, whereas Julia would have abandoned them. On the
proposal under consideration, Julia and Melanie are not blameworthy
for the exact same set of states of affairs. Julia is blameworthy both for
violating her intellectual obligation to attend class and for her beliefs r1,
r 2 , and r 3, whereas Melanie is blameworthy only for not attending class.
Hence, the scope of their blameworthiness differs. However, the degree
of their blameworthiness is identical, for they violated the same obliga-
tion and it is a matter of luck to which beliefs their meeting the obliga-
tion would have made a difference.
33. See, for instance, Latus (2000, 151), Thomson (1993, 205), Zimmerman (1988, 56–57;
1993, 227; 2002, 560; 2006, 598–6 01).
224
225
The idea, then, is that luck may broaden the scope of one’s doxastic
blameworthiness, but that it cannot add to the degree of one’s doxastic
blameworthiness. Thus, if (9) says that one cannot be blameworthy for
beliefs that are due to luck, it is false, and if it says that one cannot be more
blameworthy for events that are due to luck (more blameworthy than if
one had not had those beliefs), it is true. This would solve the problem
of doxastic luck, for the claim that luck cannot make a difference to the
degree of one’s doxastic blameworthiness is perfectly compatible with
the claim that what happens due to luck is beyond one’s control and the
claim that we properly (rightly, deservedly) blame people for beliefs that
are due to luck.
There seem to be two main lines of attack on the approach that distin-
guishes between scope and degree of blameworthiness. First, one might
argue that if two persons S and S* are blameworthy for the same thing
but S is, in addition to that, blameworthy for something else, then S must
be more blameworthy than S*. The idea is that in this regard, blamewor-
thiness is similar to phenomena like weight, length, or temperature: if
two objects have the exact same weight, and some weight, however little,
is added to the first object, then the first object must become at least
somewhat heavier than the second. 34 The problem with this response is
that it provides no motivation for thinking that blameworthiness is like
weight or temperature. Why should we not think that blameworthiness
is rather like, say, interest? If I am not only interested in subject matter
A, but also in subject matter B, am I thereby more interested? No, I may be
interested in more subject matters. It does not follow that I am somehow
more interested than someone who is only interested in A. I conclude that
the first objection against this move is unconvincing.
Second, one might argue that the approach that distinguishes
between scope and degree of blameworthiness faces a reductio ad absur-
dum. 35 Imagine that Julia and Melanie have the same racist beliefs, but
that only Julia happens to have the opportunity to attend a racial issues
class. Julia violates the obligation to attend that class. If Melanie had
happened to have the opportunity to attend a racial issues class, she
34. Thus, for instance, Smith (2013).

35. Th is has been pointed out by, among others, Latus (2000, 153–155).
225
226
also would have violated the obligation to attend class. It seems that we
would now have to say that Julia and Melanie are equally blamewor-
thy. For it is just a matter of luck that Julia has an opportunity to meet
her obligation, whereas Melanie does not. What Melanie and Julia are
blameworthy for and what renders them equally blameworthy is that
they are both such that if they had an opportunity to meet their intel-
lectual obligation by attending a racial issues class, they would violate
that obligation.
If we also take into account aretaic luck, mechanism luck, and evi-
dential luck, so the objection continues, the picture becomes even more
complicated and less plausible. It is a matter of luck that I happen to be
open-m inded and that I happen to live now. I could have been narrow-
minded and I could have lived during World War II. If it is true that if
I had been narrow-m inded and had lived during World War II, I would
have freely absorbed Nazi doctrine and formed all sorts of heinous
beliefs, then it follows that I am equally blameworthy as someone who
in fact was narrow-m inded, lived during World War II, freely absorbed
Nazi doctrine, and formed all sorts of heinous beliefs. This is because
we are both such that if we had been narrow-m inded and lived during
World War II, we would have absorbed Nazi ideology and formed all
sorts of heinous racist beliefs. I now turn out to be blameworthy for
being such that I would have formed all sorts of beliefs in radically differ-
ent historical circumstances, in other parts of the universe, in scenarios
in which my intellectual character is radically different, and so forth,
and that seems to be a dire consequence. 36
What could we say in defense of the strategy that distinguishes
between scope and degree of blameworthiness? First, some philoso-
phers, such as Nicholas Rescher, have argued that there is no such thing
as aretaic luck, that we have the virtues and vices that we have essen-
tially, that we would not be the same person if we lacked the virtues
36. One could, of course, deny that there are such things as true counterfactuals about
what individuals would have done or believed in different circumstances, either
because there are no such counterfactuals or because they are not true or false.
Many philosophers, though, think that there are such true counterfactuals, so in
what follows I assume that it makes sense to talk about the truth or falsehood of such
counterfactuals.
226
227
and vices we have or if we had virtues and vices we lack. 37 It is not clear
what this view says about mechanisms, but it seems that they fall on the
contingent rather than the essential side, for people can, say, become
blind and thereby lose the capacity of forming beliefs on the basis of
visual perception. Hence, the idea seems to be that we should draw a
line between mechanism luck, evidential luck, and consequential luck
on the one hand and aretaic luck on the other. The suggestion, then,
is that whereas the former are compatible with blameworthiness, the
latter is not.
It seems to me that, whether or not this is true for paradigmatic moral
virtues and vices, such as friendliness and sincerity, it is false for intel-
lectual virtues and vices. 38 Many people acquire or abandon intellectual
virtues and vices in the course of their lives. Persons who were raised in
a small, conservative, narrow-m inded community can learn to be criti-
cal and open-m inded. They cannot merely learn to resist a tendency to
be narrow-m inded, they can actually work on it in such a way that after
some time they no longer have such a tendency: open-m indedness has
become a well-established character trait. Some people who were once
narrow-m inded become open-m inded, and sloppy evidence gatherers
can through training become intellectually thorough. People’s intellec-
tual personality can change. It does not follow that they become a dif-
ferent person. After all, they change intellectually. Therefore, this first
response to the objection fails.
Second, one could claim with Zimmerman that Julia and Melanie
are equally blameworthy, without admitting that they are blamewor-
thy for being such that both of them would have violated the same
intellectual obligations if they had been in particular circumstances.
Here, the idea is that although they are both blameworthy in virtue of
something—namely, the violation of an intellectual obligation in partic-
ular circumstances—t hey are not blameworthy for something; they are
37. F
or this response, see Hurley (1993, 198), Rescher (1993, 155; 1995, 30–31, 155–158).
38. Perhaps all virtues and vices, including intellectual virtues and vices, are moral.
I do not want to take a stance on this issue here. What I want to argue is that even
if paradigmatic moral virtues and vices are such that one has them essentially (which
I doubt), this is not true for intellectual virtues and vices, such as open-m indedness and
thoroughness.
227
228
blameworthy tout court. 39 I find it hard to make sense of this response.

If one is not blameworthy for something, then one is blameworthy for
nothing. And if one is blameworthy for nothing, then one is not blame-
worthy at all, in the same way as someone who is responsible for nothing
is not responsible at all. The notion of blameworthiness simpliciter is not
one that we find in ordinary discourse and as long as we have not been
given a description of what blameworthiness simpliciter is supposed to
be, I find it hard to even grasp the concept.
This is not to deny that ordinary discourse is full of claims about how
good or bad people are. And it does make sense to say that someone is
morally good or morally bad simpliciter. One is not morally good or bad
for something. But notice that if we invite someone to explicate in terms
of blameworthiness what she means by someone’s being a bad person,
she will say that someone is a bad person because she is blameworthy
for performing certain bad actions or blameworthy for having a certain
corrupted character. We would be baffled if someone were to explain
someone’s being a bad person in terms of that person’s being blamewor-
thy without being blameworthy for anything, such as her actions, her
moral character, or her beliefs.
Third, one could bite the bullet and say that we are indeed blame-
worthy for all these things, such as the fact that if I had been narrow-
minded and if I had lived during World War II, I would have formed all
sorts of heinous racist beliefs. Zimmerman, for instance, suggests that
my responsibility for what I freely do in other possible worlds and for the
consequences of those things constitute my essential moral worth: I am
responsible for being such that I freely do such and such in some pos-
sible world.40 We could say, then, that my essential intellectual worth is
constituted by my choices to meet or violate intellectual obligations in
other possible worlds, including possible worlds in which I am a person
with radically different intellectual virtues and vices, in which I have
radically different doxastic mechanisms (including, perhaps, doxas-
tic mechanisms that I can now not even conceive of), and in which my
39. S ee Zimmerman (1993, 228; 2002, 564–565; 2006, 605).

40. Th
us, Greco (1995, 94).
228
229
evidential circumstances are radically different from what they are in

the actual world.
In response, I would like to grant that there is perhaps such a thing
as essential moral worth and essential intellectual worth. Perhaps there
is some sense in which I am such that if I had lived in the Mayan society,
had been narrow-m inded, and had been a priest of the Mayan religion,
I freely would have suppressed my moral doubts about human sacrifices
and would, by violating that intellectual obligation, have maintained
certain morally bad beliefs. But am I blameworthy for being such? I think
we would not blame a person for this, even if we were fully informed
about all these counterfactual truths. Somehow, what happens in such
far away possible worlds is irrelevant to one’s responsibility (including
blameworthiness) in the actual world.
My solution to the problem of doxastic luck is closely related to this
observation. In the remainder of this section, I argue that paying close
attention to what it is for some event to be due to luck saves the strategy
that distinguishes between scope and degree of blameworthiness. That
paying attention to the nature of luck helps to solve the problem of luck
has also been suggested by Pritchard.41 However, here I develop this
suggestion in much more detail.42 Since the modal nature of luck plays a
crucial role in my solution, I call it the modal solution to the problem of
doxastic luck.
In section 6.2, I argued that luck is a matter of what happens in close
or nearby possible worlds, something that has been acknowledged in the
literature on epistemic luck but that is often overlooked in the literature
on the problem of moral luck. Since luck concerns what happens in the
actual world and nearby possible worlds, we can safely ignore scenar-
ios that are radically different from the actual world and we can focus
on nearby possible worlds, that is, on how things are or easily could
have been.
41. S ee Pritchard (2005, 260; 2006, 9).

42. Statman (1991) argues that there is a close analogy between moral and epistemic luck,
but he does not discuss how the analysis of epistemic luck that epistemologists provide
might help to solve the problem of moral luck.
229
230
This is not to deny that we sometimes suffer from mechanism luck,

evidential luck, or aretaic luck. There are close possible worlds in which
I lack intellectual curiosity—or, at least, at some time there was a close
possible world in which I lack the character trait of intellectual curiosity.
For instance, I could have been born without a disposition to be intel-
lectually curious. There is a close possible world, or at some time there
was a close possible world, in which I cannot see or in which my eyesight
is much worse or slightly better than it is now. And there is a close world
in which my evidence on certain topics, such as my evidence on herons,
is different from what it is now. But there is no close possible world in
which several intellectual character traits of mine are different, in which
my cognitive mechanisms function radically differently, or in which
my evidence base is largely different from what it is in the actual world.
There are such possible worlds, but they are not nearby: things could not
easily have been that way.
This squares well with our considered verdicts in all sorts of cases. If
Jack lives during World War II and would betray a Jew if he knew where
one was hiding and it is a matter of good luck (good luck for the Jew,
bad luck for Jack) that he does not know a Jew’s hiding place, then, it
seems, he is equally blameworthy as his brother Stephen, who betrays a
Jew because he happens to know some Jew’s hiding place. Of course, we
blame them for (partly) different states of affairs. We blame Stephen, but
not Jack, for betraying a Jew. But their degree of blameworthiness seems
to be identical. If we were to know of some third person, Alex, who hap-
pens to live in Australia about sixty years later, that he would freely have
betrayed a Jew if he had had different character traits and if he had lived
some sixty years earlier in Nazi-occupied Holland, we would not blame
Alex as much as Stephen and Jack, if we would blame him at all. This is
because the world in which Alex does such a thing is not a nearby world,
but a world that is far away, and Alex’s not betraying a Jew is, therefore,
not a matter of luck. Similarly, what intellectual obligations people vio-
late in far away rather than nearby possible worlds and what beliefs they
form as a result of that makes no significant difference, or perhaps no dif-
ference at all, to the degree of their blameworthiness in the actual world.
In this way, we can in the relevant scenarios accommodate all the
intuitions that are articulated in the problem of doxastic luck. First, we
230
231
are at least sometimes blameworthy for our beliefs, because we are at

least sometimes blameworthy for the violation of our intellectual obliga-
tions from which those beliefs issue. Second, (virtually) all of our beliefs
are subject to luck: usually consequential luck, but often also mecha-
nism luck, evidential luck, and aretaic luck. Finally, luck is not incompat-
ible with blameworthiness, for you can be blameworthy for a belief even
if it is a matter of luck that you have that belief.43 It merely seems that the
two are incompatible, because doxastic luck can make a difference to the
scope of one’s beliefs, even if it cannot make a difference to the degree of
one’s blameworthiness. We can maintain our intuition that what hap-
pens beyond our control does not increase or decrease our degree of dox-
astic blameworthiness.
Let me put a bit more flesh to the skeleton of my modal solution
to the problem of doxastic luck by considering three worries. First,
one might wonder what Julia and Melanie are blameworthy for if
Melanie is as blameworthy as Julia. As I pointed out above, the notion
of blameworthiness simpliciter is rather unattractive, so we must say
that they are blameworthy for something. I reply that Melanie is only
blameworthy for violating her intellectual obligation to attend a racial
issues class, whereas Julia is blameworthy for violating her intellec-
tual obligation to attend a racial issues class and for some of her racist
beliefs.44
Second, one may worry that my account of doxastic responsibil-
ity and doxastic luck may solve the problem of doxastic luck, but that it
does not solve a closely related problem with doxastic blameworthiness.
Imagine that Jenny is born in a peaceful democracy, whereas Tim is born
in Nazi Germany. Imagine also that Tim freely and culpably violates his
intellectual obligation to reflect on his Nazi beliefs, whereas Jenny does
not, but would have done so if she had been born in Nazi Germany. If
what I argued is right, it is not a matter of luck that she does not perform
43. Th at consequential luck, whether doxastic or non-doxastic, does not excuse, is claimed
by Bennett (1995, 58–61), Rescher (1993, 154–155), Richards (1993, 169).
44. Nathan Hanna objects to this line of reasoning that someone can be blameworthy to
different degrees in two different scenarios, even though the same counterfactual is
true of him in both scenarios; see Hanna (2014). Elsewhere, I have argued that his cases
fail to make this claim plausible; see Peels (2015a).
231
232
those racist actions, but it still seems beyond Tim’s control that he is born
in Nazi Germany and beyond Jenny’s control that she is born in a peaceful
democracy. How can we properly assess them differently from a norma-
tive perspective if whether or not they hold racist beliefs depends at least
partly on factors that are beyond their control?
Since this problem does not involve the notion of luck, it is somewhat
different from the problem of moral luck that I discussed above. Let us
call this new problem the problem of control. I will be explicit as I can,
even if that means being a bit tedious at points. First, obviously, Jenny
is not responsible for being born in a democracy and Tim is not respon-
sible for being born in Nazi Germany, for those are things that are beyond
their control. Second, if Tim has sufficient reasons to act otherwise and
acts freely in failing to meet his intellectual obligation, he is blameworthy
for that and for the ensuing racist beliefs. Third, for Jenny, growing up in
Nazi Germany is a state of affairs that obtains in a possible world that is
far away from the actual world, in which she lives in a peaceful democ-
racy. It seems undeniable that growing up in a Nazi society would deeply
affect her beliefs, character, and desires. Thus, even if is true that Jenny is
such that if she had grown up in Nazi Germany, she would have formed
heinous racist beliefs, she has no control over being such, for no matter
what she does in the actual world, things in the possible world in which
she grows up in Nazi Germany are so different that what she does and
believes in the actual world will make no difference to what she does and
believes in the possible world in which she grows up in Nazi Germany. In
this regard, Jenny differs from, say, Julia and Melanie, who both have con-
trol over whether or not they meet their intellectual obligation to attend a
racial issues class (or, at least, so I have assumed in the example).
Now, one could try to circumvent this response by construing the
scenario in a slightly different way. In this alternative scenario, Jenny
also lives in a peaceful democracy, but this time someone in that democ-
racy has invented a time machine that can bring someone back to Nazi
Germany. In this scenario, it is true that if she were to enter the time
machine, she would travel back in time to Nazi Germany while her char-
acter, beliefs, and desires remain constant. Now, it seems that Jenny has
control over being such that she would freely violate her intellectual
obligations if she were in Nazi Germany, for if she becomes a loving and
232
233
friendly person in the actual world, then it seems she would not freely
violate her intellectual obligation if the time machine were to bring her
back in time to Nazi Germany, or she might not even have such an intel-
lectual obligation since she would have no racist beliefs that would give
rise to such an intellectual obligation.
This seems right to me. However, we should note that our intuitions
about whether or not she is blameworthy seem to change accordingly: if
Jenny could travel back in time to Nazi Germany while her personality
remains the same and if she would freely violate her intellectual obliga-
tions if she were to do that, then, it seems, she is blameworthy for being
such that she would do that and for the ensuing racist beliefs. In fact, if
this is a real option for Jenny, then it seems that we may well blame her
as much as someone who actually lived in Nazi Germany and freely vio-
lated certain intellectual obligations and formed racist beliefs as a result
of that. Thus, depending on how we construe the scenario, Jenny either
has no control over being such that she would freely violate her intellec-
tual obligation if she lived in Nazi Germany and is not blameworthy for
that, or she has control over being such that she would freely violate her
intellectual obligation if she lived in Nazi Germany and is blameworthy
for that. Both options are perfectly compatible with the position that
I have advocated, namely that what happens beyond one’s control can
make a difference to the scope, but not to the degree, of one’s doxastic
blameworthiness.
Just to be clear: I do not claim that in the alternative scenario it is a
matter of luck that Jenny travels back to Nazi Germany. The government
might select only one out of every million inhabitants to travel back in
time, so it might be the case that in virtually all nearby possible worlds,
Jenny does not live part of her life in Nazi Germany. All I am saying is
that in this scenario, control and blameworthiness seem to come and
go together. Now that, in this alternative scenario, Jenny has control
over whether she is such that she would violate her intellectual obliga-
tions in Nazi Germany, she seems responsible for whether or not she is
such. She does not have control over whether or not she does certain
things in possible worlds that are much further away, worlds in which
she has a different character, different beliefs, and different desires. But
then we do not seem to hold her responsible for that. Thus, neither in
233
234
solving the problem of luck nor in solving the problem of control does
the approach in terms of the distinction between scope and degree that
I have defended face a reductio.
Third, one may object to my modal solution to the problem of dox-
astic luck as follows: do we not blame people more if bad consequences
obtain, such as immoral beliefs, as the result of the violation of an intel-
lectual obligation? Maybe we do and maybe we should. It does not fol-
low that one is more blameworthy if, due to luck, a bad consequence
obtains as the result of the violation of an obligation. There seem to be
several reasons why we sometimes blame such people more than those
who do not suffer from a similar kind of doxastic bad luck. First, we are
often quite irrational and blame someone for something where, upon
further reflection, we would not blame that person, or at least not to that
degree. We easily slip from merely evaluative negative attitudes, such as
disapproval, to normative negative attitudes, such as resentment and
blame.45 Second, our epistemic deficiencies often play a pivotal role in
our practice of blaming people: if some bad consequence obtains, we
will thereby know for sure that that person took a certain risk, whereas
we are not always certain of that if no bad consequence obtains.46 Finally,
overtly blaming S more than S* ought to be distinguished from S’s being
more blameworthy than S*. It can be justified for all sorts of reasons, such
as educational purposes, to overtly blame someone more than someone
else for their beliefs, even if they have violated the exact same intellec-
tual obligations and are, therefore, equally blameworthy.47
6.6 C ONCLUSION
In this chapter, I have considered the doxastic excuse of luck. I argued

that the actualization of some state of affairs Σ is lucky for some person
45. S lote (1992, 120) rightly notices this.

46. Th
us also Richards (1993, 169–178).
47. Jensen (1993, 132–133) also notes this. As Nafsika Athanassoulis points out, this is
especially true for blame in judicial contexts, for legal punishment is often given at least
partly in order to educate the perpetrator and to deter others from committing the same
kind of crime; see Athanassoulis (2005, 59).
234
235
S if and only if the occurrence of Σ is beyond S’s intentional control,

significant to S, and Σ could easily have failed to obtain. With luck thus
understood, there turn out to be four kinds of luck that (often) imply
doxastic luck: mechanism luck, evidential luck, aretaic luck, and conse-
quential luck. To the extent that mechanism luck, evidential luck, and
aretaic luck provide a doxastic excuse, they are reducible to the excuse
of force or ignorance.
Consequential luck does not excuse at all, but a particular variety
of consequential luck shows that one is blameworthy for a belief only if
it is non-accidentally related to the violation of the original intellectual
obligation from which that belief issued.
I provided a modal solution to the problem of doxastic luck. The
problem is that it simultaneously seems that we are blameworthy for at
least some of our beliefs, that virtually all of our beliefs are subject to
luck, and that luck and blameworthiness are incompatible. The problem
can be solved by distinguishing between the scope and the degree of
one’s blameworthiness and by also reminding ourselves that luck is a
matter of what happens in the actual world and nearby possible worlds,
rather than in far away possible worlds. Hence, which beliefs issue from
one’s violation of an intellectual obligation might make a difference to
whether or not one is blameworthy for holding those beliefs, but not to
the degree of one’s doxastic blameworthiness. Thus, the fact that we
sometimes believe responsibly and sometimes blameworthily is per-
fectly compatible with the fact that which beliefs we hold is inevitably
a matter of luck.
235
236
237
APPENDIX
Responsible Belief and

Epistemically Justified Belief
In chapter 2, I pointed out that the deontological conception of epistemic justi-

fication faces the problem of doxastic involuntarism. On the deontological con-
ception, epistemically justified belief is belief for which one is blameless because
one has met or at least not violated any obligations. In that chapter, I bracketed
the issue of epistemic justification and confined myself to responsible belief. Now
that I have given an analysis of responsible belief, the question naturally arises of
precisely how responsible belief, as I analyzed it, relates to epistemically justified
belief. The aim of this Appendix is to sketch some lines of thought that provide at
least the beginning of an answer to this question. I argue that something close to
responsible belief may well be identical to epistemically justified belief.1
Since we are talking about epistemic justification rather than justification tout
court, let us in the analysans of epistemic justification confine ourselves to epis-
temic intellectual obligations. Given the account of responsible belief that I have
provided in this book, the idea that responsible belief just is epistemically justi-
fied belief then looks as follows:
S epistemically justifiedly believes that p if and only if (i) S believes that

(1)
p, (ii) S has certain epistemic intellectual obligations, (iii) there is some
belief-influencing action or series of actions A that S could have per-
formed such that if S had performed A, S would not have believed that p,
1. I have also suggested and argued this elsewhere; see Peels (2015c; 2016a).
237
238
A ppendix
and (iv) S has not violated any original intellectual obligation to which
S’s belief that p is non-accidentally related such that if S had met that
obligation, then certain belief-influencing factors would have changed
in such a way that S would not have believed that p, or (v) S has violated
such an obligation, but is excused for that by force or ignorance.2
We should note that a deontological conception in terms of intellectual obligations

à la (1) meets quite a few worries that we find in the literature. Let me give just one
example. Nottelmann objects to the deontological conception of epistemic justifi-
cation that if it is understood as epistemically permissible belief, there will be noth-
ing specifically deontological about that conception of epistemic justification. This
is because if a belief is epistemically permissible, it is permitted by one’s epistemic
reasons, that is, by one’s evidence. And one can cash out the idea that one’s belief is
permitted by one’s reasons or evidence without appeal to any deontic notions, such
as blame, responsibility, or obligation. 3 Whether or not this objection is sound,
the account of epistemic justification under consideration—proposition (1)—
circumvents this objection, because it is not cashed out in terms of epistemic rea-
sons, but in terms of epistemic intellectual obligations, and, clearly, whether or not
one has met such obligations is a deontological matter—t hat is true by definition.4
Alston, however, argues that a deontological conception based on voluntary
influence, such as (1), is deficient as a conception of epistemic justification. And it
is deficient for reasons other than the argument from doxastic involuntarism that I
discussed in chapter 2. Alston’s argument against something like (1) runs as follows.
First, he emphasizes that the account under consideration is one in terms of
epistemic rather than moral or prudential intellectual obligations. This squares
well with what I said above. 5 Subsequently, Alston points out that there are two
main strands of concepts of epistemic justification, those of externalism and inter-
nalism. Since each kind admits of many varieties, and Alston cannot discuss all
of them, he focuses on two popular accounts. On the externalist account that he
treats, to be justified in believing that p is both to be in a position that is favorable
toward reaching the epistemic goal by having a belief that is based on an adequate
ground, that is, a ground such that basing one’s belief on it renders it likely that
one’s belief is true, and to have a belief that is based on a ground that is fairly read-
ily accessible to one. On the internalist account that he discusses, to be justified
in believing that p is for it to be true that from one’s perspective, one’s belief is
2. A gain, the analysans of (1) should be read as: (i) ∧ (ii) ∧ (iii) ∧ [(iv) ∨ (v)].
3. See Nottelmann (2013, 2229).
4. A lternatively, Nottelmann may object that if the deontological conception of epistemic
justification is cashed out in terms of intellectual obligations, it is not clearly a concep-
tion of distinctively epistemic justification. I return to this worry below.
5. S ee Alston (1989b, 83–8 4; 1989d, 116).
238
239
A ppendix
based on an adequate (truth-conducive) ground. Thus, Alston’s presentation of

externalist and internalist accounts of epistemic justification can be summarized
slightly more formally as follows:
S justifiedly believes that p iff (i) S’s belief that p is based on an ade-

(2)
quate ground (its being based on that ground is truth-conducive)
and (ii) that ground is fairly readily accessible to S.
S justifiedly believes that p iff from S’s perspective, S’s belief that p is
(3)
based on an adequate ground.6
Alston criticizes (1) by arguing that it is co-extensional neither with (2) nor with
(3). Thus, he gives a long list of examples that are meant to show that a subject
can (a) be justified on (1) and yet not form a belief in a truth-conducive way, or
(b) form a belief in a truth-conducive way and yet not be justified on (1), or (c) be
justified on (1) and yet not form a belief that is truth-conducive from her own
perspective, or (d) form a belief that is truth-conducive from her own perspective
and yet not be justified on (1).7 Here, I will not discuss the examples, but simply
assume that they show that (1) is co-extensional with neither (2) nor (3). Alston
concludes on this basis that (1) cannot play the central role that we demand of
epistemic justification. 8
This argument of Alston has gone largely unnoticed in the literature. By far
most epistemologists have focused on the argument from doxastic involuntarism,
as I discussed it in c hapter 2. Hamid Vahid, however, has paid explicit attention to
this argument. His response can be summarized as follows. First, on the basis of
an earlier article of Alston,9 he defines the deontological conception of epistemic
justification as follows:
S is justified in believing that p iff (i) S believes that p, (ii) S believes

(4)
that she has adequate evidence for p, and (iii) S’s belief that she has
adequate evidence for p is based on adequate grounds.10
Subsequently, he convincingly argues that by definition, (4) does not entail truth-
conduciveness. For, clearly, one can have adequate evidence that one has adequate
6. I think (2) and (3) are somewhat problematic; few externalists and internalists would
be willing to embrace these accounts as they stand. My criticisms, however, do not
hinge on this.
7. For the entire argument, see Alston (1989d, 145–152). Elsewhere, I have sketched the
examples in detail; see Peels (2016a).
8. F or the entire argument, see Alston (1989d, 145–152).
9. See Alston (1989b).
10. S ee Vahid (1998, 289). My (4) is a fully spelled-out version of his (DJd).
239
240
A ppendix
evidence for p, even if one’s evidence for p is in fact inadequate (adequacy after all,
does not entail truth). But why, Vahid asks, would that count against (4)? That
seems to be the case only if we assume that truth-conduciveness is an essential
part of epistemic justification and that seems to beg the question. What we rather
seem to have in epistemology is a variety of different but equally valuable concep-
tions of epistemic justification. Trying to reduce this plurality is a project that is
unlikely to succeed.
For two reasons, Vahid’s response to Alston does not seem fully satisfying to
me. First, Vahid’s response may or may not save (4) from Alston’s argument. We
should remember, however, that Alston’s argument is explicitly directed against
a deontological conception of epistemic justification that is spelled out in terms
of influence and (4), in opposition to (1), is not.11 For all we know, therefore, (4),
even if it is truly deontological, falls prey to the argument from doxastic invol-
untarism that I presented in chapter 2. Of course, that as such does not count
against Vahid’s criticism of (4). But it does mean that Vahid’s criticism may miss
the target, which is (1)—a n account of epistemic justification in terms of doxastic
influence, one that is crucially different from (4).
Second, Alston’s point is not merely that (1) can come apart from truth-
conduciveness, but that it gives different verdicts from those given by the main
externalist and internalist conceptions of justification. That (1) can come apart
from truth-conduciveness is only one of the four arguments that he provides by
comparing (1) on the one hand with (2) and (3) on the other—above, I men-
tioned (b)–(d) in addition to (a).
Nevertheless, I believe that there is a grain of truth in what Vahid says, viz.,
that Alston’s argument against (1) displays a certain arbitrariness. I will explain
this below.
Alston seems to acknowledge that in order to answer the question of whether
(1) counts as a viable conception of epistemic justification, we need some way of
identifying what we have in mind by being “epistemically justified” in holding
certain beliefs. In addition to explaining “epistemic” in terms of the Jamesian goal
of having true rather than false beliefs, Alston provides the required specification
by assuming that either (2) or (3) gives the correct account of epistemic justifica-
tion. It seems to me that this strategy begs the question. For the deontological
conception of epistemic justification, as spelled out in (1), can plausibly be inter-
preted as a rival view of what it is to be epistemically justified in having some
particular belief. Depending on whether the intellectual obligations we are pre-
sumed to have are spelled out externalistically or internalistically—i n chapter 3,
I argued that most of our epistemic intellectual obligations should be understood
internalistically—(1) is a specific externalistic or internalistic account of epis-
temic justification that is to be considered as an alternative to (2) and (3). That
11. S ee Alston (1989d, 143).
240
241
A ppendix
(1) does not match (2) or (3) nor the main intuitions about epistemic justifica-
tion of those who embrace (2) or (3)—such as that an epistemically justified
belief should be based on a truth-conducive ground or that it should be based on
a ground that is truth-conducive from the subject’s perspective—does not count
against (1).
This point is strengthened by the fact that something similar seems to apply
to certain other accounts of epistemic justification that we find in the literature.
Take Richard Feldman’s and Earl Conee’s evidentialist account of epistemic
justification:
S is justified in believing that p iff believing p fits S’s evidence.

(5)
They stress that this does not require that S bases her belief that p on the evidence
that she has that fits her belief that p—t hat, they say, would be an instance of a
well-founded belief, which is stronger than epistemically justified belief. Nor is it
required that from S’s perspective her belief is based on an adequate ground—t he
cognitive subject in question may have no opinion on that matter, not even dispo-
sitionally. All one needs is that one’s belief does fit one’s evidence.12 Like (1), (5) is
best interpreted as a rival account of epistemic justification. That it gives verdicts
different from those given by (2) and (3), therefore, does not count against it. But
then it also does not count against (1) or any other account of epistemic justifica-
tion in terms of our influence on our beliefs that it gives different verdicts from
those given by (2) and (3).
Is there anything that can be said in Alston’s defense? Two things come to
mind. First, one might point out that Alston has also criticized the deontological
conception of epistemic justification elsewhere, namely in his piece “Concepts
of Epistemic Justification.” This is true, but we should notice that in that article
he rejects this conception because meeting something like (1) does not render it
likely that one’s belief is true. According to Alston, only a belief that is likely to be
true can count as epistemically justified.13 And, as I said, this begs the question,
for adherents of the deontological conception might very well maintain that it is
a rival conception of epistemic justification and that likelihood of truth is not a
necessary condition for epistemic justification. Moreover, we should notice that
an internalist conception of epistemic justification à la (3) also does not render it
likely that one’s belief is true. For it may well be that from one’s perspective, one’s
belief that p is based on an adequate ground (a ground that renders it likely that
one’s belief is true), while in fact it is not. It is even possible that one believes on
the basis of an adequate ground that one’s belief that p is based on an adequate
12. S ee Conee and Feldman (2004).

13. S ee Alston (1989b, 84).
241
242
A ppendix
ground, even though it is not (for being adequately grounded does not imply
being true).
Second, Alston himself has defended a particular conception of epistemic
justification—h is so-called internalist externalism—a nd that conception does
not match (1).14 However, in defending that position, Alston again assumes that
a belief is epistemically justified only if it is based on a ground that is favorable
relative to the aim of having true rather than false beliefs, which he explains as: “it
must be sufficiently indicative of the truth of the belief it grounds. In other terms,
the ground must be such that the probability of the belief ’s being true, given that
ground, is very high.”15 As I said, however, that begs the question. For it more
or less straightforwardly implies that any deontological conception of epistemic
justification is mistaken. And it also implies that any internalist conception that
resembles (3) is untenable. Alston, therefore, fails to provide a theoretically neu-
tral criterion—t hat is, a criterion that does not depend on a particular analysis
of epistemic justification—t hat an account of epistemic justification must meet.
Is there another criterion? As I said, it seems that that criterion is not pro-
vided by (2) and (3). Also, ordinary language cannot guide us here, for outside of
philosophical discourse, people hardly ever talk about beliefs being epistemically
justified. Are there any other criteria? I can think of three such criteria. I would
like to emphasize that I do not contend that we should espouse these criteria for
evaluating accounts of epistemic justification. Some of them might even be to
some extent question-begging—I will not take a stance on that. What I would like
to consider is whether (1) can meet these criteria, whether or not they are correct.
If the opponent of the deontological conception of epistemic justification finds
none of them convincing, I challenge him or her to come up with a more convinc-
ing criterion for epistemic justification.
CRITERION 1: SHOWING A CONCERN

WITH THE JAMESIAN GOAL
First, one could stress Alston’s point that a concept of epistemic justification
should manifest a certain concern with the Jamesian goal of having true rather
than false beliefs and suggest that (1) fails to do so, since epistemic justification
as spelled out in (1) does not guarantee true belief nor does it make true belief
likely.16
14. S ee Alston (1989a).

15. A lston (1989a, 232).
16. Th is view is widely shared, as rightly noticed by Riggs (2003, 342–3 45).
242
243
A ppendix
My reply is twofold. First, why should one think that showing a concern with
the Jamesian goal of having true rather than false beliefs should guarantee or
make it likely that one holds true beliefs? One of Alston’s own examples can help
us to see why that does not seem necessary. A student who is cognitively unable
to grasp an argument of Locke may be deontologically justified in believing that,
according to Locke, everything is a matter of opinion, although, given his cogni-
tive capacities, he is unlikely to form true beliefs on these matters. Thus, even
though he might be justified on (1), his belief on Locke’s philosophy is neither
true nor likely to be true. Still, it seems the student, if he fulfills his intellectual
obligations, such as gathering the relevant evidence by reading Locke’s books and
critically reflecting on them, thereby displays a concern with the Jamesian goal.
His aim, after all, is to find out the truth—i.e., hold a true belief—about Locke’s
philosophy. Thus, the deontological conception of epistemic justification as
spelled out in (1) does show a concern with the Jamesian goal of having true rather
than false beliefs.
Second, there is more to be said about the relation between (1) and the
Jamesian goal. In order to see that, let us first turn to (2) and (3). How do these
concepts of epistemic justification satisfy the requirement of showing a concern
with the Jamesian goal? (2) does so in virtue of the fact that, by definition, a justi-
fied belief is one that is likely to be true. For on (2) a justified belief is one that is
based on an adequate ground, and a ground is adequate only if it is such that basing
a belief on that ground makes it likely that that belief is true. Notice that it is not
required that one’s ground is fully adequate, that is, that it entails that one’s belief
is true. It is widely agreed that false beliefs can be justified. What (2) says is that
a justified belief B is a belief based on a ground G such that in a sufficiently large
portion of cases (presumably, across possible worlds) in which B is based on G, B
is true. Therefore, the relation between justification and the Jamesian goal of hav-
ing true rather than false beliefs that is implicit in (2) is that of rendering attaining
it likely.
As to (3), the fact that from S’s perspective B is based on an adequate ground
G clearly does not always render it likely that B is true rather than false. G may in
fact be wholly inadequate, so that, no matter what S believes about G, it remains
highly unlikely that in having B she reaches the Jamesian goal. Here, Alston does
not take this to count against anything like (3) and that, I think, is because there
is some important relation between (3) and the Jamesian goal. For most of the
time we do have an accurate idea of whether or not our beliefs are based on an
adequate ground. We know that believing something on the basis of an encyclope-
dia article is a generally reliable way of forming beliefs, whereas hallucination and
tarot reading are generally inadequate grounds to base one’s beliefs on. The rela-
tion between (3) and the Jamesian goal, therefore, amounts to something along
the following lines: if a belief B is justified on (3), then, generally it is likely that B
is true.
243
244
A ppendix
Finally, how should we think of the relation between (1) and the Jamesian
goal? Clearly, gathering more evidence, reflecting on one’s beliefs, and exemplify-
ing particular intellectual virtues, such as thoroughness and open-m indedness,
generally render it likely that one reaches the Jamesian goal concerning some par-
ticular proposition. Of course, this will not always be the case. Alston gives a cou-
ple of examples that nicely illustrate this point. I already mentioned the student
who reads Locke, but is cognitively unable to grasp his views. Two other exam-
ples that he gives are the following. A tribesman may be deontologically justified
in believing in witchcraft, but since his belief is based on radically inadequate
grounds, it is not likely to be true. And someone may be deontologically justified
in believing someone’s testimony, even if that person is utterly unreliable, so that
basing one’s beliefs on his testimony does not render it likely that they are true.17
In these cases one’s meeting rather than violating one’s (epistemic) intellectual
obligations does not make it any more likely that one reaches the Jamesian goal.
I submit that it will nevertheless be true that generally—in a large portion
of cases—meeting one’s (epistemic) intellectual obligations is a reliable means
of reaching the Jamesian goal concerning the propositions in question. Thus, if
the tribesman had not suffered from cultural isolation, he most likely would not
have believed in witchcraft. If he still had, his belief probably would have been
blameworthy and, thus, would not have been justified on (1). If the student had
not been cognitively limited, he most likely would not have believed that, accord-
ing to Locke, everything is a matter of opinion. If he still had, his belief probably
would have been blameworthy and, thus, not justified on (1). Generally, it seems,
deontological justification renders it likely that one attains the Jamesian goal. In
this regard, doxastic responsibility is not any different from responsibility for our
actions. Surgeons may act responsibly by doing what seems best to them in situ-
ations that are medically so complicated that they are highly unlikely to perform
that action which is most beneficial. Politicians may act responsibly in taking
certain measures that seem to them the best ones available, although, given the
complicated economic circumstances, they are unlikely to provide a good solu-
tion to the problem at hand. From such exceptional situations it does not follow
that, generally, to perform that action which seems best to one is not a reliable way
of bringing about a state of affairs that is objectively good (beneficial). If circum-
stances are sufficiently favorable, acting responsibly and believing responsibly are
reliable ways of reaching the objective good (the moral good, the truth, etc.).18
17. See Alston (1989b, 95–96; 1989d, 145–149). There has been some discussion about
whether the cognitive subjects in these examples, as described by Alston, are in fact
blameless for their beliefs. See, for instance, Steup (1988, 78–79). For present purposes
I assume that the examples do indeed show that there are or could be subjects who are
deontologically justified in holding certain beliefs, although their beliefs are not justi-
fied on (2) and (3).
18. Thus also Heil (1983, 362–363).
244
245
A ppendix
This relationship between (1) and the Jamesian goal is significant for the issue
at hand, for it means that if one pursues the Jamesian goal, then one should gen-
erally value a belief ’s being justified on (1). Again, the relation between (1) and
the Jamesian goal is a complicated one rather than the straightforward relation of
entailment. But so are the relations between (2) and (3) on the one hand and the
Jamesian goal on the other. (1) and (3) are especially similar in that they render
the attainment of the Jamesian goal likely only if certain conditions are met.
One may worry that, although (1) shows some concern with the Jamesian
goal of having true rather than false beliefs, beliefs that satisfy (1) need not count
as epistemically rather than practically (morally, prudentially) justified beliefs.
Imagine that one could take a pill that induces true beliefs on some topic. It seems
that (1) implies that the beliefs that issue from such an action are epistemically
rather than pragmatically justified.
In response, let me point out that the adherent of (1) is not committed to the
idea that we have an intellectual obligation to perform just any action that some-
how favors the Jamesian goal, let alone an epistemic intellectual obligation to do
so. In c hapter 3, I provided an account of which epistemic intellectual obligations
we have. Swallowing a pill that induces true beliefs was clearly not among them.
CRITERION 2: BEING ANALOGOUS

TO JUSTIFICATION IN OTHER REALMS
A second criterion for judging a conception of epistemic justification is that it
should be relevantly analogous to conceptions of justification in other domains,
such as that of action. For one might think that it is implausible that when it
comes to actions justification means one thing, whereas when it comes to beliefs
it amounts to something completely different.
Let us assume that this is correct. What important characteristics of justified
actions can we then distinguish? I think there are at least two such characteris-
tics. First, for an action to be justified is for it to be such that in performing that
action, one does not violate any rules, duties, or obligations—whether those rules
are legal, moral, or prudential. An action’s being justified is, therefore, a matter of
its being permitted by the relevant rules of principles. Second, an action’s being
justified implies that one is blameless for it—at least, if one is justified because one
has violated neither objective nor subjective obligations. If Jim was justified in
pulling the trigger, then it is inappropriate for someone fully informed about the
situation to blame him for that.
Let us now see whether (1) satisfies these two requirements. First, on (1),
there are certain rules, duties, or obligations that I am subject to, viz., intellectual
obligations, and those obligations are relevantly related to my beliefs. In fact,
245
246
A ppendix
it is clear that on (1), I justifiedly believe that p if and only if my believing that
p is permitted because I have not violated any relevant obligations. Second, as
I said above, the relevant intellectual obligations can be spelled out in different
ways, externalistically or internalistically. In c hapter 3, I argued that we have
both objective and subjective intellectual obligations, and in c hapter 5, I argued
that whether or not one believes responsibly is a radically subjective matter in
the sense that it crucially depends on the subject’s occurrent, dormant, and tacit
beliefs.
What is important for the present point, though, is that it seems that these
intellectual obligations can be spelled out in such a way that one’s being subject
to them renders one responsible for (not) meeting them and for the consequences
of (not) meeting them, so that one is blameworthy if one fails to meet them or if
something is a consequence of failing to meet them. I conclude that (1) also meets
this second standard of judging a concept of epistemic justification for beliefs.19
Whether (2) and (3) meet this requirement is something that I will not discuss
here. For all that I wanted to show is that if one demands that epistemic justifica-
tion for beliefs be relevantly analogous to justification for actions, (1) meets this
requirement.
Thus, (1), properly understood, might very well meet the criterion of being
relevantly analogous to justification in other realms in that it satisfies two condi-
tions that are considered necessary for justification, namely permission by the
relevant rules and blamelessness.
THIRD CRITERION: BEING NECESSARY

FOR KNOWLEDGE
Third, one could claim that the aim of epistemology is to give an account of knowl-
edge and that epistemic justification should be necessary for knowledge in order
to be relevant to epistemological concerns. My response to this line of reasoning
is threefold.
First, I agree that an important aim of epistemology is to analyze knowledge,
but I see no reason to think that that is its sole aim. Recent literature in epistemol-
ogy shows an interest in all sorts of phenomena that do not directly have to do with
knowledge, such as the aim of belief, understanding, memory, trust, testimony,
hope, faith, assuming, acceptance, intellectual virtues like open-m indedness and
19. A nd it seems that Alston, given his acknowledgment that the term “justified” is most
naturally understood deontologically, both with respect to actions and with respect to
beliefs, would have to agree with this; see Alston (1989d, 115–116, 143).
246
247
A ppendix
thoroughness, and intellectual vices like narrow-m indedness and intellectual

laziness.
Second, several epistemologists take it that there is such a thing as epistemic
justification that is not necessary for knowledge and that should nevertheless be
analyzed by epistemologists. 20 The two criteria provided above and particular
examples of what one takes to be cases of justified belief could perhaps serve to
identify this kind of epistemic justification.
Third, even if epistemic justification is necessary for knowledge, there is some
reason to think that (1) might be able to meet this standard. Imagine that I believe
a true proposition p on some evidence base E and that I have an epistemic intellec-
tual obligation to gather evidence E* concerning p, as I characterized such obliga-
tions in chapter 3. Imagine also that if I were to gather E*, my total evidence base
would change in such a way that I would no longer believe that p. Imagine, finally,
that I am lazy and, therefore, culpably violate my intellectual obligation to gather
E*. Then, do I know that p? It is not clear that I do. In nearby possible worlds, in
which I meet my obligation, I would not believe that p. If my belief is that acciden-
tal, one might doubt whether my belief could count as an instance of knowledge.
The same seems true for intellectual obligations with respect to the function-
ing of my doxastic mechanisms. Imagine that I believe that there is a mountain
goat at the top of the hill. Imagine also that I have an intellectual obligation to
squint in such cases, since I correctly believe that I have often been mistaken
about things like these in the past and that squinting makes my perceptual beliefs
more likely to be true in such circumstances. Imagine, finally, that if I were to
squint, I would come to believe that it is a heap of snow rather than a mountain
goat. Then, if I am lazy and do not squint, it is not clear that I know that there
is a mountain goat, even if there happens to be one. And it seems that mutatis
mutandis the same could be said about all other intellectual obligations that we
are subject to. 21
There is an example, concocted by Alvin Plantinga, which is meant to refute
the claim that deontological justification, as, for instance, spelled out in (1), is nec-
essary for knowledge, so let me discuss that example here. Imagine that I falsely,
but blamelessly, believe that any time I seem to see something red, there really is
no red object present. Every time I seem to see something red I remind myself of
this and disbelieve that I see something red, but one time, being tired, upon see-
ing a red London bus, I give in and blameworthily believe that there is a red bus
in front of me. According to Plantinga, this belief is epistemically blameworthy,
20. A lston himself is an adherent of this view. See Alston (1989c, 172–182; 1989d, 144).
Audi (2011, 270–2 82) and Foley (2005a) also defend this view.
21. I n fact, we can find this idea in the literature. According to Ernest Sosa, for instance,
one can lose knowledge by diachronic epistemic negligence, such as being dogmatically
narrow-m inded or failing to gather further evidence; see Sosa (2014, 81, 86).
247
248
A ppendix
because it does not satisfy (1), but nevertheless counts as knowledge.22 I agree
that such a belief might be blameworthy on (1)—t he case does not seem to be
described in sufficient detail to be sure—but I have to admit that I fail to see why
it would count as an instance of knowledge. For if I had so much as thought about
whether my grounds to believe that there is a red bus in front of me are adequate,
I would have been aware of the defeater. And being aware of the defeater, I would
have withheld this belief. In a wide class of close possible worlds, therefore, I lack
this belief. But if my belief is that accidental, it is hard to see how it could count as
an instance of knowledge.
Let me stress that this is plausible only if we understand the intellectual
obligations mentioned in (1) as epistemic intellectual obligations. Imagine, for
instance, that I have a moral intellectual obligation not to gather any evidence on
the love affairs of my niece. If I nevertheless culpably spy on my niece, I may very
well be blameworthy for the ensuing beliefs. Obviously, it does not follow that
those beliefs do not count as instances of knowledge. 23 It is crucial, therefore, that
we understand the intellectual obligations involved in (1) as epistemic intellectual
obligations, that is, obligations that in some sense “aim” at the Jamesian goal of
having true rather than false beliefs (in c hapter 3, I have spelled out what aiming
at the Jamesian goal could amount to).
One may wonder whether epistemic blameworthiness is blameworthiness in
the full-blooded deontological sense. After all, as I argued in chapters 4 and 5,
in daily life we are often excused and, hence, blameless for not meeting our epis-
temic intellectual obligations by moral and prudential factors. One may think,
therefore, that in the strict sense of the word, there is no such thing as purely
epistemic blameworthiness. Here, I need not take a stance on this thorny issue.
Perhaps (1) concerns only epistemic blameworthiness in this contrived sense.
Fortunately, that would not count against (1) as a conception of epistemic justifi-
cation. For what I have tried to show here is that there is a plausible interpretation
of (1) on which it can count as a necessary condition for knowledge.
A lot will depend here on precisely how the epistemic intellectual obliga-
tions to perform belief-i nfluencing actions are spelled out. Imagine, for instance,
that S believes that p on some evidence base E. And imagine that more investiga-
tion would turn up counter-evidence E*, so that S would no longer believe that p.
Imagine finally, however, that S has no good reason to think that further inves-
tigation will turn up such counterevidence. It seems that in this situation S may
very well know that p, although S could have gathered further evidence on p such
that if S had done so, S would no longer have believed that p. Adherents of (1),
22. See Plantinga (1993, 45). For a highly similar example, see Alston (1989c, 179).
23. Th at the violation of moral obligations does not entail the absence of knowledge has
been rightly pointed out by Bergmann (2000, 93). Only, his view is cast in terms of
doxastic rather than intellectual obligations.
248
249
A ppendix
therefore, should not contend that we have an obligation always to gather further
evidence on any proposition and always to work on our intellectual virtues and
vices. But there is no reason to think that the adherent of (1) is committed to such
a strong view on our intellectual obligations. If my account of our intellectual
obligations that I provided in c hapter 3 is correct, for instance, then the scope of
our intellectual obligations is much more restricted.
Yet I think that (1) is not necessary for knowledge, for at least two reasons.
First, condition (iii) of (1) says that there is or was some belief-i nfluencing action
or series of actions that one could have performed such that if one had done
that, one would not have believed that p. I defended this condition in detail in
chapter 4: responsible belief requires the ability to believe otherwise. However,
there are many propositions that I know without being able to believe otherwise.
I know that I exist, that there is an external material world, and that 2 + 3 = 5, even
if I could not have failed to believe these things. It follows that I can know that p
and, hence, be epistemically justified in believing that p (again, if epistemic jus-
tification is necessary for knowledge), even if I do not epistemically responsibly
believe that p because I could not have believed otherwise. What the adherent
of the deontological conception of epistemic justification will, therefore, have to
say is that epistemically justified belief is epistemically blameless rather than epis-
temically responsible belief.
Second, clause (v) of (1), concerning excuses, also causes trouble. As the
example of the red bus that I discussed above illustrates, if I have violated cer-
tain epistemic intellectual obligations in coming to believe that p, then it seems
that I do not know that p. But I can still epistemically responsibly believe that
p, namely if I am excused for the violation of those epistemic intellectual obliga-
tions. Hence, what is necessary for knowledge is not just epistemically blameless
belief, but belief that is epistemically blameless because one has not violated any
epistemic intellectual obligations in coming to hold or in maintaining that belief
rather than because one is excused for having violated some epistemic intellectual
obligations in coming to hold or in maintaining that belief.
This means that if epistemic justification as defined by (1) is to count as a
necessary condition for knowledge, it will have to be qualified in at least these
two regards.
In summary, we have seen that Alston’s arguments against (1) crucially beg
the question, since his criterion for identifying epistemic justification fails to be
theoretically neutral. Of course, the other three criteria that I mentioned may
not be entirely theoretically neutral either. The second criterion, for instance,
suggests that (2) and (3) are not viable as conceptions of epistemic justification.
However, these three criteria are at least theoretically neutral in that they do not
commit one to one or another of two particular accounts of epistemic justifica-
tion. I have argued that (1) meets two standards that one might come up with and
that are theoretically neutral in this sense. First, it manifests a concern with the
249
250
A ppendix
Jamesian goal of having true rather than false beliefs and, second, it can explain
how justified actions and justified beliefs are relevantly analogous. Epistemically
responsible belief, as defined by (1), is not necessary for knowledge. However, for
all we have seen here, something close to it is necessary for knowledge, namely
belief that is epistemically blameless because one has not violated any epistemic
intellectual obligations in coming to hold it or in maintaining it. Hence, even
though this does not show that epistemic justification should be understood
deontologically, it provides us with enough reason to treat a sophisticated deon-
tological conception of epistemically justified belief as epistemically responsible
belief as a serious rival of certain externalist and internalist accounts of epistemic
justification. For, something very close to responsible belief may well be identical
to epistemically justified belief.
250
251
BIBLIOGRAPHY
Adams, Robert M. (1985). “Involuntary Sins,” Philosophical Review 94.1, 3–31.

——— . (1999). Finite and Infinite Goods: A Framework for Ethics (Oxford: Oxford
University Press).
Adler, Jonathan E. (1987). “Luckless Desert Is Different Desert,” Mind 96.382,
247–2 49.
———. (2002). Belief ’s Own Ethics (Cambridge, Mass.: MIT Press).
Alston, William P. (1989a). “An Internalist Externalism,” in Epistemic
Justification: Essays in the Theory of Knowledge (Ithaca: Cornell University
Press), 227–2 45.
—— —. (1989b). “Concepts of Epistemic Justification,” in Epistemic
Justification: Essays in the Theory of Knowledge (Ithaca: Cornell University
Press), 81–114.
———. (1989c). “Justification and Knowledge,” in Epistemic Justification: Essays
in the Theory of Knowledge (Ithaca: Cornell University Press), 172–182.
———. (1989d). “The Deontological Conception of Epistemic Justification,”
in Epistemic Justification: Essays in the Theory of Knowledge (Ithaca: Cornell
University Press), 115–152.
———. (1991). Perceiving God: The Epistemology of Religious Experience
(Ithaca: Cornell University Press).
—— —. (1996). “Belief, Acceptance, and Religious Faith,” in Jeff Jordan
(ed.), Faith, Freedom, and Rationality: Philosophy of Religion Today
(Lanham: Rowman and Littlefield), 3–27.
———. (2005). Beyond “Justification”: Dimensions of Epistemic Evaluation
(Ithaca: Cornell University Press).
251
252
B ibliography
Andre, Judith. (1993). “Nagel, Williams, and Moral Luck,” in Daniel Statman
(ed.), Moral Luck (Albany: State University of New York Press), 123–129.
Anscombe, Gertrude E.M. (1963a). Intention, 2nd ed. (Oxford: Blackwell).
———. (1963b). “The Two Kinds of Error in Action,” The Journal of Philosophy
60.14, 393–4 01.
Aristotle. (2003). The Nicomachean Ethics, transl. H. Rackham (Cambridge,
Mass.: Harvard University Press).
Athanassoulis, Nafsika. (2005). Morality, Moral Luck, and Responsibility: Fortune’s
Web (New York: Palgrave Macmillan).
Audi, Robert. (1979). “Weakness of Will and Practical Judgment,” Noûs 13.2,
173–196.
———. (1986). “Acting for Reasons,” Philosophical Review 95.4, 511–5 46.
———. (1994). “Dispositional Beliefs and Dispositions to Believe,” Noûs 28.4,
419–434.
———. (2001). “Doxastic Involuntarism and the Ethics of Belief,” in Matthias
Steup (ed.), Knowledge, Truth, and Duty: Essays on Epistemic Justification,
Responsibility, and Virtue (Oxford: Oxford University Press), 93–111.
———. (2011). Epistemology: A Contemporary Introduction to the Theory of
Knowledge, 3rd ed. (New York/L ondon: Routledge).
Austin, John L. (1979). “A Plea for Excuses,” in James O. Urmson and Geoffrey
J. Warnock (eds.), Philosophical Papers, 3rd ed. (Oxford: Clarendon Press),
175–2 04.
Axtell, Guy. (2011a). “From Internalist Evidentialism to Virtue Responsibilism,”
in Trent Dougherty (ed.), Evidentialism and Its Discontents (Oxford: Oxford
———. (2011b). “Recovering Responsibility,” Logos and Episteme 2.3, 429–454.
———. (2012). “(More) Springs of My Discontent,” Logos and Episteme 3.1,
131–137.
Beardsley, Elizabeth L. (1979). “Blaming,” Philosophia 8.4, 573–583.
Bennett, Jonathan. (1980). “Accountability,” in Zak Van Straaten (ed.),
Philosophical Subjects: Essays Presented to P.F. Strawson (Oxford: Clarendon
Press), 14–47.
———. (1995). The Act Itself (Oxford: Clarendon Press).
Bergmann, Michael. (2000). “Deontology and Defeat,” Philosophy and
Phenomenological Research 60.1, 87–102.
Bogdan, Radu J. (1986). “The Manufacture of Belief,” in Radu J. Bogdan (ed.),
Belief: Form, Content, and Function (Oxford: Clarendon Press), 149–184.
BonJour, Laurence. (1980). “Externalist Theories of Empirical Knowledge,”
in Peter A. French, Theodore E. UelingJr., and Howard K. Wettstein
(eds.), Midwest Studies in Philosophy, Vol. V: Studies in Epistemology
(Minneapolis: University of Minnesota Press), 53–73.
252
253
B ibliography
———. (1985). The Structure of Empirical Knowledge (Cambridge, Mass.: Harvard

University Press).
Booth, Anthony R. (2006). “Can There Be Epistemic Reasons for Action?,”
Grazer Philosophische Studien 73.1, 133–144.
———. (2009a). “Compatibilism and Free Belief,” Philosophical Papers 38.1, 1–12.
—— —. (2009b). “Motivating Epistemic Reasons for Action,” Grazer
Philosophische Studien 78.1, 265–271.
———. (2012). “All Things Considered Duties to Believe,” Synthese 187.2,
509–517.
———. (2014). “On Some Recent Moves in Defence of Doxastic Compatibilism,”
Synthese 191.8, 1867–1880.
———. (unpublished). “Doxastic Influence and Epistemic Duty.”
Booth, Anthony R., and Rik Peels. (2010). “Why Responsible Belief Is Blameless
Belief,” The Journal of Philosophy 107.5, 257–2 65.
———. (2012). “Epistemic Justification, Rights, and Permissibility,” Logos and
Episteme 3.3, 405–411.
Botterell, Andrew. (2009). “A Primer on the Distinction between Justification
and Excuse,” Philosophy Compass 4.1, 172–196.
Braithwaite, Richard B. (1967). “The Nature of Believing,” in A. Phillips Griffiths
(ed.), Knowledge and Belief (Oxford: Oxford University Press), 28–4 0.
Brandt, Richard B. (1958). “Blameworthiness and Obligation,” in Abraham I.
Melden (ed.), Essays in Moral Philosophy (Seattle: University of Washington
Press), 3–39.
———. (1959). Ethical Theory: The Problems of Normative and Critical Ethics
(Englewood Cliffs, N.J.: Prentice-Hall).
———. (1969). “A Utilitarian Theory of Excuses,” Philosophical Review 78.3,
337–361.
Browne, Brynmor. (1992). “A Solution to the Problem of Moral Luck,” The
Philosophical Quarterly 42.168, 345–356.
Buckareff, Andrei A. (2004). “Acceptance and Deciding to Believe,” Journal of
Philosophical Research 29, 173–190.
———. (2014). “Deciding to Believe Redux,” in Jonathan Matheson and Rico Vitz
(eds.), The Ethics of Belief: Individual and Social (Oxford: Oxford University
Press), 33–50.
Burrington, Dale E. (1999). “Blameworthiness,” Journal of Philosophical Research
24, 505–527.
Card, Claudia. (1990). “Gender and Moral Luck,” in Owen Flanagan and Amelie
Rorty (eds.), Identity, Character, and Morality (Cambridge, Mass.: MIT
Press).
Chisholm, Roderick M. (1977). Theory of Knowledge, 2nd edn. (Englewood Cliffs,
N.J.: Prentice-Hall).
253
254
B ibliography
———. (1991). “Firth and the Ethics of Belief,” Philosophy and Phenomenological
Research 51.1, 119–128.
Chrisman, Matthew. (2008). “Ought to Believe,” The Journal of Philosophy 105.7,
346–370.
Chuard, Philippe, and Nicholas Southwood. (2009). “Epistemic Norms without
Voluntary Control,” Noûs 43.4, 599–632.
Church, Jennifer. (2002). “Taking It to Heart: What Choice Do We Have?,” The
Monist 85.3, 361–380.
Clarke, Murray. (1986). “Doxastic Voluntarism and Forced Belief,” Philosophical
Studies 50.1, 39–51.
Clifford, William K. (1901). “The Ethics of Belief,” in Lectures and Essays
(London: Macmillan and Co), 163–2 05.
Cohen, L. Jonathan. (1992). An Essay on Belief and Acceptance (Oxford: Clarendon
Press).
Conee, Earl, and Richard Feldman. (2004). Evidentialism: Essays in Epistemology
(Oxford: Clarendon Press).
Corlett, J. Angelo. (2008). “Epistemic Responsibility,” International Journal of
Philosophical Studies 16.2, 179–2 00.
Coyne, Margaret U. (1985). “Moral Luck?,” The Journal of Value Inquiry 19.4,
319–325.
David, Marian. (2005). “Truth as the Primary Epistemic Goal: A Working
Hypothesis,” in Matthias Steup and Ernest Sosa (eds.), Contemporary Debates
in Epistemology (Oxford: Blackwell), 296–312.
De Ridder, Jeroen. (2013). “Epistemic Dependence and Collective Scientific
Knowledge,” Synthese 191.1, 1–17.
Donagan, Alan. (1977). The Theory of Morality (Chicago: University of Chicago
Press).
Dougherty, Trent. (2011). “Re-R educing Responsibility: Reply to Axtell,” Logos
and Episteme 2.4, 625–632.
———. (2012a). “Internalist Evidentialism and Epistemic Virtue: Re-R eply to
Axtell,” Logos and Episteme 3.2, 281–2 89.
—— —. (2012b). “Reducing Responsibility: An Evidentialist Account of
Epistemic Blame,” European Journal of Philosophy 20.4, 534–5 47.
—— —. (2014). “The ‘Ethics of Belief ’ is Ethics (Period): Reassigning
Responsibilism,” in Jonathan Matheson and Rico Vitz (eds.), The Ethics of
Belief: Individual and Social (Oxford: Oxford University Press), 146–166.
Dretske, Fred. (2000). “Entitlement: Epistemic Rights Without Epistemic
Duties?,” Philosophy and Phenomenological Research 60.3, 591–6 06.
Driver, Julia. (1989). “The Virtues of Ignorance,” The Journal of Philosophy 86.7,
373–384.
Engel, Meylan. (1992). “Is Epistemic Luck Compatible with Knowledge?,”
Southern Journal of Philosophy 30.2, 59–75.
254
255
B ibliography
Engel, Pascal. (2009). “Epistemic Responsibility without Epistemic Agency,”

Philosophical Explorations 12.2, 205–219.
Feinberg, Joel. (1962). “Problematic Responsibility in Law and Morals,”
Philosophical Review 71.3, 340–351.
———. (1970). Doing and Deserving: Essays in the Theory of Responsibility
(Princeton: Princeton University Press).
Feldman, Richard. (1988a). “Epistemic Obligations,” in James E. Tomberlin
(ed.), Philosophical Perspectives 2: Epistemology (Atascadero, Cal.: Ridgeview
Publishers), 235–256.
—— — . (1988b). “Subjective and Objective Justification in Ethics and
Epistemology,” The Monist 71.3, 405–419.
———. (2000). “The Ethics of Belief,” Philosophy and Phenomenological Research
60.3, 667–695.
—— — . (2001). “Voluntary Belief and Epistemic Evaluation,” in Matthias
—— — . (2008). “Modest Deontologism in Epistemology,” Synthese 161.3,
339–355.
Fields, Lloyd. (1991). “A Moral Basis of Excuses,” International Journal of Applied
Philosophy 6.1, 11–2 0.
—— — . (1994). “Moral Beliefs and Blameworthiness,” Philosophy 69.270,
397– 415.
Fischer, John M. (1985–1986). “Responsibility and Failure,” Proceedings of the
Aristotelian Society 86, 251–270.
———. (1994). The Metaphysics of Free Will: An Essay on Control (Oxford: Blackwell).
Fischer, John M., and Mark Ravizza. (1991). “Responsibility and Inevitability,”
Ethics 101.2, 258–278.
———. (1992). “Responsibility, Freedom, and Reason,” Ethics 102.2, 368–389.
———. (1993a). “Introduction,” in John M. Fischer and Mark Ravizza (eds.),
Perspectives on Moral Responsibility (Ithaca: Cornell University Press), 1–41.
———. (1993b). “Responsibility for Consequences,” in John M. Fischer and
Mark Ravizza (eds.), Perspectives on Moral Responsibility (Ithaca: Cornell
University Press), 322–3 47.
———. (1998). Responsibility and Control: A Theory of Moral Responsibility
(Cambridge: Cambridge University Press).
FitzPatrick, William J. (2008). “Moral Responsibility and Normative Ignorance:
Answering a New Skeptical Challenge,” Ethics 118.4, 589–613.
Flanagan, Owen. (1990). “Virtue and Ignorance,” The Journal of Philosophy 87.8,
420– 428.
Foley, Richard. (1978). “Inferential Justification and the Infinite Regress,”
American Philosophical Quarterly 15.4, 311–316.
255
256
B ibliography
———. (2005a). “Justified Belief as Responsible Belief,” in Matthias Steup and

Ernest Sosa (eds.), Contemporary Debates in Epistemology (Oxford: Blackwell),
313–326.
———. (2005b). “Response to Wolterstorff,” in Matthias Steup and Ernest Sosa
(eds.), Contemporary Debates in Epistemology (Oxford: Blackwell), 338–3 41.
Frankfurt, Harry G. (1969). “Alternate Possibilities and Moral Responsibility,”
The Journal of Philosophy 66.23, 829–839.
———. (1993). “What We Are Morally Responsible For,” in John M. Fischer and
Mark Ravizza (eds.), Perspectives on Moral Responsibility (Ithaca: Cornell
Frankish, Keith. (2009). “Partial Belief and Flat-Out Belief,” in Franz Huber and
Christoph Schmidt-Petri (eds.), Degrees of Belief (Dordrecht: Springer), 75–93.
Friedman, Jane. (2013). “Suspended Judgment,” Philosophical Studies 162.2,
165–181.
Gardner, John. (2009). “The Logic of Excuses and the Rationality of Emotions,”
Journal of Value Inquiry 43.3, 315–338.
Gettier, Edmund L. (1963). “Is Justified True Belief Knowledge?,” Analysis 23.6,
121–123.
Gibbard, Allan. (1990). Wise Choices, Apt Feelings: A Theory of Normative Judgment
(Oxford: Clarendon Press).
Ginet, Carl. (2001). “Deciding to Believe,” in Matthias Steup (ed.), Knowledge,
Truth, and Duty: Essays on Epistemic Justification, Responsibility, and Virtue
(Oxford: Oxford University Press), 63–76.
Glover, Jonathan. (1970). Responsibility (London: Routledge and Kegan Paul).
Goldman, Alvin I. (1970). A Theory of Human Action (Princeton: Princeton
University Press).
———. (1978). “Epistemology and the Psychology of Belief,” The Monist 61.4,
525–535.
———. (1986). Epistemology and Cognition (Cambridge, Mass.: Harvard
University Press).
Goldman, Holly S. (1976). “Dated Rightness and Moral Imperfection,” The
Philosophical Review 85.4, 449–4 87.
Greco, John. (1995). “A Second Paradox Concerning Responsibility and Luck,”
Metaphilosophy 26.1–2 , 81–96.
———. (2003). “Virtue and Luck, Epistemic and Otherwise,” Metaphilosophy
34.3, 353–366.
Guerrero, Alexander A. (2007). “Don’t Know, Don’t Kill: Moral Ignorance,
Culpability, and Caution,” Philosophical Studies 136.1, 59–9 7.
Haack, Susan. (2001). ““The Ethics of Belief ” Reconsidered,” in Matthias
256
257
B ibliography
Haddock, Adrian, and Fiona Macpherson. (2008). “Introduction: Varieties

of Disjunctivism,” in Adrian Haddock and Fiona Macpherson (eds.),
Disjunctivism: Perception, Action, Knowledge (Oxford: Oxford University
Press), 1–2 4.
Haji, Ishtiyaque. (2000a). “Alternate Possibilities and Responsibility,” Journal of
Social Philosophy 31.3, 259–2 67.
———. (2000b). “Control Requirements for Moral Appraisals: An Asymmetry,”
The Journal of Ethics 4.4, 351–356.
Haksar, Vinit. (1964). “Aristotle and the Punishment of Psychopaths,” Philosophy
39.150, 323–3 40.
Hall, Richard J., and Charles R. Johnson. (1998). “The Epistemic Duty to Seek
More Evidence,” American Philosophical Quarterly 35.2, 129–139.
Hanna, Nathan. (2014). “Moral Luck Defended,” Noûs 48.4, 683–698.
Harper, William. (1996). “Knowledge and Luck,” The Southern Journal of
Philosophy 34.3, 273–2 83.
Hart, Herbert L.A. (1970). Punishment and Responsibility: Essays in the Philosophy
of Law (Oxford: Clarendon Press).
Hawthorne, John, and Jason Stanley. (2008). “Knowledge and Action,” The
Journal of Philosophy 105.10, 571–590.
Heil, John. (1983). “Doxastic Agency,” Philosophical Studies 43.3, 355–364.
———. (1984). “Doxastic Incontinence,” Mind 93.369, 56–70.
———. (1992). “Believing Reasonably,” Noûs 26.1, 47–61.
Heller, Mark. (2000). “Hobartian Voluntarism: Grounding a Deontological
Conception of Epistemic Justification,” Pacific Philosophical Quarterly 81.2,
130–141.
Helm, Paul. (1994). Belief Policies (Cambridge: Cambridge University Press).
Hetherington, Stephen. (2002). “Epistemic Responsibility: A Dilemma,” The
Monist 85.3, 398–414.
Hieronymi, Pamela. (2005). “The Wrong Kind of Reason,” The Journal of
Philosophy 102.9, 437–457.
———. (2006). “Controlling Attitudes,” Pacific Philosophical Quarterly 87.1, 45–74.
———. (2008). “Responsibility for Believing,” Synthese 161.3, 357–373.
Houlgate, Laurence D. (1968). “Knowledge and Responsibility,” American
Hurley, Susan L. (1993). “Justice without Constitutive Luck,” in A. Phillips
Griffiths (ed.), Ethics: Royal Institute of Philosophy Supplement 35
(Cambridge: Cambridge University Press), 179–212.
Huss, Brian. (2009). “Three Challenges (and Three Replies) to the Ethics of
Belief,” Synthese 168.2, 249–271.
Jäger, Christoph. (2004). “Epistemic Deontology, Doxastic Voluntarism, and the
Principle of Alternate Possibilities,” in Winfried Löffler and Paul Weingartner
257
258
B ibliography
(eds.), Knowledge and Belief: Proceedings of the 26th International Wittgenstein

Symposium (Vienna: öbv&hpt Verlagsgesellschaft), 217–227.
James, William J. (1979). “The Will to Believe,” in The Will to Believe and Other
Essays in Popular Philosophy (Cambridge, Mass.: Harvard University
Press), 13–33.
Jensen, Henning. (1993). “Morality and Luck,” in Daniel Statman (ed.), Moral
Luck (Albany: State University of New York Press), 131–140.
Joyce, G. H. (1914). “Invincible Ignorance,” in James Hastings (ed.), Encyclopaedia
of Religion and Ethics, Vol. 7 (Edinburgh: T. & T. Clark), 403–4 04.
Katzoff, Charlotte. (1996). “Epistemic Obligation and Rationality Constraints,”
The Southern Journal of Philosophy 34.4, 455–470.
———. (2000). “Counter-evidence and the Duty to Critically Reflect,” Analysis
60.1, 89–96.
Kavka, Gregory S. (1983). “The Toxin Puzzle,” Analysis 43.1, 33–36.
Kelly, Thomas. (2002). “The Rationality of Belief and Some Other Propositional
Attitudes,” Philosophical Studies 110.2, 163–196.
Kim, Kihyeon. (1994). “The Deontological Conception of Epistemic Justification
and Doxastic Voluntarism,” Analysis 54.4, 282–2 84.
Kornblith, Hilary. (1982). “The Psychological Turn,” Australasian Journal of
Philosophy 60.3, 238–253.
—— — . (1983). “Justified Belief and Epistemically Responsible Action,”
———. (2001). “Epistemic Obligation and the Possibility of Internalism,” in
Abrol Fairweather and Linda Zagzebski (eds.), Virtue Epistemology: Essays
on Epistemic Virtue and Responsibility (Oxford: Oxford University Press),
231–2 48.
———. (2002). Knowledge and Its Place in Nature (Oxford: Clarendon Press).
Lackey, Jennifer. (2008). “What Luck Is Not,” Australasian Journal of Philosophy
86.2, 255–2 67.
Latus, Andrew. (2000). “Moral and Epistemic Luck,” Journal of Philosophical
Research 25, 149–172.
———. (2003). “Constitutive Luck,” Metaphilosophy 34.4, 460–475.
Le Morvan, Pierre. (2011). “On Ignorance: A Reply to Peels,” Philosophia 39.2,
335–3 44.
———. (2012). “On Ignorance: A Vindication of the Standard View,” Philosophia
40.2, 379–393.
———. (2013). “Why the Standard View of Ignorance Prevails,” Philosophia 41.1,
239–256.
Le Morvan, Pierre, and Rik Peels. (2016). “The Nature of Ignorance: Two Views,”
in Rik Peels and Martijn Blaauw (eds.), The Epistemic Dimensions of Ignorance
(Cambridge: Cambridge University Press), forthcoming.
Leon, Mark. (2002). “Responsible Believers,” The Monist 85.3, 421–435.
258
259
B ibliography
Levy, Neil. (2007). “Doxastic Responsibility,” Synthese 155.1, 127–155.

———. (2009). “What, and Where, Luck Is: A Response to Jennifer Lackey,”
Australasian Journal of Philosophy 87.3, 489–497.
Locke, John. (1975). An Essay Concerning Human Understanding, ed. Peter H.
Nidditch (Oxford: Clarendon Press).
Lycan, William G. (1988). Judgement and Justification (Cambridge: Cambridge
University Press).
McCain, Kevin. (2008). “The Virtues of Epistemic Conservatism,” Synthese
164.2, 185–2 00.
McHugh, Conor. (2014). “Exercising Doxastic Freedom,” Philosophy and
Mele, Alfred R. (1987). “Akratic Belief,” in Irrationality: An Essay on Akrasia, Self-
Deception, and Self-Control (Oxford: Oxford University Press), 109–120.
Meylan, Anne. (2013). Foundations of an Ethics of Belief (Berlin: Walter de
Gruyter).
———. (2015). “The Legitimacy of Intellectual Praise and Blame,” Journal of
Philosophical Research 40, 189–2 03.
Montmarquet, James A. (1993). Epistemic Virtue and Doxastic Responsibility
(Lanham: Rowman and Littlefield).
——— . (1995). “Culpable Ignorance and Excuses,” Philosophical Studies
80.1, 41–49.
———. (1999). “Zimmerman on Culpable Ignorance,” Ethics 109.4, 842–8 45.
———. (2008). “The Voluntariness of Virtue—a nd Belief,” Philosophy 83.325,
373–390.
Moore, Adrian W. (1990). “A Kantian View of Moral Luck,” Philosophy 65.253,
297–321.
Morillo, Carolyn R. (1984). “Epistemic Luck, Naturalistic Epistemology and
the Ecology of Knowledge or What the Frog Should Have Told Dretske,”
Philosophical Studies 46.1, 109–129.
Moser, Paul K. (1989). Knowledge and Evidence (Cambridge: Cambridge
University Press).
Murray, J. Clark. (1914). “Ignorance,” in James Hastings (ed.), Encyclopaedia of
Religion and Ethics, Vol. 7 (Edinburgh: T. & T. Clark), 103–105.
Nagel, Thomas. (1993). “Moral Luck,” in Daniel Statman (ed.), Moral Luck
(Albany: State University of New York Press), 57–71.
Naylor, Margery B. (1985). “Voluntary Belief,” Philosophy and Phenomenological
Research 45.3, 427–436.
Nickel, Philip J. (2010). “Voluntary Belief on a Reasonable Basis,” Philosophy and
Nottelmann, Nikolaj. (2007). Blameworthy Belief: A Study in Epistemic
Deontologism (Dordrecht: Springer).
259
260
B ibliography
———. (2013). “The Deontological Conception of Epistemic Justification: A

Reassessment,” Synthese 190.12, 2219–2241.
Nottelmann, Nikolaj, and Rik Peels. (2013). “Some Metaphysical Implications
of a Credible Ethics of Belief,” in Nikolaj Nottelmann (ed.), New Essays on
Belief: Constitution, Content, and Structure (New York: Palgrave Macmillan),
230–250.
Nussbaum, Martha C. (1993). “Luck and Ethics,” in Daniel Statman (ed.), Moral
O’Saughnessy, Brian. (1980). The Will: A Dual Aspect Theory
Oshana, Marina A.L. (1997). “Ascriptions of Responsibility,” American
Owens, David. (2000). Reason without Freedom: The Problem of Epistemic
Normativity (London: Routledge and Kegan Paul).
———. (2002). “Epistemic Akrasia,” The Monist 85.3, 381–397.
Peels, Rik. (2010a). “Epistemic Desiderata and Epistemic Pluralism,” Journal of
Philosophical Research 35, 193–2 07.
———. (2010b). “The Ethics of Belief and Christian Faith as Commitment to
Assumptions,” Religious Studies 46.1, 97–107.
———. (2010c). “What Is Ignorance?,” Philosophia 38.1, 57–67.
———. (2011a). “Ignorance Is Lack of True Belief: A Rejoinder to Le Morvan,”
Philosophia 39.2, 345–355.
———. (2011b.) “Tracing Culpable Ignorance,” Logos and Episteme 2.4, 575–582.
———. (2012). “The New View on Ignorance Undefeated,” Philosophia 40.4,
741–750.
—— —. (2013a). “Belief- Policies Cannot Ground Doxastic Responsibility,”
Erkenntnis 78.3, 561–569.
———. (2013b). “Does Doxastic Responsibility Entail the Ability to Believe
Otherwise?,” Synthese 190.17, 3651–3669.
———. (2014a). “Against Doxastic Compatibilism,” Philosophy and Phenomenological
Research 89.3, 679–702.
———. (2014b). “What Kind of Ignorance Excuses? Two Neglected Issues,”
———. (2015a). “A Modal Solution to the Problem of Moral Luck,” American
———. (2015b). “Believing at Will Is Possible,” Australasian Journal of Philosophy
93.3, 524–5 41.
———. (2015c). “Let’s Bite the Bullet on Deontological Epistemic Justification:
A Response to Robert Lockie,” Social Epistemology Review and Reply
Collective 4.12, 42–49.
—— —. (2016a). “Responsible Belief and Epistemic Justification,” Synthese,
forthcoming.
260
261
B ibliography
———, ed. (2016b). Perspectives on Ignorance from Moral and Social Philosophy
(London: Routledge).
Peels, Rik, and Anthony Booth. (2014). “Why Responsible Belief Is Permissible
Belief,” Analytic Philosophy 55.1, 75–88.
Peels, Rik, and Martijn Blaauw, eds. (2016). The Epistemic Dimensions of Ignorance
Plantinga, Alvin. (1983). “Reason and Belief in God,” in Alvin Plantinga and
Nicholas Wolterstorff (eds.), Faith and Rationality: Reason and Belief in God
(Notre Dame: University of Notre Dame Press), 16–93.
—— — . (1990). “Justification in the Twentieth Century,” Philosophy and
Phenomenological Research 50, Supplement, 45–71.
———. (1993). Warrant: The Current Debate (Oxford: Oxford University Press).
Pollock, John L., and Joseph Cruz. (1986). Contemporary Theories of Knowledge,
2nd ed. (Lanham: Rowman and Littlefield).
Price, Henry H. (1954). “Belief and the Will,” Proceedings of the Aristotelian
Society, Supplementary Volume 28, 1–2 6.
Prior, Elizabeth W. (1985). Dispositions (Aberdeen: Aberdeen University Press).
Pritchard, Duncan. (2005). Epistemic Luck (Oxford: Clarendon Press).
———. (2006). “Moral and Epistemic Luck,” Metaphilosophy 37.1, 1–25.
Pryor, James. (2001). “Highlights of Recent Epistemology,” The British Journal for
the Philosophy of Science 52.1, 95–124.
Railton, Peter. (1997). “On the Hypothetical and Non- Hypothetical in
Reasoning about Belief and Action,” in Garrett Cullity and Berys Nigel
Gaut (eds.), Ethics and Practical Reason (Oxford: Oxford University
Press), 53–79.
Ramsey, Frank P. (1931). The Foundations of Mathematics, ed. Richard B.
Braithwaite and George E. Moore (London: Routledge and Kegan Paul).
Rescher, Nicholas. (1993). “Moral Luck,” in Daniel Statman (ed.), Moral Luck
(Albany: State University of New York Press), 141–166.
———. (1995). Luck: The Brilliant Randomness of Everyday Life (Pittsburgh:
University of Pittsburgh Press).
Richards, Norvin. (1993). “Luck and Desert,” in Daniel Statman (ed.), Moral
Riggs, Wayne D. (2003). “Balancing Our Epistemic Goals,” Noûs 37.2, 342–352.
Rivera-L ópez, Eduardo. (2006). “Can There Be Full Excuses for Morally Wrong
Actions?,” Philosophy and Phenomenological Research 73.1, 124–142.
Rosen, Gideon. (2003). “Culpability and Ignorance,” Proceedings of the Aristotelian
Society 103.1, 61–8 4.
———. (2004). “Skepticism about Moral Responsibility,” in John Hawthorne
(ed.), Philosophical Perspectives 18: Ethics (Oxford: Blackwell), 295–313.
———. (2008). “Kleinbart the Oblivious and Other Tales of Ignorance and
Responsibility,” The Journal of Philosophy 105.10, 591–610.
261
262
B ibliography
Ross, W. David. (1939). Foundations of Ethics (Oxford: Clarendon Press).

———. (2002). The Right and the Good, ed. Philip Stratton-L ake (Oxford:
Clarendon Press).
Russell, Bruce. (2001). “Epistemic and Moral Duty,” in Matthias Steup (ed.),
Knowledge, Truth, and Duty: Essays on Epistemic Justification, Responsibility,
and Virtue (Oxford: Oxford University Press), 34–4 8.
Ryan, Sharon. (2003). “Doxastic Compatibilism and the Ethics of Belief,”
Philosophical Studies 114.1–2 , 47–79.
Ryle, Gilbert. (1949). The Concept of Mind (London: Hutchinson’s University
Library).
Scanlon, Thomas M. (1998). What We Owe to Each Other (Cambridge, Mass.:
Harvard University Press).
Schellenberg, John L. (2005). Prolegomena to a Philosophy of Religion (Ithaca:
Cornell University Press).
Schleifer-McCormick, Miriam. (2015). Believing Against the Evidence: Agency and
the Ethics of Belief (New York: Routledge).
Schlick, Moritz. (1962). Problems of Ethics (New York: Dover Publications).
Schmitt, Frederik F. (2014). “Group Belief and Acceptance,” in Sara Rachel Chant,
Frank Hindriks, and Gerhard Preyer (eds.), From Individual to Collective
Intentionality: New Essays (Oxford: Oxford University Press), 61–88.
Schwitzgebel, Eric. (2002). “A Phenomenal, Dispositional Account of Belief,”
Noûs 36.2, 249–275.
Scott-K akures, Dion. (1994). “On Belief and the Captivity of the Will,” Philosophy
and Phenomenological Research 54.1, 77–103.
Shah, Nishi. (2002). “Clearing Space for Doxastic Voluntarism,” The Monist 85.3,
436–4 45.
Sinnott-A rmstrong, Walter. (1984). “‘Ought’ Conversationally Implies ‘Can’,”
Slote, Michael. (1992). From Morality to Virtue (Oxford: Oxford University Press).
Smith, Angela M. (2000). “Identification and Responsibility,” in Ton van den
Beld (ed.), Moral Responsibility and Ontology (Dordrecht: Kluwer Academic
Publishers), 233–2 46.
———. (2005). “Responsibility for Attitudes: Activity and Passivity in Mental
Life,” Ethics 115.2, 236–271.
Smith, Holly M. (1983). “Culpable Ignorance,” Philosophical Review 92.4,
543–571.
———. (2011a). “Non-Tracing Cases of Culpable Ignorance,” Criminal Law and
———. (2011b). “The Moral Clout of Reasonable Beliefs,” in Mark Timmons
(ed.), Oxford Studies in Normative Ethics, Vol. 1 (Oxford: Oxford University
Press), 1–25.
262
263
B ibliography
———. (2013). “Negligence,” in Hugh LaFollette (ed.), International Encyclopedia

of Ethics (Oxford: Blackwell), 3565–3571.
———. (2016). “Tracing Cases of Culpable Ignorance,” in Rik Peels (ed.),
Perspectives on Ignorance from Moral and Social Philosophy (London:
Routledge), 95–119.
Sosa, Ernest. (2014). “Knowledge and Time: Kripke’s Dogmatism Paradox
and the Ethics of Belief,” in Jonathan Matheson and Rico Vitz (eds.),
The Ethics of Belief: Individual and Social (Oxford: Oxford University
Press), 77– 8 8.
Statman, Daniel. (1991). “Moral and Epistemic Luck,” Ratio 4.2, 146–156.
———. (1993). “Introduction,” in Daniel Statman (ed.), Moral Luck (Albany: State
University of New York Press), 1–34.
Steup, Matthias. (1988). “The Deontic Conception of Epistemic Justification,”
Philosophical Studies 53.1, 65–8 4.
——— . (2000). “Doxastic Voluntarism and Epistemic Deontology,” Acta
Analytica 15.1, 25–56.
———. (2001). “Introduction,” in Matthias Steup (ed.), Knowledge, Truth, and
Duty: Essays on Epistemic Justification, Responsibility, and Virtue (Oxford:
Oxford University Press), 3–18.
———. (2008). “Doxastic Freedom,” Synthese 161.3, 375–392.
———. (2012). “Belief Control and Intentionality,” Synthese 188.2, 145–163.
———. (2016). “Believing Intentionally,” Synthese, forthcoming.
Stocker, Michael. (1971). “‘Ought’ and ‘Can’,” Australasian Journal of Philosophy
49.3, 303–316.
———. (1982). “Responsibility Especially for Beliefs,” Mind 91.363, 398–417.
Strawson, Peter F. (1959). Individuals: An Essay in Descriptive Metaphysics
(London: Methuen).
———. (1974). “Freedom and Resentment,” in Freedom and Resentment and Other
Essays (London: Methuen), 1–25.
———. (1992). Analysis and Metaphysics: An Introduction to Philosophy (Oxford:
Oxford University Press).
Stump, Eleonore. (1993). “Intellect, Will, and the Principle of Alternate
Possibilities,” in John M. Fischer and Mark Ravizza (eds.), Perspectives on
Moral Responsibility (Ithaca: Cornell University Press), 237–2 62.
Swinburne, Richard. (2001). Epistemic Justification (Oxford: Clarendon Press).
Thomason, Richmond H. (1981). “Deontic Logic and the Role of Freedom in
Moral Deliberation,” in Risto Hilpinen (ed.), New Studies in Deontic Logic
(Dordrecht: Reidel), 177–186.
Thomson, Judith J. (1993). “Morality and Bad Luck,” in Daniel Statman (ed.),
Moral Luck (Albany: State University of New York Press), 195–215.
Tidman, Paul. (1996). “Critical Reflection: An Alleged Epistemic Duty,” Analysis
56.4, 268–276.
263
264
B ibliography
Unger, Peter. (1968). “An Analysis of Factual Knowledge,” The Journal of

———. (1975). Ignorance: A Case for Scepticism (Oxford: Clarendon Press).
Vahid, Hamid. (1998). “Deontic vs. Nondeontic Conceptions of Epistemic
Justification,” Erkenntnis 49.3, 285–301.
Van Inwagen, Peter. (1983). An Essay on Free Will (Oxford: Clarendon Press).
Van Woudenberg, René. (2009). “Ignorance and Force: Two Excusing Conditions
for False Beliefs,” American Philosophical Quarterly 46.4, 373–386.
Walker, Margaret Urban. (1993). “Moral Luck and the Virtues of Impure Agency,”
in Daniel Statman (ed.), Moral Luck (Albany: State University of New York
Press), 235–250.
Wallace, R. Jay. (1994). Responsibility and the Moral Sentiments (Cambridge,
Mass.: Harvard University Press).
Watson, Gary. (1993). “Responsibility and the Limits of Evil: Variations on a
Strawsonian Theme,” in John M. Fischer and Mark Ravizza (eds.), Perspectives
on Moral Responsibility (Ithaca: Cornell University Press), 119–148.
Weatherson, Brian. (2008). “Deontology and Descartes’ Demon,” The Journal of
Philosophy 105.9, 540–569.
Wedgwood, Ralph. (2007). The Nature of Normativity (Oxford: Clarendon Press).
Wieland, Jan Willem, and Philip Robichaud, eds. (2016). Responsibility: The
Epistemic Condition (Oxford: Oxford University Press).
Williams, Bernard A.O. (1973). “Deciding to Believe,” in Problems of the
Self: Philosophical Papers 1956–1972 (Cambridge: Cambridge University
Press), 136–151.
———. (1993). “Moral Luck,” in Daniel Statman (ed.), Moral Luck (Albany: State
University of New York Press), 35–55.
Williamson, Timothy. (2007a). “On Being Justified in One’s Head,” in
Mark Timmons, John Greco, and Alfred R. Mele (eds.), Rationality and
the Good: Critical Essays on the Ethics and Epistemology of Robert Audi
(Oxford: Oxford University Press), 106–122.
———. (2007b). The Philosophy of Philosophy (Oxford: Blackwell).
Wolf, Susan. (1980). “Asymmetrical Freedom,” The Journal of Philosophy 77.3,
151–166.
———. (1990). Freedom within Reason (Oxford: Oxford University Press).
———. (1993). “The Importance of Free Will,” in John M. Fischer and Mark
Ravizza (eds.), Perspectives on Moral Responsibility (Ithaca: Cornell
Wolterstorff, Nicholas. (1996). John Locke and the Ethics of Belief
—— —. (2005). “Obligation, Entitlement, and Rationality,” in Matthias
Steup and Ernest Sosa (eds.), Contemporary Debates in Epistemology
(Oxford: Blackwell), 326–338.
264
265
B ibliography
———. (2008). Justice: Rights and Wrongs (Princeton: Princeton University

Press).
———. (2010a). “Entitlement to Believe and Practices of Inquiry,” in Practices
of Belief: Selected Essays, Vol. 2, ed. Terence Cuneo (Cambridge: Cambridge
———. (2010b). “Ought to Believe—Two Concepts,” in Practices of Belief: Selected
Essays, Vol. 2, ed. Terence Cuneo (Cambridge: Cambridge University
Press), 62–85.
Zagzebski, Linda. (1994). “Religious Luck,” Faith and Philosophy 11.3, 397–413.
———. (2001). “Must Knowers Be Agents?,” in Abrol Fairweather and Linda
Zagzebski (eds.), Virtue Epistemology: Essays on Epistemic Virtue and
Responsibility (Oxford: Oxford University Press), 142–157.
Zimmerman, Michael J. (1986). “Negligence and Moral Responsibility,” Noûs
20.2, 199–218.
———. (1988). An Essay on Moral Responsibility (Totowa, N.J.: Rowman and
Littlefield).
———. (1993). “Luck and Moral Responsibility,” in Daniel Statman (ed.), Moral
———. (1996). The Concept of Moral Obligation (Cambridge: Cambridge
University Press).
———. (1997). “Moral Responsibility and Ignorance,” Ethics 107.3, 410–426.
———. (2002). “Taking Luck Seriously,” The Journal of Philosophy 99.11, 553–576.
———. (2006). “Moral Luck: A Partial Map,” Canadian Journal of Philosophy
36.4, 585– 6 08.
———. (2008). Living with Uncertainty: The Moral Significance of Ignorance
———. (2016). “Ignorance as a Moral Excuse,” in Rik Peels (ed.), Perspectives
on Ignorance from Moral and Social Philosophy (London: Routledge), 77–94.
265
266
â•‡ 267
INDEX
ability asymmetry thesis, 134, 155, 157–â•‰162

to believe at will, 61–â•‰66 doxastic, 157–â•‰162
to believe otherwise, 81, 132–â•‰163, 249 attitude
to meet an intellectual obligation, affective, 18, 23
171–â•‰173 doxastic, 30, 42, 67, 91, 99, 156, 193
accidentality, 203, 215–â•‰218 evaluative, 18, 20, 26, 47, 83, 157, 184
account normative, 18–â•‰23, 28, 47, 83, 157–â•‰159,
constitutive, 27 184, 197, 222
in terms of necessary and sufficient objective, 18, 222
conditions, 3–â•‰6, 15, 25, 27 reactive, 18
accountability, 25 Audi, Robert, 36, 41, 89
action Austin, J.L., 5
belief-â•‰influencing, 98, 101, 111, 156, 158
habitual, 74, 77, 195 badness
Adler, Jonathan, 8, 221 epistemic, 102–â•‰108, 122–â•‰123, 130,
advice, doxastic, 4 148, 170
akrasia, 31, 189–â•‰194 objective, 96–â•‰98, 105, 122, 125,
Alston, William, 34, 52–â•‰55, 67–â•‰68, 86, 92, 197, 217
117, 238–â•‰2 39, 242–â•‰2 44 subjective, 97–â•‰98, 116, 123–â•‰124, 218
analysis basic law V, 36–â•‰37, 39, 41
connective, 3 beauty, 47, 82, 159
in terms of necessary and sufficient behaviorism, 24, 26, 35
conditions, 3–â•‰6, 15, 25, 27 belief
appraisal acquisition of, 68
legal doxastic, 7 actual, 149
negative, 17, 157 blameless, 43–â•‰51, 129, 142, 157,
neutral, 17, 22 173, 249
positive, 17, 157 blameworthy, 7–â•‰8, 15, 46, 51, 86, 116,
Aristotle, 133, 164, 192 134, 157–â•‰163, 185–â•‰197, 200, 220
267
268
I ndex
belief (Cont.) doxastic, 71–81, 150–152

Combination Account of, 37–43 about omissions, 134
degrees of, 4, 168 strict doxastic, 150–152
dormant, 11, 14–15, 18, 31–43, 141, 159, compulsion, 130, 135, 162, 194
163, 177, 181, 193–198, 246 condition
goal of, 246 necessary and sufficient, 3–6 , 15, 25, 27
irrational, 7, 103, 108, 122, 141 Conee, Earl, 8, 241
maintenance, 68–71 Coningham, Arthur, 1, 100, 188
occurrent, 11, 14–15, 28–43, 51, 114, consequence, 96, 125, 134, 138, 147, 156,
177, 189, 193–197, 246 177, 199–2 01, 234
praiseworthy, 11, 15, 43–51, 48, 129, consequence-particular, 152–155
140, 157–159 consequence-u niversal, 152–155
property of, 148–150 conservatism, epistemic, 44–45, 142
responsible, 43, 118–119, 121, 143, control
184, 217 causal, 57–58, 203
subconscious, 114 compatibilist, 8, 11, 61, 63, 88
suspension of, 113, 166, 169, 171–172 direct doxastic, 11, 61–6 6
tacit, 31–43, 51, 140–141, 159, 166, 177, indirect doxastic, 66–72
193–198, 246 intentional, 58–59, 61, 66–67, 77–78,
belief-policy, 94–95 86, 118, 202–2 03, 205, 207
Bennett, Jonathan, 199 over intentions, 77–78
Bieber, Justin, 102 the problem of, 232–2 34
Blamelessness Thesis, 43–51 voluntary, 52–55, 61, 83
blameworthiness control view, 26–27
degree of, 144, 165, 168, 224 counterfactual, 3, 35–36, 152–156, 180,
scope of, 224 219, 226, 229
simpliciter, 227–228 Coyne, Margaret, 221
Bogdan, Radu, 31–32 Crick, Francis, 118
BonJour, Laurence, 103 cricket captain, 46, 158
Booth, Anthony, xii, 13, 49, 70, 79, 107, 110
boundary case, 6, 140, 205 defeater, 45, 50–51, 248
Breivik, Anders, 2 degrees, 4, 128, 133, 137, 141, 144,
Browne, Brynmor, 222 168, 224
Delay Argument, 65–6 6
Cap Arcona incident, 1–2 demand
Card, Claudia, 203 doxastic, 85–87
Chrisman, Matthew, 84–85 deontologism, 4, 53, 82, 96, 109,
Chuard, Philippe, 85–87 237–2 50
Cohen, Jonathan, 34 modest, 82
commensurability, 126 desert, 19
communism, 43 determinism, 21, 57, 64, 202
community disbelief, 30, 39, 42, 91, 181
epistemic, 84–85 discrepancy
compatibilism doxastic, 106–109, 113, 123
about actions, 58, 81, 134, 150–152, disposition
155–157 belief-forming, 31, 94, 151
about consequences, 134, 156 to believe, 31, 36, 41, 140
268
269
I ndex
DNA, 118 Gettier, Edmund, 5, 69, 167, 209

Dougherty, Trent, 114–115 Gettier-case, 69, 209
Ginet, Carl, 61
epistemology, 3, 92, 199, 240, 246 Glover, Jonathan, 26
social, 3 goal
essence, 27, 228 of belief, 246
evidence, 9, 35, 48, 50, 62, 117, 145, 161, epistemic, 105, 238
196, 227, 241 Jamesian, 102, 116, 240, 242–245, 248, 250
evidence gathering, 117, 145, 161, 196, 227 God, 55, 70, 113–114, 141, 161
evidentialism, 241 goodness, 96, 244
evil demon scenario, 46–47
excuse Haarsma, Ieke, xi
broad, 10, 128–129, 139, 142, 183–184 Hall, Richard, 103
doxastic, 8–12, 31, 124–130, 162–163, Heil, John, 66, 89
198, 234–2 35 Heller, Mark, 73–74
full, 144–147, 171–178, 182, 186–188 Helm, Paul, 94–95
narrow, 10, 128, 131, 139, 184 Hetherington, Stephen, 142
partial, 127–128, 144–145, 171, 177, Hieronymi, Pamela, 78, 110
182, 186–188
externalism, 49–51, 238–2 42, 246 ideal
extremism, 2 epistemic, 82–8 4, 87
ignorance
factor blameless, 12, 125, 165, 188, 198
belief-i nfluencing, 11–12, 90–96, 100, blameworthy, 31, 175, 178, 185–197
120, 212–213, 216, 218 deep, 118, 169, 171–182, 198, 211
Feinberg, Joel, 26 disbelieving, 174, 178–182
Feldman, Richard, 68–72, 81–82, 241 as lack of knowledge, 165–167, 198
Feldman-case, 69–72 as lack of true belief, 166
Fischer, John Martin, 81, 150, 153 suspending, 171, 173, 175, 177–182, 198
FitzPatrick, William, 12, 189–192 implication
force deflationary, 8
diachronic, 142 inflationary, 8
full, 136–137 improbability, 199, 205
partial, 135–137 incommensurability, 126
synchronic, 138 indoctrination, 4, 74
foresight influence
lack of, 12, 67, 147, 156, 165, 175–181, doxastic, 11, 15, 27, 46, 67, 75, 90, 114,
187–188, 198 118, 134, 148, 202, 208, 238
Frankfurt, Harry, 134, 150–162, 219 intention, 58, 73, 76–8 0, 189, 212
Frankfurt scenario, 150–151 intentionality
Frankfurt-style scenario, 134, weak, 73–74
150–156, 219 internalism, 49, 238–2 42, 246
Franklin, Rosalind, 118 involuntarism
Frege, Gottlob, 36–37, 39, 41 doxastic, 9, 44, 46, 53, 55, 59–61, 135,
Friedman, Jane, 42 163, 237–2 40
functionalism, 42 conceptual doxastic, 55
fundamentalism, 2, 13 psychological doxastic, 55
269
270
I ndex
irrationality, 122–123, 141, 150, 168, 178, mechanism

197, 234 doxastic, 91–96, 102–108, 161, 209,
219–220, 228
James, William, 102 mens rea, 7
Johnson, Charles, 103 merit, 19
judgment Meylan, Anne, 89
evaluative doxastic, 28 modality, 191, 229, 231, 235
normative doxastic, 28 Montmarquet, James, 8, 189–192
suspension of, 30, 42, 77, 91, 110–111, Morillo, Carolyn, 203
171, 174–175 motivation, 25, 121, 182
justification
deontological, 4, 53, 163, 237–250 nature
epistemic, 4, 13, 53, 110, 117, 141, 163, epistemic, 73–74
237–250 New View on ignorance, 166–168, 198
externalist, 238, 242 Nickel, Philip, 61
internalist, 238 non-accidentality, 200, 215–218
Nottelmann, Nikolaj, xii, 8, 38, 66, 123,
Katzoff, Charlotte, 111–113 148–149, 151, 163, 238
Kavka, Gregory, 77–78
Kim, Kihyeon, 103 obligation
knowledge, 3, 5, 28, 33, 41, 81, 104, 165–169, all-t hings-considered, 56–57, 110, 116,
179, 199, 246–247 119, 121, 126, 137–138, 169–175
knowledge-how, 58 contingent, 100–101
Kornblith, Hilary, 82 contractual, 82
doxastic, 11, 52–88, 100, 150
Lackey, Jennifer, 205–2 06 epistemic, 101–116, 121, 170
Latus, Andrew, 206, 224 intellectual, 9, 12–13, 90, 96–100
ledger view, 25–2 6 moral, 110, 114–116, 126, 170,
Leon, Mark, 142 210, 248
Locke, John, 243–2 44 non-contingent, 101
luck objective, 105, 127
alethic, 209 pro tanto, 56, 102, 110, 114, 116, 119,
aretaic, 226–227, 230–2 31, 235 126–127, 137, 171
bad, 200, 206, 210–211, 221–222, prudential, 71, 109–110, 115
230, 234 role, 81–82, 87
consequential, 13, 213–218, 227, subjective, 97, 104, 125, 197, 245
231–2 35 omission, 67, 107, 109, 132, 134–135, 150,
constitutive, 211 155, 157
evidential, 13, 210–212, 218, 231, 235 Oshana, Marina, 25
good, 221, 230 ought implies can, 54, 109
mechanism, 209–210, 212–213, 218, ought-to-be, 84–85
226, 230–2 31 ought-to-do, 84–85
property, 208–2 09 Owens, David, 8, 111
Lycan, William, 39 Owens, Jonathan, 78
McCain, Kevin, 45 paradigm case, 6, 135, 140

McHugh, Conor, 77, 79 Paris terrorist attacks, 2
270
271
I ndex
Peels, Rik, 4, 13, 29, 38, 49, 55, 72, 95, 133, doxastic, 14– 43
164, 166–168, 170, 188, 218, 231, epistemic, 17, 114, 237, 249–250
237, 239 group, 3
person, 227 individual, 3, 202
personality, 227 Ledger Account of, 25–2 6
Philipse, Herman, xiv legal, 16–17, 234, 245
Plantinga, Alvin, 247 ministerial, 16–17
possible world, 3, 74, 200, 204–2 06, moral, 16
210–219, 228–2 35 normative, 16–17
closeness of, 200, 204–2 06, 210–219, original, 117–118, 207, 216–217
228–2 35 prudential, 16–17, 26, 70–72, 110
praiseworthiness, 12, 43, 45–4 8, 139, role, 16–17
157–162, 224 reward, 24, 77, 80
Praiseworthiness Thesis, 43–51 rightness, 157
Price, Henry, 66 rule
Principle of Credulity, 45 of action, 84
Pritchard, Duncan, 201, 229 of criticism, 84
privacy, 99, 101, 119, 248 Russell, Bertrand, 36
problem Russell, Bruce, 76
of control, 232–2 34 Ryan, Sharon, 59–61, 73–76
of doxastic luck, 218–2 33
the modal solution to the problem scenario
of, 229 Frankfurt, 150–151
of moral luck, 221 Frankfurt-style, 134, 150–156, 219
punishment, 16, 24 New Evil Demon, 47
Schleifer-McCormick, Miriam, 89
racism, 30, 94, 144, 151–155, 197, 214–2 35 Schlick, Moritz, 24
rationality, 19, 81, 148 Sellars, Wilfrid, 84
Ravizza, Mark, 81, 150, 153 Shah, Nishi, 78
reason, 8, 25, 56, 62–6 6, 72, 75–82, 110– Sinnott-A rmstrong, 61
112, 121, 159, 179, 195–196, 238 situatedness
epistemic, 110–111, 238 cognitive, 42, 90–96, 98, 130, 138
reason-responsiveness, 9, 26, 73–74 skepticism, 45, 113
reductio, 102, 225, 234 Sosa, Ernest, 247
reduction, 3 Southwood, Nicholas, 85–86
regress, 12, 31, 78, 165, 185–197 speech act, 17, 124
reliability, 92–93, 102, 106–107 Stalin, Joseph, 189
Rescher, Nicholas, 226 Standard View on ignorance,
resemblance 165–169, 198
family, 5 state of affairs, 16, 18, 68–69, 96, 101,
responsibility 125–128, 148–150, 187,
Accountability Account of, 25 200–2 03, 244
Appraisal Account of, 14–2 8, 43 Steup, Matthias, xii, 61–65, 73–81,
causal, 16 111, 141
for consequences, 117–118, Strawson, Peter, 3, 15, 18, 21–22
207–2 08, 216 subjectivity, 19, 97–9 9, 104, 107, 116,
derivative, 117–118, 207–2 08, 216 123–125, 196–198, 207, 217–218
271
272
I ndex
suspension intellectual, 9, 92, 94, 106, 145, 191,

of belief, 113, 166, 169, 171–172 211, 247
of judgment, 30, 42, 77, 91, 110–111, virtue
171, 174–175 intellectual, 9, 75, 90–93, 102, 107, 138,
Swinburne, Richard, 45 161, 196, 211, 244, 246, 249
moral, 227
Talisker 57° North, 127
thought, 11, 29–43 warrant, 19, 69, 104, 123, 130, 166–169
time-i ndex, 41 Watson, James D., 118
Toxin Puzzle, 77–78 Weatherson, Brian, 44, 46–51, 158
Wedgwood, Ralph, xii
Up-to-me-ness Argument, 63–6 6 Wiesenthal, Simon, 19
Williams, Bernard, 55
Vahid, Hamid, 239–2 40 Wolf, Susan, 158–160
value, 129, 160, 203, 206 Wolterstorff, Nicholas, xii, 69
epistemic, 26, 245 worth
moral, 26, 70 essential, 228
prudential, 70 intellectual, 228
truth, 110, 144 wrongness, 157, 189
Van Inwagen, Peter, 152–155
Van Woudenberg, René, xii Zimmerman, Michael, 12, 165, 189, 191–196,
vice 203, 224, 227
272
273
274

Peels, Rik - Responsible Belief - A Theory in Ethics and Epistemology (2017, Oxford University Press)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Peels, Rik - Responsible Belief - A Theory in Ethics and Epistemology (2017, Oxford University Press)

Uploaded by

Copyright:

Available Formats

i

Published in the United States of America by Oxford University Press

© Oxford University Press 2017

All rights reserved. No part of this publication may be reproduced, stored in

You must not circulate this work in any other form

CIP data is on file at the Library of Congress

Introduction: Why Responsible Belief Matters 1

2. The Problem: Doxastic Control

4.5 C ontrol over or Influence on Properties

Appendix: Responsible Belief and Epistemically Justified Belief 237

provided a stimulating research environment in which working on this

belief-​policies in c­ hapter 3 summarize what I argue in “Belief-​Policies

I.1 R ESPONSIBLE BELIEF

responsibly or culpably. It seems that if he was not culpable for hold-

I ntroduction : W hy R esponsible B elief M atters

The purpose of this book is to develop an account of responsible belief.

doxastic responsibility of groups of people would not remove the doxas-

5. For more on my view on epistemic justification, independently of how it relates to

I ntroduction : W hy R esponsible B elief M atters

Let me stress that the aim of this book is to provide an account of

7. See Gettier (1963).

One may object that an analysis in terms of necessary and sufficient

I argued above that it matters whether or not we believe responsibly. But

10. S ee Alston (1989d).

I ntroduction : W hy R esponsible B elief M atters

something is the case or that something is true. I cannot now choose to

I ntroduction : W hy R esponsible B elief M atters

compatibilist doxastic control are mistaken. We are indeed responsible

First, as I pointed out above, without an analysis of doxastic excuses,

I ntroduction : W hy R esponsible B elief M atters

This book is structured as follows. In ­chapter 1, I provide analyses of

obligations.” Also, I argue that, roughly, one responsibly believes that p

I ntroduction : W hy R esponsible B elief M atters

to be lucky for some person is for it to be beyond her intentional control,

Some of the difficulties that plague an account of responsible belief

D oxastic R esponsibility: W hat I s I t ?

What is it to be responsible? Before answering this question, let me be

position to answer that question if we have first acquired a firmer grip

1. Thus also Hart (1970, 211–​2 30).

D oxastic R esponsibility: W hat I s I t ?

responsibility and, especially important for our purposes, epistemic

(a) “You are responsible(i) for what you say.”

2. See Strawson (1974, 4–​13).

D oxastic R esponsibility: W hat I s I t ?

by which we hold people up to certain norms. I prefer the term “evalua-

3. Th is is also pointed out by Feinberg (1970, 127), Watson (1993, 145).

culpable. Also, that S is the proper object of normative attitudes does

D oxastic R esponsibility: W hat I s I t ?

attitude of neutral appraisal toward someone for φ-​ing, then I could in

7. See Oshana (1997, 80).

D oxastic R esponsibility: W hat I s I t ?

not follow that an analysis of being admirable in terms of one’s relation

out what it is to be responsible in terms of such attitudes, then such an

8. See Schlick (1962, 151–​158).

D oxastic R esponsibility: W hat I s I t ?

Second, some philosophers think that to be responsible is to be

10. S ee Oshana (1997, 77–​8 0).

marks add to one’s debit. To be responsible is to have positive or negative

12. S ee Feinberg (1970, 30–​31) and Glover (1970, 64–​65).

D oxastic R esponsibility: W hat I s I t ?

or essence of responsibility.13 Let me explain what I mean by this by way

In most circumstances, these will not be normative doxastic judgments,

D oxastic R esponsibility: W hat I s I t ?

2. The Problem: Doxastic Control

4.5 C ontrol over or Influence on Properties

belief-policies in c hapter 3 summarize what I argue in “Belief-Policies

I.1 R ESPONSIBLE BELIEF

5. For more on my view on epistemic justification, independently of how it relates to

7. See Gettier (1963).

10. S ee Alston (1989d).

This book is structured as follows. In chapter 1, I provide analyses of

1. Thus also Hart (1970, 211–2 30).

2. See Strawson (1974, 4–13).

3. Th is is also pointed out by Feinberg (1970, 127), Watson (1993, 145).

attitude of neutral appraisal toward someone for φ-ing, then I could in

7. See Oshana (1997, 80).

8. See Schlick (1962, 151–158).

10. S ee Oshana (1997, 77–8 0).

12. S ee Feinberg (1970, 30–31) and Glover (1970, 64–65).

OB: S occurrently believes that p at some time t iff S at t consid-

DB: S dormantly believes that p at t iff (i) S at t does not consider

17. S ee Bogdan (1986).

DB′: S dormantly believes that p at t iff (i) S does not consider p

20. See his example of Tom offending Norma: Audi (1994, 419–420).

DB″ S dormantly believes that p at t iff (i) S at t does not consider

TB″ S tacitly believes that p at t iff (i) S at t does not occurrently

It seems to me that the Combination Account—which I take to include

24. See Lycan (1988, 60).

25. Th is objection might be inspired by Goldman (1978, 526).

27. S ee Audi (1994, 422).

29. S ee Friedman (2013).

BT: S responsibly believes that p iff S blamelessly believes that p.

32. F or a defense of this thesis, see McCain (2008).

35. For his presentation of such scenarios, see Pryor (2001, 117).

In c hapter 1, I argued that to believe responsibly is to be the proper

2.3 O BLIGATIONS AND CONTROL

S has causal control over φ-i ng iff S can φ and S can ~φ.6

7. Here, I assume that my inability to φ or my ignorance about how to φ is not itself

S at t has an obligation to φ only if φ-i ng is normally under S’s

9. See Ryan (2003, 49–57).

2.4 F IRST RESPONSE: WE HAVE DIRECT

12. S ee Sinnott-A rmstrong (1984); Stocker (1971).