Emotion and Value (2015)

Emotion and Value
Emotion and Value
EDITED BY
Sabine Roeser
Cain Todd
1
1
Great Clarendon Street, Oxford, ox2 6dp,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© the several contributors 2014
The moral rights of the authors have been asserted
First Edition published in 2014
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2014940450
ISBN 978–0–19–968609–4
Printed and bound by
CPI Group (UK) Ltd, Croydon, cr0 4yy
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
In memory of Peter Goldie
Acknowledgements
The idea for this volume arose from the happy coincidence of three conferences that
took place independently in three different countries in May 2011, all focusing in some
way or another on emotion and value. Cain Todd and Julien Deonna organized a con-
ference on ‘Emotion, Self, and Time’ at the University of Geneva, in collaboration with
the Swiss Centre for Affective Sciences and the Thumos Research Group, which was
generously funded by the Swiss National Science Foundation. Sabine Roeser organ-
ized the conference ‘Moral Emotions and Intuitions’ in The Hague, generously funded
via the Philosophy Department at TU Delft and via her research project on ‘Moral
Emotions and Risk Politics’, funded by NWO (Netherlands Organization for Scientific
Research). Sabine Roeser’s research team members Dr Jessica Nihlén Fahlquist and
Sofia Kaliarnta were also involved in the organization, and Henneke Filiz provided
for excellent organizational support. Peter Goldie and Moritz Müller organized the
workshop ‘Emotion: Phenomenology and Content’ at the University of Manchester’s
Centre for Emotion and Value, funded by a grant from the Mind Association.
This convergence, amongst others, alerted us to the lively, and growing, philosophi-
cal interest in these issues and motivated us to further pursue the fruitful discussions
that took place at these events. This volume contains chapters by several of the speakers
at these conferences, and we would like to thank all of our contributors for providing
such original and stimulating perspectives on a range of different issues concerning
the many connections between emotion and value. We would also like to thank Peter
Momtchiloff for his support and encouragement in bringing this volume to fruition,
and the anonymous referees provided by OUP for their helpful comments at the initial
stage of planning the collection.
Our work in editing this volume has been made possible by a number of research
grants and bodies. Cain Todd’s work on the volume took place while being funded
by the Swiss National Science Foundation (SNF), at first as an official member of the
Thumos Research Group at the University of Geneva and the Swiss Centre for Affective
Science (CISA), and latterly through the SNF at the University of Fribourg on the pro-
ject ‘Imagination, Emotion, and Value’. Sabine Roeser’s work was funded by an NWO
VIDI grant on ‘Moral Emotions and Risk Politics’ (grant nr 276-20-012).
Finally, and most importantly, we must thank Peter Goldie, whose original idea it
was to put together this edited collection. He was to be one of the editors, and a con-
tributor to the volume, but sadly he passed away while the proposal was in preparation.
We dedicate this volume to the memory of Peter, whose rich and nuanced work on
emotion, and on value, has been so important and influential, as is attested by many of
viii Acknowledgements
the chapters collected here. Few have done more for philosophical research on emo-
tion. He was a wonderful writer, an inspiring and generous speaker, and a mentor and
friend for many of us working in this field. He is very much missed.
Sabine Roeser and Cain Todd
Contents
List of Contributors xi
1. Emotion and Value: Introduction 1

art I. Emotion and the Nature of Value
P
2. In What Sense Are Emotions Evaluations? 15
Julien A. Deonna and Fabrice Teroni
3. Evaluative Phenomenology 32
Michelle Montague
4. Emotion, Attention, and the Nature of Value 52
Michael Brady
5. Emotions as Unitary States 72
Jonathan Dancy
6. Relatively Fitting Emotions and Apparently Objective Values 90
Cain Todd
art II. Emotion, Evaluation, and Justification
P
7. Emotion, Evaluative Perception, and Epistemic Justification 107
Adam C. Pelser
8. Why Recalcitrant Emotions Are Not Irrational 124
Sabine A. Döring
9. Surprise 137
Adam Morton
10. Emotions Fit for Fiction 146
Greg Currie
art III. Emotion, Value, and the Self
P
11. Emotional Self-Trust 169
Linda Zagzebski
12. Self-Empathy and Moral Repair 183
Nancy Sherman
13. Emotions and the Virtues of Self-Understanding 199
Michael Lacewing
x Contents
14. Emotion and Agency 212

Jan Slaby and Philipp Wüschner
15. Evaluating Existential Despair 229
Matthew Ratcliffe
Index 247
List of Contributors
Michael Brady
University of Glasgow
Greg Currie
University of York
Jonathan Dancy
University of Reading/ University of Texas at Austin
Julien A. Deonna
University of Geneva
Sabine A. Döring
Tübingen University
Michael Lacewing
Heythrop College, University of London
Michelle Montague
University of Texas at Austin
Adam Morton
University of Alberta
Adam C. Pelser
United States Air Force Academy
Matthew Ratcliffe
Durham University
Sabine Roeser
Delft University of Technology
Nancy Sherman
Georgetown University
Jan Slaby
Free University Berlin
Fabrice Teroni
University of Bern
xii List of Contributors
Cain Todd
Lancaster University
Philipp Wüschner
Free University Berlin
Linda Zagzebski
University of Oklahoma
1
Emotion and Value
Introduction
The last few decades have witnessed a growing interest in, and expanding literature on
the emotions, in a number of domains, and from a number of different perspectives.
There have been monographs exploring the historical and cultural dimensions of the
emotions, a great deal of empirical work in psychology and neuroscience concerning
the nature and neural bases of emotion, philosophical theories about the relationship
between emotion and other mental states, and the role of the emotions in disclosing
and constituting various types of value.1 Although the primary focus of the present
collection is on this last topic, the issues touched upon are diverse and wide-ranging,
drawing on many different areas of philosophy as well as on relevant empirical work
undertaken in other disciplines. The contributions that follow all explore various
important connections between emotion and value, including the role that emotions
play in evaluative thought and experience, and in the acquisition of evaluative knowl-
edge. The nature of such experience, directed at both self and world, and its epistemic
credentials constitute the key themes that unite all of the contributions. As many of
the issues discussed have hitherto been pursued largely independently in a number
of separate philosophical domains, such as the philosophy of mind, metaethics, aes-
thetics, value theory, and the philosophy of emotions, we hope that one of the virtues
of this book will be to bring together this cutting-edge research into a more unified
overview of the field.
It is fair to say that philosophers and psychologists working on the emotions
have reached something of a consensus about the complex, dual but interrelated
nature of the affective and cognitive components of these states. Emotions are both
world-directed, disclosing what is of value in general, and yet also fundamentally
1
For a brief overview of some of this literature, see the references at the end of this introduction.
Particularly useful bibliographical information and overviews of the relevant fields can be found in
Davidson et al. (2003); Sander and Scherer (2009); Goldie (2012); Deonna and Teroni (2012).
2 Sabine Roeser and Cain Todd
self-directed, providing access to our own subjective experiential states and evaluative
stances. Perhaps most distinctively, emotions comprise and are concerned in particu-
lar with bodily states of the subject. They thus involve the heart and the head, body and
mind, the world outside and the world within. These dichotomies have been central
to the various philosophical conceptions of emotion that have previously been articu-
lated. Although few today would argue that emotions must be simply reducible to one
or the other of their non-cognitive or cognitive components, the precise make-up of
their ‘Janus-faced’ nature remains a matter of debate, and this is reflected in the diverse
roles that emotions are held to play in different areas of philosophy.2
One traditional view in metaethics, as well as in moral psychology and decision the-
ory, has been that emotions are opposed to rationality, objectivity, and justification.
Indeed, one way of categorizing metaethical theories would be under the broad head-
ings of rationalism or sentimentalism respectively.3 Rationalists generally maintain
that moral thought has to be based in reason in order to lead to justified moral judge-
ments. Sentimentalists emphasize instead the dependence of value on affective states
such as emotions, but insofar as they share the idea that emotions are non-cognitive,
they are prone to question the objectivity attributed to moral judgements by their
opponents.4
A similar dichotomy in approaches can be witnessed in decision theory. The nor-
mative approach holds that well-grounded decisions are made in accordance with
rationality, while the alternative empirical approach holds that, as a matter of fact, peo-
ple make many decisions in an emotional, unconscious, irrational way, where these
notions are equated. This latter approach has been popularized, for example, by Nobel
Prize winner Daniel Kahneman (2011), in his recent book Thinking Fast and Slow.
Empirical research in moral psychology also frequently presupposes such a dualistic
framework, for example in the highly influential work of neuropsychologist Joshua
Greene (Greene 2003, 2007) and social psychologist Jonathan Haidt (Haidt 2001;
Greene and Haidt 2002). These scholars argue that because emotions play an important
role in moral judgements, this undermines the credibility of these judgements, based
as they supposedly are on irrational, unconscious motives. These views also resonate
in the common appeals to be rational and not to be emotional in decision-making,
since being emotional is by definition taken to be inferior.
In contrast with such pessimistic assessments of the connection between emotion
and rationality, however, many contemporary emotion scholars in philosophy and
psychology now think that emotions are an important source of practical rationality.
After a long period of relative neglect by analytic philosophers, Robert Solomon (1976)
2
The apt description of emotions as ‘Janus-faced’ comes from de Sousa (2007).
3
Clearly, of course, this is an oversimplification even of the main fault-line that runs between Humean
and Kantian approaches to ethics, but it does reflect important differences in the ways emotions have been
viewed in metaethical theories.
4
For a rich and sophisticated account of these issues, and in particular how sentimentalism fails to map
neatly onto the divide between realists and expressivists in metaethics, see D’Arms and Jacobson (2006).
Emotion and Value: Introduction 3
broke new ground and initiated a wave of philosophical research into the emotions, by
arguing that they are cognitive states, specifically judgements of value. The attribution
of a cognitive nature to emotions has been highly influential and has informed the
majority of philosophical accounts of the emotions ever since.5 Outside the confines of
philosophy, the research of the neuropsychologist Antonio Damasio (1994), for exam-
ple, has been highly influential. He showed that people with damage to their amygdala
in the temporal lobe lose their capacity to feel emotions and at the same time are no
longer capable of making concrete moral and practical judgements.
It is thus now widely accepted that emotions play an important epistemological role
in evaluative thought and experience. However, moving away from narrowly centred
judgement-based theories, philosophical discussion has more recently been keen to
stress that unlike merely intellectual, propositional evaluative knowledge, emotions
uniquely provide (or purport to provide) direct experiential acquaintance with evalu-
ative states of affairs. Their distinctive phenomenal character, bodily feelings, valence,
motivational pull, and world-directed intentionality, all serve to connect us, as it were,
directly with the evaluative world.6 The values they have been claimed to reveal, or to
constitute, range from the aesthetic and moral to the epistemic, psychological, and
self-reflective. Naturally, the questions of whether and how emotions perform this
function, and the exact scope of the relevant values encompassed, are at the centre of
deep philosophical dispute; and hence at the heart of this volume.
One common theme in the recent literature has been an appeal to perception as a use-
ful model for illuminating emotion. Indeed, perceptual theories have come to encom-
pass both accounts of emotion that emphasize these states’ cognitive components (e.g.
Roberts 2003), as well as accounts that emphasize their essentially bodily nature (e.g.
Prinz 2004). Such theories are attractive insofar as our perceptual states too are world-
directed and yet possess a distinctive phenomenal character; and, it has seemed to
many, these two features are somehow intrinsically connected. Emotional phenomenal
character and intentional content, like perceptual character and content, seem to go
hand-in-hand. Moreover, philosophers sympathetic to the perceptual model have been
keen to emphasize purported similarities between the epistemic roles of perception and
emotion—the former in justifying sensory knowledge of the world, the latter in justify-
ing evaluative knowledge.7 In terms of the general analogy with perception, it is also
worth noting that philosophy here echoes certain developments in affective science,
where ‘appraisal theories’ of emotion have exercised a recent dominance.8
Representatives of these various positions on the connection between emotion and
value can be found in several contributions to this book, such as those by Döring and
5
Prominent treatments include, for example: Rorty (1980); Lyons (1980); de Sousa (1987); Gordon (1987);
Greenspan (1988); Nussbaum (2001).
6
Amongst such works are: Goldie (2000); Tappolet (2000); Roberts (2003); Prinz (2004); Helm (2007);
Roeser (2011).
7
For an excellent overview of perceptual theories see Deonna and Teroni (2012: ch. 6).
8
For an overview see Davidson et al. (2003); Sander and Scherer (2009).
Pelser, as too can objections to such views. For, inevitably, such accounts of emotion
have their critics, who are keen to undermine the supposed analogies between per-
ception and emotion in terms of the relationship between intentional content and
phenomenal character, as well as to highlight the poorer epistemic credentials of the
emotions in grounding evaluative knowledge. Alternative positive accounts of the
evaluative nature of emotion can be found in the chapters by Deonna and Teroni,
Brady, Dancy, Todd, and Montague in this volume.
In addition to addressing these numerous general connections between emo-
tion and value, sometimes at a level of relative abstraction, much of the work under-
taken in this volume has important implications for current research in philosophical
domains that are directly concerned with specific types of value (and sometimes with
specific emotions) such as those present in metaethics, moral psychology, and aesthet-
ics. Prominent philosophical issues here include the role that emotions play in practi-
cal rationality, in self-understanding and well-being, in moral judgement, and in the
appreciation of fiction. These apparently disparate themes are all addressed in several
of the contributions to the volume, and are bound by one particular thread, namely,
the normative nature of emotions in general—a topic approached under many differ-
ent guises depending on subject matter and inclination: ‘appropriateness’, ‘fittingness’,
‘truth’, ‘objectivity’, and ‘rationality’ are all notions employed by philosophers in trying
to understand the particular normative force of the connection between emotion and
value. We leave it for the reader to try and trace this often delicate and complex thread
throughout the subsequent chapters in more detail, but in the remainder of this intro-
duction we will provide a brief overview of the chapters and highlight some of the sali-
ent interconnections.
The book is organized as follows. Part I addresses the role of emotion in our under-
standing of the nature of value. Part II examines the possible justificatory role emo-
tions may play in evaluation. Part III examines the special role that emotions play in
relation to the self, specifically to self-evaluation and self-experience.
Part I: Emotion and the Nature of Value

Whether and how emotions resemble perception, how emotional experiences repre-
sent value, and the nature of the connection between the various elements constituting
emotional phenomenology and evaluative intentional content are all issues pursued in
the chapters collected in Part I of this volume.
In the chapter, ‘In What Sense are Emotions Evaluations?’, Julien Deonna and
Fabrice Teroni consider whether emotions are, as many have maintained, kinds of
evaluations. Addressing some of the main theories supporting such a view, including
the perceptual theory of emotions, they argue that each is unsatisfactory. According
to Deonna and Teroni, this is in large part because of a mistaken presupposition on
which such theories rest, namely that emotions possess only one type of content. In
response, they outline their own novel theory of emotion—the attitudinal theory—
which, they claim, avoids the various problems afflicting other accounts by holding
that emotions should not be conceived as evaluations in terms of what they represent,
but rather in terms of the sort of attitude subjects take towards what is represented.
They then explore what sorts of attitudes emotions are and defend the idea that they
are felt bodily attitudes.
Michelle Montague, in her chapter ‘Evaluative Phenomenology’, also defends the
idea that emotions are evaluations, but without assimilating them to perceptions or
traditional propositional attitudes. She argues for the strong claim that emotions are
essentially experiential evaluative representations. In doing so, however, she rejects
the common idea that emotional phenomenology can be reduced to some type of
sensory or bodily phenomenology, and holds instead that emotions have their own
distinctive sui generis kind of phenomenology, which she calls ‘evaluative phenom-
enology’. Recalling some of the key themes raised by Deonna and Teroni, the novelty
of Montague’s position lies in tying an emotion’s evaluative phenomenology inextri-
cably to its intentionality, and thereby illuminating the sui generis way in which emo-
tions represent evaluative properties. In the process, she offers a detailed account
of the nature of evaluative phenomenology, and how it differs from other types of
phenomenology.
Drawing on recent empirical evidence, Michael Brady’s chapter ‘Emotion, Attention,
and the Nature of Value’ explores how attention impacts upon the accuracy of the
emotional assessments of one’s evaluative situation. In particular, he is concerned to
highlight differences between positive and negative emotional responses in account-
ing for the different effects of valence on attentional breadth. Broadening his theme
to encompass the nature of emotional evaluation, Brady claims that the empirical
evidence lends some support to Nozick’s view that positive affect involves seeing the
world or one’s life as unified, integrated, and coherent, and that negative affect involves
seeing these things as fragmented, and lacking in unity or harmony. An important
implication, Brady suggests, is that value is a matter of organic unity, which in turn
supports the idea that a useful coping strategy when bad things happen is to attempt to
integrate such things into the pattern of one’s life. In examining this important connec-
tion between emotion and attention, therefore, Brady contends that we thereby learn
something important about the nature of the values and the content of the evaluations
associated with emotional experience.
In his contribution, ‘Emotions as Unitary States’, Jonathan Dancy tackles the funda-
mental issue of whether emotions are unitary states. This is a difficult problem given
that emotions seem to contain many different components, including feelings, beliefs,
perceptions, and desires. Arguing that various current proposals are unsatisfactory,
Dancy suggests that the assembly of elements constituting the moral emotions are uni-
fied by a normative relation. Agreeing with Peter Goldie’s account that emotions are
plausibly thought of as processes, he argues that the coherence of such states consists in
a particular reason-relation that is articulated in terms of the appropriateness of each

of an emotion’s elements to the circumstances that concern it.
This normative notion of appropriateness resurfaces in several of the following
chapters (specifically in those of Todd, Döring, Currie, and Zagzebski) and is a key
component of so-called ‘fitting attitude’ or ‘sentimentalist’ analyses of value, in which
emotions and other affective states are held to play a central role. Parallel to, but hith-
erto largely independent of the discussions concerning the nature and epistemic status
of emotional evaluative experiences, an increasing number of philosophers working
in value theory have become attracted to these analyses in which it is held that value
or evaluative concepts can be somehow illuminated in terms of appropriate or fitting
affective or emotional responses.
Cain Todd’s chapter ‘Relatively Fitting Emotions and Apparently Objective Values’
examines the connection posited between emotion and value by so-called sentimental-
ist theories. Todd aims to show that such theories are threatened by relativism insofar
as, contrary to their aim, they fail to offer an informative way of objectively specify-
ing the nature of ‘fittingness’ when confronted with essentially contestable evaluative
concepts. He examines one recent attempt to avoid these problems, proposed by Justin
D’Arms, which relies on the class of natural emotions to provide a plausible analysis of
certain evaluative concepts. Todd argues, however, that D’Arms’s ‘rational sentimen-
talism’ rests on an inaccurate conception of emotional phenomenology and to that
extent fails to secure the sentimentalist enterprise of specifying a notion of objective
fittingness. In conclusion, Todd offers an alternative account of the phenomenology
of emotions—one which appeals to the notion of ‘apparent objectivity’, but which has
as a consequence that any version of sentimentalism based upon the emotions must
contend with a relativistic notion of fittingness.
How the notion of fittingness discussed here relates to the notion of epistemic justi-
fication and to the rationality of our emotional responses is an important question that
resurfaces in some of the discussions in Part II and Part III.
Part II: Emotion, Evaluation, and Justification

Some of the epistemic issues touched upon in the contributions to Part I of the vol-
ume become the primary focus of attention in several of the chapters in Part II, which
is broadly concerned with the general epistemology of emotions, encompassing two
separate but related topics: (i) the role of emotions in justifying evaluative judgements
and grounding evaluative knowledge; (ii) justifying the appropriateness of certain
emotional responses themselves.
Adam C. Pelser’s chapter ‘Emotion, Evaluative Perception, and Epistemic
Justification’ is concerned with the epistemic status and role of emotion in evaluation.
Discussing perceptual theories of emotion, he focuses on the purported resemblance
between emotion and sense perception insofar as both states are often taken to give
rise to beliefs. Noting that such beliefs are generally trusted, he is chiefly concerned
with the question of whether, in the case of emotion experience, the resulting beliefs
are ever epistemically justified. Confronting certain sceptical views, Pelser argues that
emotion is indeed a basic source of epistemic justification. More specifically, he claims
that emotions themselves can non-inferentially confer justification on the beliefs to
which they give rise. Labelling this claim the justificatory thesis, Pelser offers a sus-
tained and systematic defence of the epistemic justification of emotion-based beliefs,
and in the process upholds some form of the perceptual theory of emotion.
Sabine Döring too, in her chapter ‘Why Recalcitrant Emotions Are Not Irrational’,
defends a perceptual model of emotion, focusing on an influential objection to
this model posed by Bennett Helm, who points out that while recalcitrant percep-
tions are not irrational, recalcitrant emotions are. Döring argues against this that
although recalcitrant emotions share a cognitive conflict between experience and
judgement with recalcitrant perceptions, in neither case does such conflict amount
to self-contradiction, and hence suffice for irrationality. Rather, she contends, recal-
citrant emotions typically lead to practical conflict, and this is what explains the mis-
guided intuition that they are therefore irrational. Offering a complex exploration of
the issues of justification, cognitive conflict, and rationality, Döring’s chapter touches
on recurrent themes in several of the chapters in the final section of the volume.
While discussion of the issues raised in the previous chapters has, where it has been
relevant, primarily been concerned with moral emotions, the final contributions in
this section, by Adam Morton and Greg Currie, focus on different kinds of emotions.
In his chapter ‘Surprise’, Adam Morton addresses the nature of surprise and focuses
on its rationality, specifically on the question of why it sometimes make sense to be
surprised when improbable things occur. Morton is also interested in understanding
the value and disvalue of surprise, given that we both desire and fear the unexpected.
In discussing the contrastive nature of this and similar emotional states, and the phe-
nomenology of surprise, his chapter touches on the nature of emotional justification in
general and its connection to other mental states, such as beliefs and desires.
Greg Currie’s chapter on ‘Emotions Fit for Fiction’ offers an important additional ori-
entation, to the domain of fiction and aesthetic value. Currie explores the similarities
and differences between the nature and appropriateness of our emotional responses
to fiction and towards the real world. Holding that fictions are representations of real-
ity rather than alternatives to it, he argues that the emotions we have in response to
fictions are appropriate because of how things are represented, not because of how
they are. In this crucial respect they differ from the emotions we direct at events and
things in the real world. However, Currie notes, there are many ways for emotions to
be appropriate, and in one of them truth sometimes matters even for fictive emotions.
Specifically, it matters when we have reason to think that a representation is confusing
us about what is true, and getting what he calls ‘an emotional free-ride’ in consequence.
He argues, in conclusion, that the responses we are intended to have to a fictional rep-
resentation can be very far indeed from the responses we would, or should, have to
similar events in the real world, even if they come to us via testimony or some other
form of representation. Yet the notion of appropriate emotional responses to fiction
too is intrinsically connected to the evaluative nature of such responses
Currie’s chapter thus remains directly engaged with many of the themes raised in
Part II: with the appropriateness of emotional responses, with their phenomenology,
and with the role they play in justifying and explaining our evaluative reactions.
Part III: Emotion, Value, and the Self

This part explores the role that emotions play in the evaluation and experience of the
self. The chapters included here concern the epistemic legitimacy of the first-person
point of view and the role emotions can play in undermining, as much as strengthen-
ing, that point of view. Most of the chapters in this part are also concerned with psy-
chological aspects of emotions; for example, those aspects that are central to cases of
trauma, guilt, and existential despair.
In her chapter ‘Emotional Self-Trust’ Linda Zagzebski defends the rationality of
emotional self-trust, analogous to an account she has developed on epistemic self-
trust. She holds that just as beliefs can be true or false, emotions can be appropriate and
inappropriate, fitting or unfitting. Zagzebski argues that the fittingness of emotions
cannot be reduced to the fittingness of beliefs; specifically, emotions entail an evalu-
ative perspective that cannot be reduced to beliefs. When it comes to the justification
of our emotions, we have to follow similar patterns as in the justification of our beliefs.
In the end, all self-reflection and justification is circular, as we cannot prove its reli-
ability from an independent source. We have to start justification by trusting certain of
our beliefs and emotions; we cannot start from a void. A responsible epistemic agent
employs what Zagzebski calls ‘epistemic conscientiousness’, which involves trying our
best to make our emotions fit their objects in a way that is parallel to epistemic consci-
entiousness, which involves trying to make our beliefs true. In both cases we have no
non-circular guarantee that our faculties fit their objects. This is where we have to trust
their disposition to be fitting. As noted previously this notion of fittingness is a key one
in understanding the nature of emotions and their epistemic role, and one that directly
connects Zagzebski’s chapter with several other chapters in the volume.
Nancy Sherman’s chapter ‘Self-Empathy and Moral Repair’ too is concerned with
the first-person point of view, specifically with self-empathy and how it can contrib-
ute to overcoming moral injury after war. In cases of traumatic experiences, it can be
difficult to accept one’s actions and emotions. Self-empathy, Sherman contends, can
help to cope with feelings of shame and guilt. In order to illustrate these ideas, she
discusses two examples of soldiers who have experienced traumatic dilemmas at war.
The first example is that of a major who tries to behave in a humane way to the surviv-
ing family members of civilian casualties, but his attempts to show his compassion
are disrupted by bureaucracy. He experiences the resulting shame as more damaging
to his agency than his experiences in combat. The second example is that of a captain
who feels shame for not having prevented the death of a member of his group through
an accident. Even though these soldiers were not directly to blame for what happened,
they nevertheless felt responsible. Sherman argues that the notion of self-empathy can
help one to adopt a benevolent, forgiving stance towards oneself that can contribute to
coping with such difficult experiences and negative emotions.
Similar themes of self-assessment and perspective are central to Michael Lacewing’s
chapter ‘Emotions and the Virtues of Self-Understanding’, which studies the impor-
tant but hitherto largely neglected relation between moral enquiry and developing
self-understanding. Lacewing, like Zagzebski, is concerned with the trustworthiness
of our own perspective, and he sees emotions as an important source of our knowledge
of reasons for actions. However, he notes, they can also be misleading in unconscious
ways that are hard to access, for example, due to certain defence mechanisms. The
question then is how such mechanisms can be deconstructed in order to improve the
epistemic role of emotions. Lacewing discusses the importance of intellectual cour-
age and communal enquiry to this task, focusing on the connection between defence
mechanisms and the rejection of parts of the self deemed inappropriate. He argues,
in a way that echoes Sherman’s main thesis, that acceptance and compassion can play
important roles in deconstructing defence mechanisms and hence in improving and
enriching moral reflection.
The chapter ‘Emotion and Agency’ by Jan Slaby and Philipp Wüschner discusses
metaphysical concerns with the role of emotions in constituting value that connects it to
the chapters in Part I. However, its main focus is on how value is related to the perspec-
tive of agents as caring subjects, as subjects for whom things matter. They thus examine
the connection between emotions, caring, and existential commitment. Specifically,
they argue that human emotions are active engagements with the world rather than
passively undergone experiences. The authors develop an account according to which
emotions constitute as well as detect value. They argue, further, that we can make sense
of this apparent paradox by considering the point of view of a socially engaged person
and the way in which she is emotionally situated. Understanding the phenomenology
of an emotion is crucial to making sense of the way that emotions impact on our agency
and how that in turn contributes to our further emotional experiences. They close their
discussion by reflecting on what they call the active-affective ‘minimal self ’ and how its
erosion in, for example, severe depression supports their account.
This leads us to the final chapter of the volume, ‘Evaluating Existential Despair’ by
Matthew Ratcliffe, which has depression as its key focus and incorporates a detailed
analysis of passages from Tolstoy’s memoir, A Confession. Ratcliffe explores the seem-
ingly revelatory nature of some depression experiences and, in the process, draws
attention to various important connections between evaluative beliefs and affec-
tive feelings. He starts his chapter by discussing a posthumously published essay by
Peter Goldie, in which Goldie discusses the experience of someone whose intellectual
life has ‘gone cold’. This person still knows that a book is worth reading, but lacks the
emotional engagement necessary for setting oneself to read, write, and develop one’s
ideas. Ratcliffe expands on this to offer a broader discussion of the phenomenology of
‘existential despair’, which he understands as ‘a painful sense that no human activity
of any kind could ever be of any worth’. He raises the pressing question of whether or
not such a predicament involves a correct evaluation of human life, and offers a partial
response to the challenge of existential despair by focusing on distinctively interper-
sonal forms of concern.
Conclusion
The contributions to this volume testify to the rich, complex nature of emotions and
their close, manifold links with a range of different values. They demonstrate not
merely the growing significance of current philosophical research in this area, but
also that many of the purported traditional dichotomies between, for example, rea-
son, objectivity, and justification on the one hand, and irrationality, subjectivity, and
unwarrantedness on the other, are no longer viable. Emotions, whatever their limita-
tions, are an indispensable source of evaluative thought and experience, a key com-
ponent of practical rationality, and a crucial ingredient in self-understanding and
well-being. We hope that this volume provides some inspiration for future research in
this rapidly expanding area, but also that, given the centrality of emotions to our evalu-
ative conception of ourselves and our world, many outside the confines of academia
may draw some insights from the chapters within.
References
Damasio, A. (1994). Descartes’ Error: Emotion, Reason, and the Human Brain (New York:
Putnam’s).
D’Arms, J. and Jacobson, D. (2006). ‘Sensibility Theory and Projectivism’. In The Oxford
Handbook of Ethical Theory, ed. D. Copp (Oxford: Oxford University Press), 186‒218.
Davidson, R., Scherer, K., and Goldsmith, H. (eds.) (2003). The Oxford Handbook of Affective
Sciences (Oxford: Oxford University Press).
de Sousa, R. (1987). The Rationality of Emotion (Cambridge, MA: MIT Press).
de Sousa, R. (2007). ‘Truth, Authenticity, and Rationality’. Dialectica 61: 323–45.
Deonna, J. and Teroni, F. (2012). The Emotions: A Philosophical Introduction (London: Routledge).
Goldie, P. (2000) The Emotions: A Philosophical Exploration (Oxford: Oxford University Press).
Goldie, P. (ed.) (2012). The Oxford Handbook of the Philosophy of Emotion (Oxford: Oxford
University Press).
Gordon, R. (1987). The Structure of Emotions: Investigations in Cognitive Philosophy
(Cambridge: Cambridge University Press).
Greene, J. D. (2003). ‘From Neural “Is” to Moral “Ought”: What Are the Moral Implications of
Neuroscientific Moral Psychology?’ Nature Reviews Neuroscience 4: 847–50.
Greene, J. D. (2007). ‘The Secret Joke of Kant’s Soul’. In Moral Psychology, Vol. 3: The Neuroscience
of Morality: Emotion, Disease, and Development, ed. W. Sinnott-Armstrong (Cambridge,
MA: MIT Press), 2–79.
Greene, J. D. and Haidt, J. (2002). ‘How (and Where) Does Moral Judgment Work?’ Trends in
Cognitive Sciences 6: 517–23.
Greenspan, P. (1988). Emotions and Reasons: An Inquiry into Emotional Justification
(New York: Routledge).
Haidt, J. (2001). ‘The Emotional Dog and its Rational Tail: A Social Intuitionist Approach to
Moral Judgment’. Psychological Review 108: 814–34.
Helm, B. (2007). Emotional Reason: Deliberation, Motivation, and the Nature of Value
Kahneman, D. (2011). Thinking Fast and Slow (New York: Farrar, Straus & Giroux).
Lyons, W. (1980). Emotion (Cambridge: Cambridge University Press).
Nussbaum, M. (2001). Upheavals of Thought: The Intelligence of Emotions (Cambridge: Cambridge
University Press).
Prinz, J. (2004). Gut Reactions: A Perceptual Theory of Emotion (Oxford: Oxford University
Press).
Roberts, R. (2003). Emotions: An Essay in Aid of Moral Psychology (Cambridge: Cambridge
University Press).
Roeser, S. (2011). Moral Emotions and Intuitions (Basingstoke: Palgrave Macmillan).
Rorty, A. (ed.) (1980). Explaining Emotions (Berkeley: University of California Press).
Sander, D. and Scherer, K. (eds.) (2009). The Oxford Companion to Emotion and the Affective
Sciences (Oxford: Oxford University Press).
Solomon, R. (1976). The Passions: The Myth and Nature of Human Emotion (Garden City,
NJ: Anchor Press).
Tappolet, C. (2000). Emotions et Valeurs (Paris: Presses Universitaires de France).
PA RT I
Emotion and the Nature
of Value
2
In What Sense Are Emotions
Evaluations?
Julien A. Deonna and Fabrice Teroni
Introduction
Why think that emotions are ways of evaluating? This chapter puts forward an origi-
nal account of emotions as evaluations apt to circumvent some of the chief difficulties
with which alternative approaches find themselves confronted. We shall proceed by
first introducing the idea that emotions are evaluations (sec. I). Next, two well-known
approaches attempting to account for this idea in terms of attitudes that are in and of
themselves unemotional but are alleged to become emotional when directed towards
evaluative contents are explored. According to the first approach, emotions are noth-
ing but evaluative judgments. Sec. II reminds the reader of the problems associated
with this idea: one of its consequences is to deprive creatures with limited cognitive
capacities of any sort of partaking of emotional life. According to the second approach,
which is often praised for its capacity to avoid the pitfalls facing an appeal to evaluative
judgments, emotions are perception-like experiences of evaluative properties and are
as such within the reach of creatures bereft of conceptual capacities. This perceptual
theory is taken up in sect. III, in which we explain why it remains unsatisfactory inso-
far as it shares with the evaluative judgment theory the idea that what makes emotions
evaluations is the specific contents that they have. On this basis, we proceed by outlin-
ing in sect. IV an alternative—the attitudinal theory of emotions. Its main point of
departure from current theorizing about the emotions consists in elucidating the fact
that emotions are evaluations not in terms of what they represent, but rather in terms
of the sort of attitude subjects take towards what they represent. We explore here what
sorts of attitudes emotions are and defend the idea that they are felt bodily attitudes.
I. Emotion as Access to Value

Why is the idea that emotions are apprehensions of values so attractive? Ordinary lan-
guage certainly vouchsafes for the existence of a close connection between emotions
16 Julien A. Deonna and Fabrice Teroni
and values. Several evaluative words thus appear to derive from words referring to
types of emotions: we speak of the disgusting, the amusing, the regretful, the admira-
ble, the shameful, and so forth.1 These linguistic data constitute some suggestive evi-
dence in favour of the idea that emotions relate to values. Still, what does this mean
exactly? The answer requires that we have at least a working characterization of the
notion of value at stake as well as a preliminary understanding of how values relate to
emotions. Let us consider these issues in turn.
The idea we are trying to capture appeals to values in a way that, although not strictly
technical, diverges from that in which it is used in ordinary language. People typically
talk of values in connection with abstract ideals driving their lives. It is in this sense
that friendship and solidarity may be thought of as paradigmatic values. These, how-
ever, are not the sort of entities that we are concerned with here; by “value,” we rather
designate a specific type of property which may be instantiated by objects, events,
or situations. From this viewpoint, a joke exemplifies the value of being amusing or
humorous, the disappearance of a loved one, that of constituting a loss, and so on. The
idea is thus that emotions relate to exemplifications of properties of this type, and not
to abstract ideals. Moreover, as some of the previous examples suggest, the evaluative
properties at issue are, by contrast with abstract ideals, susceptible of being either posi-
tive (the amusing, the admirable) or negative (the shameful, the disgusting).
Let us now try to provide a preliminary understanding of the manner in which emo-
tions relate to these evaluative properties. Imagine a subject finding herself on a hunt-
ing expedition. Sight and hearing allow her to apprehend an indistinct form only a few
yards away and the sound of heavy breathing, while the testimony of the gamekeeper
allows her to apprehend that she is about to be attacked by a wild boar. The core idea
shared by the approaches according to which there is an intimate connection between
emotions and values is captured by the idea that one can pursue the enumeration just
laid down as follows: and fear allows the subject to apprehend the boar’s dangerousness.
More generally, the idea is that there is an intentional relation between one’s emotions
and the values exemplified by the objects and events one may confront: different types
of emotions are apprehensions of distinct values. Thus, fright will be the apprehension
of the dangerous, anger the apprehension of the offensive, amusement the apprehen-
sion of the comical, and so forth.
The foregoing has, we hope, conveyed the sort of intuitive appeal that an approach
emphasizing the existence of a relation between emotions and values may claim. What
about its theoretical virtues? Let us begin by observing that the approach is able to give
convincing answers to some of the fundamental questions facing any theory of the
emotions. Indeed, its friends are likely to insist that appealing to values allows for the
individuation, rationalization, and assessment of emotions (see e.g. Teroni 2007).
1
This point is aptly emphasized by Mulligan (1998) and Tappolet (2000). A typology of emotions based
on their connection to values is illuminatingly laid out by Roberts (2003).
In What Sense Are Emotions Evaluations? 17
First, values have the potential to individuate emotion types insofar as they consti-
tute something that is shared by the various objects of a given emotion type. Although
different episodes of sadness may be related to particular objects as disparate as the
disappearance of a loved one, failure at an exam, the melting of a glacier, and so on, the
subject who feels sad nevertheless apprehends all these objects as losses. This is to say
that emotion types are unified at the level of their evaluative intentionality, and that we
can distinguish them by means of the distinct values to which they relate.
Second, an appeal to values allows for measuring whether or not emotions are
rational or intelligible. Observe that we understand how very different situations
can elicit emotions of the same type once we realize that, from the subject’s point of
view, they each exemplify the same evaluative property. This being the case, a given
emotion becomes intelligible once we figure out how such a point of view may have
been fostered by the circumstances (“After all, reaching out to him might have been
understood as an offense and might thus have made him angry”); otherwise it remains
unintelligible.
Finally, values play an essential role in the way we assess emotions. After all, we
are naturally inclined—and for good reasons!—to treat some emotional episodes as
appropriate, others as inappropriate. Within the present framework, this contrast is
primarily to be understood in terms of the opposition between emotions directed at
objects or situations that do exemplify the relevant values, and emotions directed at
objects or situations that do not exemplify them. Insofar as a remark was innocent,
for instance, the anger it has triggered is inappropriate. This is why it makes sense to
speak of the evaluative conditions of correctness of the emotions—a notion to which
we shall return repeatedly in what follows.2 The underlying idea consists in approach-
ing the emotions in a way that is routine in philosophical discussions of other types
of mental states. A visual experience of a red kite on the roof of the house has specific
conditions of correctness (simplifying a bit: the presence of such a kite at that par-
ticular location); analogously, an episode of fear directed at the danger represented by
a narrow mountain trail has specific evaluative conditions of correctness (the path’s
dangerousness). In both cases, the representation may turn out to be either correct or
incorrect, depending on whether or not the world actually is the way it is represented
to be.3
Philosophers have in mind precisely these three distinct roles that values may play
when they treat them as the formal objects of the emotions—yet another notion to
2
While it is perhaps natural to speak of the emotions as appropriate or inappropriate, we do not ordi-
narily speak of them as correct or incorrect. The latter are terms of art, then, and correspond to the sort of
epistemic assessment to which the emotions are subject, as opposed to other ways (prudential, moral, and so
on) in which they might also be assessed (see D’Arms and Jacobson 2000).
3
Let us observe that, if the idea of evaluative conditions of correctness invoked here is incompatible with
strong subjectivism about evaluative properties, it is perfectly compatible with quite a wide range of alterna-
tive ontological claims (such as sophisticated dispositionalist accounts as well as fitting-attitudes and realist
accounts).
which we shall return in what follows.4 However, in order to account for this as well as
for many other fundamental claims regarding the emotions, we need first of all to get
a firmer grip on a notion central to the previous observations: that of a “way of appre-
hending value.”
II. The Evaluative Judgment Theory

We are all familiar with the distinction between attitude and content.5 Mary may
for instance either believe that it is raining or believe that the weather is nice; that
is, she may have the same attitude towards two different contents. It is of course
also possible to have two different attitudes towards the same content: Mary comes
to believe that the weather is nice after having looked out of the window, whereas
Jean, who has not opened the blinds yet, desires that to be the case. It is natural to
approach the emotions in terms of that same contrast. After all, not only can we
believe or desire that it rains, we can also hope, fear, or regret it. Natural as it may
be, this idea has not been pursued at all within recent philosophical approaches to
the emotions. This is to say that, instead of trying to see where the idea that emo-
tions are distinctive sorts of attitudes might lead, the general trend has consisted in
trying to account for the emotions in terms of this or that already familiar attitude
and, correlatively, to leave to their specific contents the task of accounting for what
is emotional about them.
This is especially manifest in a conception that was already advocated in antiquity,
according to which emotions relate to values because they consist in evaluative occur-
rent beliefs or judgments. Types of emotions are then distinguished from one another
in terms of the different value attributions that feature in the contents of these judg-
ments.6 Thus, judging that one is in a dangerous situation is to be afraid of it, judg-
ing that one has committed a fault is to feel guilty about it, and so on. This approach
not only allows for the individuation of emotions by means of the evaluative concepts
deployed in the relevant judgments, it can also draw the needed distinctions between
intelligible, unintelligible, appropriate, and inappropriate emotions—it of course rec-
ommends to understand these two contrasts in the light of whatever constraints bear
4
Kenny (1963) introduces the idea of formal object into the contemporary debate. The various roles for-
mal objects may play in connection to the emotions are discussed in Teroni (2007). For skepticism about this
idea, see Deigh (1994).
5
Henceforth, we shall use “attitude” where others use “mode.” The choice is governed by the idea that
thinking of the different emotions as different attitudes we take towards the world is, as we shall argue, just
right. Being typically used within the expression “propositional attitude,” however, it has the drawback of
encouraging the claim, which we shall reject, that emotions are attitudes exclusively directed at proposi-
tions. This drawback is largely compensated, we believe, by the benefits of avoiding any confusion that might
arise between the idea of “mode” and that of “modes of presentation.” The latter, which has to do with dis-
tinctions amongst contents, will play no role in what is to come.
6
This theory, which is often said to have been defended by the Stoics, was brought back into fashion by
Nussbaum (1994) and Solomon (1988).
on the intelligibility and appropriateness of the evaluative judgments at issue. As is

widely acknowledged, however, conceiving of emotions as judgment-like attitudes
directed towards evaluative contents is hardly satisfactory. Revisiting some of the dif-
ficulties from the vantage point offered by the attitude vs. content contrast will provide
important clues on the way to a satisfactory account.
First, the proposed analysis simply does not lay out sufficient conditions for having
an emotion. It is quite common for a subject to make an evaluative judgment without
thereby being emotionally affected in the way she would have to be if the proposal were
along the right track. It is not, since judging that it is dangerous to live in Tokyo is obvi-
ously quite different from being afraid of living there. In addition, the theory cannot
account for the fact that emotions may sometimes be caused by evaluative judgments
(“Max was frightened because he had judged the situation to be dangerous”); for this
cannot be reconciled with the purported equivalence between the two. All this is illus-
trative of a more general point: the theory cannot do justice to the phenomenological
aspect of emotions—a point we shall return to.
Second, if we think that young infants and animals are capable of feeling emotions,
it is then difficult to maintain that this requires the making of evaluative judgments.
For the capacity to make judgments of this nature may be exclusive to humans who
have reached a certain stage of their cognitive development (e.g. Deigh 1994; Tappolet
2000). Let us briefly rehearse why this may be thought to be the case. According to
many, the attitude of judging, regardless of how precisely we spell that out, amounts
to asserting a propositional content or, to put it in slightly different words, to com-
mitting oneself to the truth of a given proposition. If this is the case, then judging
requires, on the one hand, the capacity to think a proposition and, on the other hand,
the capacity to commit oneself to its truth. To judge that the weather is nice thus
requires the capacity to think that the weather is nice together with the capacity to
endorse this as a fact. It is legitimate to doubt that animals and young infants are
really capable of such cognitive prowess. After all, to commit oneself to the truth of a
proposition seems to require, if not the general capacity to weigh the reasons bearing
on judgments, at least the capacity to refrain making such judgments when circum-
stances require. And such capacities do seem to be a human prerogative. This is one
reason why one may want to resist appealing to judgments in trying to understand
emotions.
We may be running ahead of ourselves, however. Perhaps the evaluative judg-
ment theory faces these difficulties chiefly because it misdiagnoses the exact nature
of emotional attitudes—judging may simply be a near miss. There are indeed popular
approaches that are closely related to the judgment theory but which deny that subjects
must commit to the truth of the relevant proposition. And, for all we have said so far,
those might very well fit the bill. If so, what sort of attitude would do the trick? Patricia
Greenspan sets her sights on the attitude of entertaining a proposition; Robert Roberts
sails in the same waters when he argues that the desired attitude is that of construing
certain facts in the light of a given value (Mary can thus construe a dog as dangerous
without thereby asserting that it is such).7 It stands to reason that this strategy allows
for a more satisfying treatment of a certain range of cases. For example, one may then
treat the situation in which, say, a subject judges that a given object is not dangerous
while being nevertheless afraid of it without having to invoke a severe form of irra-
tionality. He no longer needs to be understood as making contradictory judgments; he
may much more plausibly be said to judge that the object does not exemplify a given
value while still construing it in terms of that same value (Roberts 2003: 89–93).
That being said, this change of attitude within the framework set by the evaluative
judgment theory does not really succeed in solving the problem. After all, the require-
ment that the subject must be capable of having the relevant proposition in mind when
undergoing the emotion remains in full force. And creatures with limited cognitive
capacities may well prove unable to achieve that much, regardless of any difficulty associ-
ated with the additional requirement that one is committed to its truth.8 A proposition, as
it is generally understood within the relevant literature, is an articulated sequence of con-
cepts that is truth-assessable. As a consequence, to have a proposition in mind requires
the capacity to deploy the concepts which constitute it.9 Someone cannot, for instance,
believe that one’s husband plays the bagpipe without having the concepts of “husband”
and of “bagpipe.” Furthermore, given what is involved in mastering and deploying con-
cepts, the capacity in question is cognitively quite demanding. Mastery of a concept
seems in fact to require of a subject that she master at least some of the epistemic, seman-
tic, and logical relations tying the concept in question to other concepts.10 A subject who
masters the concept of red, for instance, must first be capable of distinguishing the cir-
cumstances in which its application is justified from those in which it is not. This is made
manifest, for instance, in her attentiveness to varying conditions of illumination that may
potentially defeat her color judgments. Moreover, such a subject must display an under-
standing of some of the relations the concept in question bears to other concepts like, say,
the concept of color (what is red cannot be green; what is red is colored) and that of a sur-
face. This explains several aspects of the normativity and rationality of conceptual states;
it underscores, for instance, the idea that a subject who makes a given judgment must
either accept its logical and epistemic implications, or otherwise agree to revise it once
she becomes aware of the difficulties it raises. Conceptual thinking is thus cognitively
7
See Greenspan (1988) and Roberts (1988, 2003: 92). Roberts emphasizes the fact that the attitude specific
to emotions does not involve an assertion of the content. He attempts a positive characterization of the emo-
tional attitude as one in which the subject views its content as having the “appearance of truth.”
8
Roberts insists on the fact that the open variable in “construe a as x” can in principle be occupied by any
type of representation (images, concepts, or even perceived objects). Where values are concerned, however,
“construing a as x” must entail a subject deploying her conceptual capacities. While Roberts advertises his
theory as a perceptual one, the point just stressed motivates its inclusion in the present section.
9
It is true that the expression “propositional content” is sometimes used to merely mean something like
truth-assessable content, and this may lead one to perceive our emphasis on concept-deployment as mis-
placed. Yet, the near consensus within the present literature regarding the conceptual nature of the content
of judgment and judgment-like attitudes warrants the more substantial characterization of propositional
content we work with in our presentation.
10
This way of developing the idea of conceptual content is that of Crane (1992). Bermudez (1998) offers an
alternative account, which is criticized in Creese and Deonna (2006).
much more demanding than the capacity to respond discriminatively to various stimuli,
a capacity which is widely, if not universally shared by animals.
Now, if these observations are met with approval when they concern colors, they
must a fortiori be accepted as regards the evaluative concepts appealed to by the theo-
ries under examination. The strategy consisting in modifying slightly the evaluative
judgment theory by allowing that a subject undergoing an emotion need not accept
the truth of the relevant proposition is then clearly unsuccessful. The fundamental
worry stems from the fact that emotions are conceived of as mental states dependent
on capacities similar to the ones we deploy when we commit ourselves to the truth
of a proposition. And this, together with its failure to accommodate the experiential
dimension of emotions, is more than enough to reject the evaluative judgment theory
and its variants. The considerations we brought to bear on this rejection also constitute
one of the chief reasons why it is nowadays widely accepted that, if emotions are ways
of apprehending values, this should be understood in completely different terms. This
is precisely the purpose of the so-called perceptual theory of emotions.
III. The Perceptual Model

The conclusion to which we have been led is that the evaluative judgment theory and
its variants fail to provide a plausible account of how emotions relate to evaluative
properties as well as for the fact that they are experiences with a salient phenomenol-
ogy. It has provided the incentive to pursue an alternative approach that consists in
trying to understand emotions on the model of perceptual experiences.11 In a nutshell,
the core idea is now that emotions must be understood as perception-like experiences
of values rather than as evaluative judgments or thoughts.
This is unquestionably an attractive move. Perceptual and emotional experiences
have both intentional as well as phenomenological aspects. If a theory of the emotions
must unify these two aspects in some convincing manner, it is then tempting to appeal
to perceptual experiences in order to account for the felt aspect of emotions. All the
more so since such a perceptual approach has the merit of getting around the two chief
difficulties concerning respectively the attitude and the content of emotions that we
have expanded on at some length while discussing the evaluative judgment theory. Let
us consider them in turn.
The attitude involved in perceiving has often been claimed to be distinct from that of
judging, and this for reasons identical to those adduced in the preceding section with
regard to the emotions.12 Our judgments about the properties of our surroundings are
11
This model, anticipated by de Sousa (1987), is explored in detail and espoused by Tappolet (2000, forth-
coming) and defended in different versions by Döring (2007, 2008), Prinz (2004), Deonna (2006), as well as
Deonna and Teroni (2008). As will become obvious, we have since then changed our minds.
12
An influential paper developing this claim is Crane (1992). Perhaps because many are convinced that
perceiving is not an attitude or a mode, positive characterizations of perceiving as an attitude or a mode are
scarce. For some insightful remarks on this topic, see Mulligan (1999) and Sturgeon (2000: ch. 1).
sometimes at odds with the deliverances of our senses (as the fabled straight stick half
immersed in water illustrates), and such situations are just as commonplace as those
in which our evaluative judgments and affective responses vis-à-vis a given object or
situation come apart. Michael is fully aware of the fact that this small bug is hardly
dangerous; yet, he shivers at its sight and swears that he is about to die. More generally,
there exists a variety of emotional illusions, amongst which there are phobias—illu-
sions of a particularly acute and recurrent nature. Regarding such cases, the perceptual
model has a clear advantage over the judgment theory: namely, that of acknowledging
the existence of such conflicts without thereby implying that the conflicted subjects
are affected by a serious form of irrationality.13 It can claim this theoretical advantage
because the sort of attitude that is characteristic of perception differs crucially from
that of judging.14 Of course, we have seen that some variants of the evaluative judgment
theory can lay claim to the same advantage, but the perceptual model has the addi-
tional merit of doing this without requiring that the subject possess complex cognitive
capacities. Let us now turn to this last consideration as it constitutes one of the key
promises of the perceptual model.
That model indeed turns out to be equally attractive when attention is turned to the
content of emotions: perceptual and emotional experiences arguably constitute two
examples of mental states endowed with non-conceptual content. By this, reference is
made to experiences that are contentful—that is, represent something—without this
depending on the subjects’ deployment (in the sense specified previously) of the con-
cepts used to specify these contents. The idea is thus to sever issues related to represen-
tation from issues related to concept mastery. And it can be profitably applied to the
kinds of experiences we are considering. Creatures with limited or no conceptual abili-
ties may perfectly well perceive a given shade of red, or feel a particular danger, without
thereby being able to meet the various requirements connected to judgments of color
or value. Although space does not permit us to examine in much detail the different
arguments in support of this idea, we shall at least briefly rehearse a few considerations
that have been put forward in its favor.
Observe first that perceptual experiences seem to be free from the epistemic and
normative constraints associated with conceptual thinking; more specifically, they
are not subject to norms. As opposed to passing a judgment about it, perceiving a
given property or object does not depend on the mastery of the kinds of epistemic,
semantic, and logical relations tying the concept in question to other concepts laid
13
This issue has been the focus of an interesting debate whose main protagonists are Brady (2007, 2009),
Döring (2008, this volume), Helm (2001), Roberts (2003), and Tappolet (2000, 2012).
14
This way of distinguishing the emotional attitude from the judgment it may give rise to makes room
for the possibility of developing within the perceptual approach an epistemology of values that assigns to
emotions a role similar to that of perceptions. The idea is that while visual or auditory perceptions can justify
judgments about visual or auditory properties, emotions play the same role as regards evaluative judgments.
We shall not address these important epistemic questions here. For a detailed treatment of the relevant
issues, see Deonna and Teroni (2012: chs. 8–10).
out in the foregoing. This contrast at the level of content explains why the fundamen-
tal norms bearing on judgments do not bear on perception. Thus, for instance, while
we are bound to accept the logical consequences of our judgments, there are no other
perceptual experiences which we should have as a result of perceiving, say, the redness
of a tomato. Indeed, the very notion of such a constraint barely makes sense. Second,
the fineness of grain which characterizes perceptual experiences darkens the pros-
pects of assimilating them to conceptual states. We never perceive properties such
as red, or crimson, but much more specific shades that we cannot conceptualize any
more than we can retain them in short-term memory. In a nutshell, the fineness of
grain of perceptual experiences exceeds our conceptual capacities. Finally, it is often
said that the acquisition of many of our concepts has its roots in perception, and this
could not be the case if perceiving any property required that one deploy the relevant
concept.15
According to the advocates of the perceptual model, the three types of consideration
just reviewed apply equally to the emotions: they are similarly free of the constraints
associated with the presence of concepts, they likewise respond discriminatively to
very subtle evaluative differences (that we could not easily articulate conceptually),
and they play a crucial role in the acquisition of evaluative concepts such as those of
danger, offense, disgrace, and so forth. Because the aforementioned cognitive con-
straints do not bear on experiences with non-conceptual content, a perceptual theory
has a real edge over its older cousin.16
Similarities between perceptual and emotional experiences can thus be found at the
level of attitude as well as at the level of content. So, why not endorse the substantial
claim according to which emotions literally are perceptions of values? As a matter of
fact, this claim becomes hard to maintain as soon as attention fixes on some key differ-
ences between emotions and perceptual experiences. We shall mention here only a few.
First of all, some will argue that perceptual experiences are underpinned by the activi-
ties of various sense organs, whereas this idea does not make much sense in the case of
emotions. Second, emotions lend themselves naturally to being classified according to
their polarity or valence (e.g. Colombetti 2005; Teroni 2011): some of them are negative
(fear, sadness, shame), others are positive (joy, pride, admiration). Yet, such a classifi-
cation makes no sense for perceptual experiences. Finally, while perception provides
us with information about our environment through mechanisms that are relatively
impervious to our motivational states,17 emotions, by contrast, are highly responsive
15
Of course, the debate between conceptualists and non-conceptualists has left philosophers of percep-
tion divided. For positions that reject the approach articulated here—that is, conceptualist positions—see,
for example, McDowell (1994) and Brewer (2002).
16
Tappolet (forthcoming) elaborates the claim that emotions have non-conceptual content, whereas
Helm (2001) emphasizes the existence of strong normative constraints bearing on affective states.
17
This claim presupposes that we carefully distinguish perceptual states from both attention and per-
ceptual judgments, as the latter are, of course, influenced by our motivational states. On this point see, in
particular, Dretske (2006).
to our desires, concerns, and character traits. Now, it is certainly possible to wonder
whether these differences really threaten the import of the analogy between emotion
and perception. They may, after all, not run very deep, and we should perhaps learn
to live with them in the light of the significant advantages delivered by the perceptual
approach. So let us turn our attention to two further differences that cannot be wel-
comed with such equanimity.
The first one regards the realm of things about which we can have emotions. This
realm is definitively not circumscribed in the way the realm of perceptual objects is. In
order for someone to be in a position to, for instance, see an object, several constraints
must be fulfilled: the object cannot be abstract, the subject must be in a given spatial
relation to it, certain lighting conditions must obtain, the field of view must be unob-
structed, some physical processes must take place, and so forth. No object can be seen
unless these conditions are met. Things are quite different with the emotions. If Sam is
afraid of the racoon right in front of him, then conditions identical to those holding for
perception must surely be met for his fear to be about the racoon. Yet although emo-
tions grounded in perceptual experiences are perhaps paradigmatic cases, many oth-
ers are about situations that are remembered, imagined, believed, supposed, and so on.
Furthermore, there can be emotions about abstract objects just as there can be emo-
tions about concrete ones—Maria may regret the lack of justice in her country, and
Michelle may hope that Darwinism is true. So, while the set of the possible objects of
emotions is entirely open, this fails to carry over to the case of perception. This puts the
perceptual model under substantial pressure. Indeed, one fundamental issue at this
stage is whether someone may properly be said to be in a perceptual relation with the
evaluative property of an object she happens to represent in a memory, a belief, a sup-
position, or an imaginative state. A positive answer to this is perhaps not inconceiv-
able,18 but definitively not very promising. More generally, the previous considerations
should make one wonder whether there is any plausibility in the claim that the relation
between emotions and values is of a perceptual nature. Many would object that values
are simply not the sort of properties one could possibly perceive. However that may be,
in order to account for the relation between emotions and values, the theory appears to
do no more than postulate the existence of a perceptual relation whose nature remains
ultimately unexplained. As we shall see, a significant advantage of the alternative the-
ory we are about to put forward consists precisely in its capacity to elucidate the rela-
tion in question.
The second difference between perceptions and emotions, which is closely related
to what we have just said, is perhaps even more significant. Perception constitutes an
autonomous way of accessing the properties and objects it represents: seeing, hearing,
or touching do not latch onto some prior apprehension of the properties or objects
It might indeed become conceivable if, following Pelser (this volume), one conceives of the content of
18
perception as being fully propositional in the substantial sense retained in the present chapter. As we have
argued, it is unclear to us how this constitutes an improvement on the judgment theory of emotions.
they give access to. In order for a subject to be in a position to see the redness of a
tomato, or hear the pitch of a given voice, no prior access to such properties needs to be
secured. Emotions work differently. They latch onto other psychological states, which,
as we have seen, may be of many distinct types (beliefs, memories, imaginative or per-
ceptual experiences), and which function as their cognitive bases. In order for one to
be afraid of a dog, one needs some representation of the dog that is logically prior
to the emotion itself: one needs to see, hear, or have beliefs about it. This difference
is of the utmost importance.19 Indeed, one of the main tasks with which a theory of
the emotions is confronted consists in offering a satisfactory explanation of the rela-
tion between emotions and their cognitive bases. To emphasize just one lesson among
several which may be drawn from the fact that emotions can have a great variety of
objects, it is now clear that we should challenge the assimilation of emotions to states
susceptible of having only one type of content—a conceptual proposition-like content
according to the evaluative judgment theory or a non-conceptual one according to the
perceptual theory. An alternative to these must be found.
The journey accomplished so far has led to the following conclusion. Although
fruitful in some respects, the analogy between emotion and perception, if pushed too
far, obscures some fundamental differences between the two types of mental states.
Two different strategies may now be envisioned. The first one consists in forsaking any
straight identification of the emotions with perceptions of values while pursuing the
kindred approach consisting in claiming that emotions, while not strictly perceptions,
are still experiences of values. We shall not discuss this strategy here, for we incline to
think that its various forms do not improve our understanding of the emotions.20 The
second, more radical strategy consists in preserving the idea that emotions relate to
values while rejecting a claim shared by the judgment theory and the perceptual the-
ory; namely, that values are represented by the emotions. The remainder of this chapter
is devoted to a brief presentation of this strategy.
IV. The Attitudinal Theory

We have seen that attempts to capture the specific relation that takes place between
emotions and values in terms of judgment-like or perception-like attitudes are likely
19
We are familiar with the idea that perception is cognitively penetrable; that is, that some features of
its content can depend on the fact that the subject possesses given conceptual abilities. This sense in which
perception may be said to have cognitive bases has nothing to do with the kind of dependence we are refer-
ring to here; that is with the fact that emotions inherit their contents, whatever these contents are, from other
mental states.
20
If reference to perception is simply meant to draw attention to just one or another feature shared by
perceptual and emotional experiences—be it a salient phenomenology, the presence of non-conceptual con-
tents, modularity or near modularity, appearance of truth, or the capacity to justify the relevant judgments
non-inferentially or anything else—then it would be better to do without such a misleading reference and
simply account for these features of emotions. For a detailed presentation and criticism of both the literal
and the more relaxed perceptual approaches to the emotions, see Deonna and Teroni (2012: ch. 6).
to fail. Now, the central source of the problems they confront is the fact that these
approaches proceed by redeploying attitudes that have nothing emotional about them
but that are alleged to somehow turn into emotions when they relate to specific evalu-
ative contents. Provided that these theories share the distinctive claim that emotions
put the subject in relation to values because they represent such properties, we have
an incentive to put this claim into question. That is to say that the difference between
emotions and other types of attitudes, as well as the difference between types of emo-
tions—at least insofar as their relation to evaluative properties is concerned—is not
to be located at the level of content, but at the level of the attitudes themselves. Although
sibylline at this stage, such an alternative approach has at least an intuitive ring about it.
Conceiving of the different types of emotions as many different attitudes rather than
one and the same attitude—that of judging, that of perceiving, or, for that matter, that
of emoting21—directed at the representation of different evaluative properties should
after all appear as plain common sense. Is it not obvious that the contrast between,
say, fear, anger, and joy is to be understood as one between different attitudes we take
towards objects, events, situations, and so forth? Is this contrast not to be located at the
same level as the contrast between, say, desiring, believing, and conjecturing—and is it
not at the same time to be carefully distinguished from the contrast between believing
a given proposition and believing a different one? We think that these two questions
should be answered affirmatively. To locate such a relation at the level of the attitude
has the additional virtue that it stands in basic agreement with the intuitive idea that
two distinct emotions can literally have the same content. After all, if Maria is amused
by a remark which irritates Michelle, it is natural to think that they are adopting differ-
ent attitudes regarding one and the same content.22 This cannot be maintained within
a perceptual model, however, within which to be amused by something consists in
perceiving its humorous character, whereas to be angry at it consists in perceiving its
offensiveness.
Still, as intuitive as such an attitudinal approach to the relation between emotions
and values may be, we should not come to a verdict before having turned our attention
to the following issues. What is the nature of emotional attitudes? In what sense are
they evaluations? To deal with these questions, we shall first briefly retrace our steps
to what we have observed, and shall on this basis articulate and defend the idea that
emotional attitudes are distinctive types of bodily attitudes. Second, we shall move on
to explain why this attitudinal theory is apt to illuminate the sense in which emotions
are evaluations.
21
Observe that positing the existence of a generic and sui generis attitude of emoting that would be shared
by all different types of emotion would, in the present context, constitute a clear step backwards, for it would
once again require that the differences among these types be made at the level of content. In the light of the
foregoing, this is not the strategy to pursue. The idea that emotions are attitudes in the sense of types of
position-taking is nicely emphasized in De Sousa (1987).
22
Goldie (2004) and Gunther (2004) object to the idea that in emotion a clear demarcation line between
attitude and content can be drawn in this way. The weaknesses of their respective criticisms are exposed
convincingly in Herzberg (2012) and in Deonna and Teroni (forthcoming).
Emotional attitudes, we have already observed, are attitudes that subjects can take
towards a variety of different contents. This is because they always inherit their content
from that of other mental states that function as their cognitive bases. One may, on the
basis of a conjecture to that effect, fear that the stock market will crash, but one may
also much more simply be afraid of a lion one sees approaching. Emotions are, in addi-
tion, attitudes that are characterized by a very salient phenomenology—it plainly feels
like something to be afraid, like another thing to be angry, and so on. To say this, how-
ever, naturally leads to the following substantial question: how are we to characterize
the phenomenology of emotional attitudes?
Although we cannot adequately defend that claim here, the answer must in our opin-
ion appeal to the way one’s body feels. Since William James, philosophers and psychol-
ogists alike have put emphasis on the vital importance of this aspect of the emotions.
We concur with them: neither introspection nor empirical investigation militates for
describing what is affective about episodes of fear, sorrow, anger, shame, joy, hope, and
so on, otherwise than in terms of the distinctive ways the body feels.23 Now, we are
aware that classical and more recent endeavors to understand emotions by appeal-
ing to their bodily phenomenology have often led to the conclusion, embraced by will
or by force, that emotions cannot acquaint us with anything but our own body. This
would obviously constitute a step backwards by comparison with the theories consid-
ered up till now, which at least make room for the plain fact that emotions are more
often than not directed towards something outside the subject’s own body. And the
issue is perhaps made especially salient within the sort of attitudinal approach we rec-
ommend: glossing an episode of fear or sadness by saying that it is an attitude directed
towards the subject’s own body indeed hardly makes sense. In any case, we shall now
see that one does not need to subscribe to such a thesis even if one insists on the bodily
nature of emotional phenomenology.
To see why this is the case, we can start with the following observation: the bodily
changes involved in emotions are felt by the subject as distinctive attitudes that are (at
least typically) directed towards external objects. In other words, while undergoing
an emotion, the subject feels herself taking a distinctive bodily attitude towards a cer-
tain object. More specifically, there is, in the words of an early and eloquent advocate
of this approach, “a consciousness of a form, a ‘Gestalt’ of multiple organic impres-
sions [ . . .] a consciousness of a global attitude of the organism” (Claparède 1928: 128).
In order to understand the nature of emotional attitudes, we then recommend tak-
ing exception to the somewhat atomistic approach to bodily feelings which is com-
mon in the recent literature, and appealing instead to a type of bodily awareness
which is both holistic and directed towards objects beyond the body. This in addition
23
The principal accounts of the emotions inspired by James found in the contemporary literature are
those of Damasio (2000) and Prinz (2004). These accounts, like ours, inherit potentially the many prob-
lems that have long been associated with James’ theory. For responses to these problems and to several oth-
ers that the presentation of our account will undoubtedly raise, see Deonna and Teroni (2012: ch. 7, and
forthcoming).
lends itself to an appealing development in terms of an aspect of emotions commonly

emphasized within contemporary psychology; namely, action readiness. The feeling
of an action readiness is in fact a key element in understanding how emotional atti-
tudes are evaluations, which is, you will recall, the second issue we have to confront.
The core idea is the following: we understand why emotions are evaluations once we
admit that they relate to values by virtue of being experiences of one’s body being ready
or poised to act in some specific manner towards a given object or situation. The notion
of bodily attitude we appeal to here is intended in quite a broad sense so as to include
the tendency to move away, to approach or move towards an object, to submit oneself
to it, to focus on it, to have one’s attention captured by it, to disengage one’s attention
from it, or even the tendency to prevent any contact with it. Here are some illustrations
of what we have in mind. Whenever one is afraid, one feels one’s body readiness to neu-
tralize something; whenever one is angry, one feels one’s body readiness to retaliate in
some way or another; episodes of shame are characterized by a felt tendency to disap-
pear from the sight of the ones who elicited this emotion; when sad, we feel our body
being deprived of the possibility of interacting with the object whose loss has elicited
our emotion. These illustrations are meant to emphasize the agential dimension of the
feel specific to emotions. And this is key to understanding the sense in which the latter
are evaluative attitudes.
This is best appreciated if we step back a little and remind ourselves of some general
features of the way the correctness conditions of mental states are fixed. These cor-
rectness conditions are the joint upshot of two distinct factors: the content and the
attitude. For instance, the proposition that John speaks Swahili may, amongst other
attitudes, be the object of a belief or of a supposition. Its truth or mere possibility enters
the conditions of correctness of the resulting mental states by virtue of the fact that it is
either believed or supposed. Correctness conditions result, then, from the conjunction
of a content, identical in both cases, and distinct attitudes which contribute a property
that the proposition featuring in the content must possess—truth in one case, possibil-
ity in the other one—for the state to be correct. This is the sense in which belief and
supposition relate to truth and possibility, and this has nothing to do with what they
represent. We believe that types of emotions relate to types of values, which are their
formal objects, in exactly the same way—fear to dangerousness, anger to offensive-
ness, and so on. The sense in which emotions are evaluative attitudes, then, is this: the
fact that they are the very attitudes they are explains why adopting them towards an
object is correct if, and only if, that object actually exemplifies the relevant evaluative
property.
Thinking of these attitudes in terms of felt action readiness provides an elucidation
of why they can play this explanatory role. Fear relates to danger insofar as it is an expe-
rience of one’s body’s readiness to reduce an object’s impact (flight, preventive attack,
immobility, and so on)— an attitude which is correct to adopt only if the object is actu-
ally dangerous. Analogously, anger relates to offensiveness insofar as it is an experi-
ence of one’s body readiness to retaliate in some way or another, an attitude which is
correct to adopt only when confronting an offense. And shame relates to one’s own
degradation insofar as one feels one’s body’s readiness to disappear into the ground,
or to move away from the sight of relevant others—an attitude which is only correct if
we are actually degraded. These examples should suffice to illustrate the sense in which
emotions are evaluative attitudes. Its fruitfulness depends of course on the possibil-
ity of providing rich and convincing descriptions of a substantial amount of emotion
types in terms of felt bodily attitudes. The hope that this approach may successfully be
pursued does not seem unwarranted to us.24
The attitudinal theory we have outlined here avoids the difficulties with which both
the judgment and the perceptual theories find themselves confronted. Like the per-
ceptual theory, the attitudinal theory maintains that the phenomenology of emotions
is essential to the way in which they relate to values. However, the two theories part
company to the extent that the latter understands the relation to values in terms of
distinct types of emotional attitudes. In other words, it denies that values need to be
represented by emotions in order to feature into their correctness conditions. This is
why the theory avoids many undesirable consequences we have had the opportunity
to discuss. Within the specific brand of attitudinal theory defended here, the distinct
evaluative correctness conditions of different emotion types trace back to distinctive
types of felt bodily attitudes. As is the case for other attitudes, the formal objects of the
emotions do not need to feature in their content. For that reason, we do not have to end
up claiming that emotions presuppose the mastery of complex evaluative concepts, or
that they are subtended by mysterious quasi-perceptual relations to values: the sense in
which they relate to the latter traces back to the nature of the relevant attitudes which
have nothing especially demanding or mysterious about them.25
References
Bermudez, J. L. (1998). The Paradox of Self-Consciousness (Cambridge, MA: MIT Press).
Brady, M. S. (2007). “Recalcitrant Emotions and Visual Illusions.” American Philosophical
Quarterly 44(3): 273–84.
Brady, M. S. (2009). “The Irrationality of Recalcitrant Emotions.” Philosophical Studies
145(3): 413–30.
Brewer, B. (2002). Perception and Reason (Oxford: Clarendon Press).
Claparède, E. (1928). “Feelings and Emotions.” In Feelings and Emotions: The Wittenberg
Symposium, ed. M. L. Reymert (Worcester: Clark University Press), 124–39.
Colombetti, G. (2005). “Appraising Valence.” Journal of Consciousness Studies 12(8–10): 103–26.
24
The number of different profiles of systematic physiological changes susceptible to being distinctively
felt is still an open empirical question, yet there are reasons to be optimistic (see e.g. Scherer 2009).
25
We are grateful to Sabine Roeser, Cain Todd, and Michele Ombrato for their comments on a previ-
ous version of this chapter. We acknowledge the support of the SNSF-sponsored Swiss National Centre for
Competence in Research (NCCR) in Affective Sciences.
Crane, T. (1992). “The Nonconceptual Content of Experience.” In The Contents of Experience:

Essays on Perception, ed. T. Crane (Cambridge: Cambridge University Press), 136–57.
Creese, A. and Deonna, J. (2006). “Les Liaisons Dangereuses or How Not to Construe
Nonconceptual Content.” European Review of Philosophy 6: 101–15.
Damasio, A. (2000). The Feeling of What Happens: Body and Emotion in the Making of
Consciousness (New York: Harcourt Brace).
D’Arms, J. and Jacobson, D. (2000). “The Moralistic Fallacy: On the ‘Appropriateness’ of
Emotions.” Philosophy and Phenomenological Research 61(1): 65–90.
de Sousa, R. (1987). The Rationality of Emotions (Cambridge, MA: MIT Press).
Deigh, J. (1994). “Cognitivism in the Theory of Emotion.” Ethics 104(4): 824–54.
Deonna, J. A. (2006). “Emotion, Perception, and Perspective.” Dialectica 60(1): 29–46.
Deonna, J. A. and Teroni, F. (2008). Qu’est-ce qu’une émotion (Paris: Vrin).
Deonna, J. A. and Teroni, F. (2012). The Emotions: A Philosophical Introduction (London and
New York: Routledge).
Deonna, J. A. and Teroni, F. (forthcoming). “Emotions as Attitudes.” Dialectica.
Döring, S. (2007). “Seeing What to Do: Affective Perception and Rational Motivation.” Dialectica
61(3): 363–94.
Döring, S. (2008). “Conflict Without Contradiction.” In Epistemology and Emotions, ed. G.
Brun, U. Doguoglu, and D. Kuenzle (Aldershot: Ashgate), 83–104.
Dretske, F. (2006). “Perception Without Awareness.” In Perceptual Experience, ed. T. S. Gendler
and J. Hawthorne (Oxford: Clarendon Press), 147–80.
Goldie, P. (2004). “Emotion, Feeling, and Knowledge of the World.” In Thinking About
Feeling: Contemporary Philosophers on Emotions, ed. R. Solomon (New York: Oxford
University Press), 91–106.
Greenspan, P. S. (1988). Emotions and Reason (New York: Routledge).
Gunther, Y. H. (2004). “Emotion and Force.” In Essays on Nonconceptual Content, ed. Y. H.
Gunter (Cambridge, MA: MIT Press), 279–88.
Helm, B. (2001). Emotional Reason: Deliberation, Motivation, and the Nature of Value
(New York: Cambridge University Press).
Herzberg, L. A. (2012). “To Blend or to Compose: A Debate about Emotion Structure.” In
Dynamicity in Emotion Concepts, ed. P. Wilson (Frankfurt: Peter Lang), 73–94.
Kenny, A. (1963). Action, Emotion and Will (London: Routledge).
McDowell, J. (1994). Mind and World (Cambridge, MA: Harvard University Press).
Mulligan, K. (1998). “From Appropriate Emotions to Values.” The Monist 81(1): 161–88.
Mulligan, K. (1999). “Perception, Particulars and Predicates.” In Consciousness and
Intentionality: Models and Modalities of Attribution, ed. D. Fisette. The Western Ontario
Series in Philosophy of Science, Vol. 62 (Dordrecht: Kluwer), 163–94.
Nussbaum, M. (1994). The Therapy of Desire: Theory and Practice in Hellenistic Ethics (Princeton,
NJ: Princeton University Press).
Prinz, J. (2004). Gut Reactions: A Perceptual Theory of Emotion (New York: Oxford University
Press).
Roberts, R. (1988). “What an Emotion Is: A Sketch.” Philosophical Review 97(2): 183–209.
Roberts, R. (2003). Emotion: An Essay in Aid of Moral Psychology (New York: Cambridge
University Press).
Scherer, K. R. (2009). “The Dynamic Architecture of Emotion: Evidence for the Component
Model Process.” Cognition and Emotion 23(7): 1307–51.
Solomon, R. (1988). “On Emotions as Judgments.” American Philosophical Quarterly
25(2): 183–91.
Sturgeon, S. (2000). Matters of Mind: Consciousness, Reason and Nature (London: Routledge).
Tappolet, C. (2000). Emotions et valeurs (Paris: Presses Universitaires de France).
Tappolet, C. (2012). “Emotions, Perceptions, and Emotional Illusions.” In Perceptual Illusions:
Philosophical and Psychological Essays, ed. C. Calabi (Basingstoke: Palgrave Macmillan),
205‒22.
Tappolet, C. (forthcoming). Emotions, Values, and Agency (New York: Oxford University Press).
Teroni, F. (2007). “Emotions and Formal Objects.” Dialectica 61(3): 395–415.
Teroni, F. (2011). “Plus ou moins: émotions et valence.” In Les ombres de l’âme: penser les émo-
tions négatives, ed. C. Tappolet, F. Teroni, and A. Konzelmann Ziv (Geneva: Markus Haller
Editions), 21–36.
3
Evaluative Phenomenology
Michelle Montague
1. Introduction
Our normal waking life involves a seemingly continuous stream of conscious-
ness. What is given in this stream? Almost everyone agrees that a world of physi-
cal objects and states of affairs involving those objects is given,1 and evaluations of
objects and states of affairs are also given—we judge, think, and feel that things are
good or bad. Even if evaluative properties are not in fact objective features of the
world, David Hume and J. L. Mackie are surely right in holding that they seem to us
to be objective features of the world, and in that sense at least they are given to us in
experience.
This givenness of objects and states of affairs in consciousness (including evalua-
tions of objects and states of affairs) essentially involves phenomenological givenness.
It is only in or through phenomenological or experiential givenness that the world of
objects and states of affairs (and evaluations of those objects and states of affairs) can
be given to us in consciousness. All conscious episodes must essentially involve, and
are conscious in virtue of having, a certain phenomenological–experiential–qualita-
tive character—a character which is such that there is, in a familiar phrase, ‘something
it is like’, experientially, for the subject of experience to experience it. As I will under-
stand the term ‘consciousness’, there is no non-phenomenological consciousness.
Notions such as Ned Block’s ‘access consciousness’ may pick out certain functional
properties that certain mental episodes have (and some of the episodes they pick out
may qualify as conscious mental episodes), but they do not pick out a special kind
of consciousness—a kind of consciousness that can possibly exist without having any
phenomenological character.2
The conscious stream involves a great deal of activity: sensing the ambient environ-
ment through the five external sensory modalities, feeling worried, happy, and sad
1
There are other kinds of objects given to consciousness, such as mathematical objects, aesthetic objects,
fictional objects, but these will not concern me here.
2
I argue that all consciousness must be phenomenological in ‘The Life of the Mind’, forthcoming.
Evaluative Phenomenology 33
about various situations in our lives, and having a great many thoughts, such as judg-
ing that the economy is getting better, wondering what will happen in the Middle East,
entertaining the possibility of vacation in Greece next year, and so on. This is not, of
course, an exhaustive list of what constitutes the stream, and there are many questions
about how we can best categorize these various conscious activities. One rough group-
ing divides them into three categories: thought, perception (including propriocep-
tion), and emotion/mood. These are not strictly mutually exclusive categories, because
the conscious activities that constitute the stream are so overlapping and interdepend-
ent, but I think that we can usefully talk about them separately, and the purpose of this
chapter is to focus on the category of conscious emotion.
I will restrict my attention to emotions that are not only conscious experiences but
are also intentional phenomena, leaving open the possibility that there may be con-
scious emotions, such as free floating anxiety or depression, that are not intentional
phenomena at all, and acknowledging the fact that we allow that attributions of emo-
tional states or conditions may be true even if the person to whom they are attributed
is not at that time consciously feeling any emotion at all, being, perhaps, in a dreamless
sleep.
It is clear that conscious emotions are complex intentional phenomena. That is, they
not only represent objects and states of affairs, but they essentially represent objects
and states of affairs in an evaluative way. I will argue that emotions represent objects
and states of affairs as being sad, just, or joyful, and that these representations are
essentially evaluative representations. That is, emotions essentially represent objects
and states of affairs as having evaluative properties. (Sometimes I will put this point by
saying that emotions represent evaluative properties. I will also use the shorter expres-
sion ‘their objects’ and ‘its object’ as short for an emotion’s representing an ‘object or
a state of affairs’.) Many philosophers claim that emotions are somehow connected
to evaluations of objects and states of affairs,3 but I am claiming something stronger.
Emotions, according to the present view, are essentially evaluative representations.4
It follows from my definition of consciousness that conscious emotions are essen-
tially phenomenological episodes, and of course most philosophers accept this. But
most views that accept that emotions essentially involve phenomenology assimilate that
phenomenology to some type of sensory or bodily phenomenology.5 According to the
current proposal, by contrast, emotions have their own distinctive sui generis kind of
phenomenology, which I call ‘evaluative phenomenology’. In section 2 I will provide a
3
See e.g. de Sousa (1987), Greenspan (1988), Roberts (2003), and Solomon (1976).
4
Note that in saying that emotions represent objects and states of affairs as having evaluative properties,
I am not committing myself to any form of moral realism. Just as an error theorist about colour may claim
that we can represent colour properties without there being colour properties (as we experience them) in the
world, I am claiming that we can represent evaluative properties without there being any such properties in
the world. It may be that this position is inconsistent with certain contemporary forms of ‘externalism’ about
content. So be it.
5
See e.g. Prinz (2004).
34 Michelle Montague
rough taxonomy of kinds of phenomenology, distinguishing evaluative phenomenol-

ogy from other kinds, most notably from sensory phenomenology and what is now
often called ‘cognitive phenomenology’.6
What further distinguishes this view of emotions from other views is that it ties an
emotion’s evaluative phenomenology inextricably to its intentionality, and in par-
ticular to how it represents the evaluative properties of its object. Solomon (1976)
and Nussbaum (1990, 1994) both argue that emotions are evaluative judgements, but
according to their views, whatever phenomenology is involved in having such evalu-
ative judgements, that phenomenology is not part of how that judgement represents
evaluative properties. In contrast, I will argue that an emotion represents its object’s
evaluative properties in virtue of its own distinctive sui generis kind of phenomenol-
ogy. It is because emotions have evaluative phenomenology that I say that they are
experiences of value rather than just representations of value.
Much of this chapter will be dedicated to explicating how evaluative phenome-
nology represents evaluative properties. A central aspect of giving this account will
involve describing in some detail what I take the general structure of phenomenology
(or consciousness) to be. In describing this structure, it will often be helpful to com-
pare conscious perception and conscious emotion.
The final issue I will consider is the question of how ‘fine-grained’ evaluative phe-
nomenology is. It is clear that sensory phenomenology is very fine-grained. That is,
sensory phenomenology is rich and varied along different dimensions. For instance,
we have colour phenomenology, shape phenomenology, and sound phenomenology,
and each one of these categories is distinguishable into further detailed phenomenol-
ogy. I will suggest that evaluative phenomenology is also fine-grained along various
dimensions.
The view of emotions offered here requires a lot of preparation whose point may not
be apparent at first; I hope it will become so. Part of the difficulty is that the issues with
which I am concerned within the theory of emotions can only be adequately discussed
by looking at wider issues in the philosophy of mind. I began this introduction with the
idea that everything that is given in consciousness requires phenomenological given-
ness—a very general claim in the philosophy of mind. This chapter is concerned with
an application of this general claim to the particular case of the emotions: in the having
of an emotion, the (either really or seemingly objectively possessed) evaluative proper-
ties of the object and states of affairs in question are given to the subject of experience
in a distinctive way. That is, the way in which evaluative properties are given in emo-
tions is very different from the way they are given in evaluative judgements, which may
lack any emotional feeling.
In setting out an account of what is given in the having of an emotion, it will be help-
ful to consider what it is for something to be given to consciousness at all. I explicate
6
This may not be an exhaustive taxonomy. For example, some have argued that there is ‘agentive phe-
nomenology’, which may be irreducible to sensory or cognitive phenomenology. See e.g. Bayne (2011).
this idea of givenness in section 2 in terms of a general theory of content which takes
proper account of the phenomenological aspect of experience, and in particular offers
a taxonomy of phenomenology, as mentioned previously. A central feature of the the-
ory is the idea that in having a conscious experience the subject is always and neces-
sarily aware of the having of that very experience. I explicate this feature in section
3, because it is essential for giving an account of the property attributions made in
experience. One of the preparatory detours made in section 4 is to give an account of
how property attributions are made in perception, and in particular how colour attri-
butions are made in visual experience. This will help clarify how evaluative property
attributions are made in the having of emotional experience. Finally, in section 5 I dis-
cuss the fine-grainedness of evaluative phenomenology.
2. Theory of Content

I call the totality of what is given to one experientially in a conscious episode ‘the con-
tent’ of that episode. The content, the total mental content, of a conscious episode is
(absolutely) everything that is given to one, experientially, in the having of the experi-
ence. It is everything one is aware of, experientially, in the having of the experience. So I
have a very inclusive notion of content—one that extends beyond the use of ‘content’ to
simply mean ‘representational content’'—that is, truth conditions or accuracy condi-
tions—that is currently common in analytic philosophy.
So, what is given to a subject in having a conscious emotion? The total mental content
of a conscious emotion (or any conscious episode) is a genus under which several spe-
cies of content fall, and those species can be categorized under three headings:
(i) Phenomenological content, or experiential-qualitative character;
Since I take it that all consciousness is a phenomenological phenomenon, it follows
that all conscious emotion is a phenomenological phenomenon. And according to the
inclusive conception of content, given a particular emotion, the subject will be aware
of the total phenomenological character of the emotion; the total phenomenological
character of the emotion is given to the subject, and thus the total phenomenological
character of the emotion will be part of the content of the experience. I take it that if
it is part of phenomenological character, it is necessarily—by definition—given to the
subject.
(ii) Representational (or intentional) content,7 the phenomenon of something’s
being about something or of something; a representation of things being a cer-
tain way (essentially for emotions, and sometimes for thoughts, this will include
evaluative representations, representations of objects, and states of affairs as
having evaluative properties).
7
I will use ‘representational’ and ‘intentional’ interchangeably.
And:
(iii) External content, typically physical objects and their properties.8 (The question
of what the external content of emotion is will partly depend on whether evalu-
ative properties are objective properties of objects and states of affairs.)
In discussing the notion of content many philosophers have taken it that it is true
by definition that ‘all content is intentional content’. According to the present view, we
can speak of ‘sensory content’, or more widely of ‘phenomenological content’, with-
out commitment to the claim that all content is intentional content. In the end, I will
argue that all phenomenological content is intentional content, but that is a substan-
tive claim, and should not be made true by definition.
Once these three kinds of content are introduced, important issues arise about the
relationship between them. I will argue that all phenomenological content is represen-
tational or intentional content, as already remarked. That is, all phenomenology, has,
essentially, intentional ‘structure’. (My reasons for asserting this are radically differ-
ent from the standard representationalist’s reasons; a point I will return to shortly.)9 It
is a further question whether all representational content is phenomenological con-
tent. Many have thought that the answer to this question is ‘no’, given the evidence for
unconscious representational content. However, I will not address this issue in this
chapter.
Before considering the relation between phenomenological content and repre-
sentational content, I want to introduce three sui generis kinds of phenomenology or
phenomenological content, which I will call ‘sensory phenomenology’, ‘cognitive phe-
nomenology’, and ‘evaluative phenomenology’. The most familiar is sensory phenom-
enology, the ‘what it is likeness’ associated with the familiar five sensory modalities; for
example, what it is like to see colours, hear sounds, smell odours. Sensory phenomenol-
ogy also includes the phenomenology typically associated with proprioception. What
is now typically called ‘cognitive phenomenology’ is a non-sensory kind of phenom-
enology paradigmatically associated with conscious occurrent thought, but also with
perception and emotion.10 Advocates of cognitive phenomenology point out that there
is something it is like to consciously think that 2 + 2 = 4, or to think that temperance is
a virtue, that is irreducible to any sensory phenomenology that may be associated with
8
In using ‘external content’ to mean, for example, physical objects our perceptions are of, I am deviating
from the standard usage of this term in contemporary philosophy of mind.
9
In the case of conscious perception, ‘representationalists’ such as Dretske (1995), Harman (1990), and
Tye (1995, 2000, 2002, 2009) argue that all phenomenological content either supervenes or is identical to
intentional content, and intentional content is further reduced to some kind of causal-cum-functional rela-
tions between states of the brain and states of the environment.
10
There is some debate about the best way to define ‘cognitive phenomenology’. Some (e.g. Smithies 2013)
define it as whatever the phenomenology is that is associated with conscious thought even if it turns out
that that phenomenology is purely sensory. I think this obscures the central question of the cognitive phe-
nomenology debate and robs the term ‘cognitive phenomenology’ of its best terminological usage; that is, to
designate a type of phenomenology that is irreducible to any sensory phenomenology.
these thoughts. Finally, ‘evaluative phenomenology’ is—so I propose—a further sui

generis (non-sensory) kind of phenomenology that is associated with conscious emo-
tions. As mentioned previously, I choose the term ‘evaluative phenomenology’ because
I believe, and am going to argue, that emotional experiences are essentially experiences
of value or as of value.
I do not have the space here to give a full account of these different kinds of phenom-
enology and the ways in which they can be present in conscious thought, conscious
perception, and conscious emotion. (There is currently a lively debate about whether
there is such a thing as cognitive phenomenology at all.)11 I will just briefly mention
a few relevant theses to make explicit the differences between these three kinds of
phenomenology.
First, consider conscious thought. According to the view of consciousness proposed
here, all consciousness is essentially a phenomenological phenomenon. It follows that
conscious thought is an essentially phenomenological phenomenon. What kind of
phenomenology is associated with conscious thought? My view is that although cog-
nitive phenomenology and sensory phenomenology may be present in every case of
conscious thought, the presence of cognitive phenomenology is necessary to make
a conscious thought specifically a conscious thought, rather than some other kind of
conscious state, and is in this way essential to conscious thought.12
Secondly, consider conscious perception (where perception is opposed to, is some-
thing essentially more than, sensation). In this case I believe that both cognitive phe-
nomenology and sensory phenomenology are fundamentally present—even if the
kind of cognitive phenomenology present is sometimes very primitive.13
Thirdly, emotion. I take it that typical emotions involve all three kinds of phenome-
nology, although once again the cognitive phenomenology present may be very prim-
itive. I should also emphasize that although bodily sensations may often be present
when we experience emotions, evaluative phenomenology is not reducible to bodily
sensations. My emotional experience of sadness may, for example, include certain
‘fatigue’-like bodily sensations, but the psychic deflation associated with sadness is
not reducible to such bodily fatigue nor any other combination of bodily sensations.
I will argue that evaluative phenomenology is intentional in a way that bodily sen-
sations like pain, fatigue, and the feeling of a rapid heartbeat are not. Briefly, evalu-
ative phenomenology can be and in most cases is intentionally directed beyond the
body, whereas bodily sensations, if they are intentional at all, are only directed at the
11
For a recent collection of essays on this issue see Bayne and Montague (2011). See also Horgan and
Tienson (2002).
12
I argue for this claim in ‘The Life of the Mind’, forthcoming.
13
In Montague (2011) I argue that our perceiving and thinking is structured in such a way that it is fun-
damentally ‘object-positing’. Object-positing is best categorized as an instance of cognitive phenomenology
and delivers the this object of our perceptual experiences, although it is less conceptually specific, for exam-
ple, than seeing a dog as a dog.
body.14 It may also be true that one may make evaluative conscious judgements with-
out any emotional feeling, and because all conscious episodes are phenomenological
and experiential, an evaluative conscious judgement may be said to be an experiential
representation of value. However, conscious emotion, with its evaluative phenomenol-
ogy, is a distinctive kind of experience of value, fundamentally experientially different
from the kind of cognitive experience involved in an evaluative conscious judgement.
This difference should be evident when we reflect on the phenomenological difference
between judging (without feeling) that a friend’s death is sad and feeling down about
the sadness of a friend’s death. I will argue that this phenomenological difference entails
that the way evaluative phenomenology represents evaluative properties is very differ-
ent from the way a conscious thought or judgement represents evaluative properties.
3. Awareness-of-Awareness
There is one more essential element to the overall content of conscious episodes—an
element which ties together the intentional and phenomenological aspects of conscious
episodes in a number of ways. The basic idea is simple and ancient—and Aristotelian. It
is that in having a visual experience of a tree in leaf (for example), the subject, in addi-
tion to being aware of the tree and any other relevant external content, is also aware of
the awareness of the tree. In having a particular conscious perceptual experience the
subject is always and necessarily also aware of that very experience itself. There is always
some sort of awareness of the experience or experiencing: conscious awareness always
involves—constitutively involves—some sort of awareness of that very awareness.15
The relationship between the awareness-of-awareness feature of experience and the
phenomenological features of experience makes explicit how the phenomenological
features of experience are intentionally structured. The phenomenological features of
an experience are that in virtue of which an experience is what it is, experientially, to the
subject who has it, with the particular qualitative character that it has. The instantiation
of a phenomenological property immediately reveals to one that one is having an experi-
ence, and so in having an experience one is immediately aware of having an experience.
Two important facts follow from this picture. First, the awareness-of-awareness fea-
ture is an essential (constitutive) feature of phenomenology; given which it follows
immediately that phenomenology is fundamentally intentionally structured. Second,
according to the awareness-of-awareness thesis, phenomenological properties them-
selves present, phenomenologically, as what they are—properties of experience. This
14
Evaluative phenomenology can be directed at the body, as is the case when one fears the cancer in their
body.
15
This view is well expounded in the Phenomenological tradition. See e.g. Brentano (1995), Husserl
(2001), Gurwitsch (1966 [1941]), and so on. See also Kriegel (2009), Rosenthal (2005), Smith (1989), and
Zahavi (2006). Locke puts it strongly by saying that ‘thinking consists in being conscious that one thinks’
(Essay Concerning Human Understanding [1689]: II.i.19); he uses ‘think’ in the wide Cartesian sense to cover
all conscious mental goings-on).
second fact shows that the oft-cited ‘transparency thesis’ in the philosophy of percep-
tion, that introspection of our perceptual experiences never reveals properties of expe-
rience themselves, is simply false. Experiences immediately reveal themselves to us as
what they are, experiences, but we are often not thinking of them as such.16
To illustrate these points in more detail, let us begin again with visual experience. Most
agree that my visual experience of a tree in leaf is about or of the tree and its shape and
colour (that it has all these things as part of its content), but many deny that it is also about
or of its own visual phenomenological character. According to the view under discus-
sion, however, there is a fundamental sense in which the visual (as opposed to say tactile)
quality of one’s seeing the tree is part of what one is aware of in having the experience.17
The greenness and shape of the tree is visually presented to me. And its being pre-
sented to me visually is part of the intentional content of the experience, on this view,
because its being a visual experience is part of what I am aware of just in having it, and
necessarily so—otherwise it would not be a visual experience at all. In other terms: in
having a visual experience, I have an experience of the experience. Most simply, I expe-
rience the experience. This may seem trivial, but it is not!
This awareness does not require the possession of the concept visual experience,
or being visual, or any colour or shape concepts. Rather, it follows immediately, from
the fact that one’s awareness of the world, in the case in question, consists partly in
one’s being visually aware of the world, that one is, in having that experience, aware
of the visual character of one’s experience. And the content of this awareness of one’s
experience’s being visual18 can be specified only in terms of (by reference to) the sen-
sory phenomenology associated with visual experiences; that is, what it is like to see
colours, shapes, and so on. So the phenomenological content essential to having an
experience is part of an experience’s intentional content, whatever else may be part of
its intentional content.
We can summarize these claims as follows:
(a) In general, in having an experience, one is aware of the experience, one experi-
ences the experience, not (as it were) just the tree (as many philosophers seem to
hold these days).19
16
There is a lot of debate about the ‘transparency thesis’, and some have argued (e.g. Kriegel 2009) that the
awareness-of-awareness thesis is compatible with it. I think there are reasons the transparency thesis seems
plausible. The sense in which phenomenological properties reveal themselves to be properties of experience
in the having of the experience is already quite elusive, but it becomes even more so when we introspect on
our experiences. When I introspect on a visual experience, for example, I am no longer having that visual
experience and so the experiential properties of that visual experience become very difficult to access. My
best access to experiential properties is when I am actually having experiences.
17
Aristotle puts the point by saying that ‘we perceive that we perceive’, but the former kind of perceiving
is obviously different from the latter: the claim is not that we see that we see. It is controversial what the first
occurrence of perception amounts to, and the present claim is simply that it will involve appeal to phenom-
enological properties, whatever its full account. See Caston (2002) for an excellent discussion.
18
Arguably there is more to an experience’s being visual than sensory phenomenology, such as cognitive
phenomenology. However, I will leave this issue aside for the purposes of this chapter.
19
The sense in which one is aware of an experience as an experience should not be understood as involving
the concept experience. I mean the phrase ‘as an experience’ in the same sense as Tye means experiencing a
So:
(b) In having a visual experience, one experiences it as a visual experience—one is
aware of it as a visual experience, one experiences the distinctively visual char-
acter, the visuality, of the experience, which is quite unlike aural character, the
aurality, of hearing, or the olfactory character of smelling and so on.
So:
(c) Experiencing the visuality of the visual experience necessarily involves expe-
riencing the phenomenology that is proprietary to visual experience, such as
colour phenomenology, visual-shape phenomenology, and so on.
We can now say that each sensory modality is a ‘way’ of experiencing the world
(leaving aside questions of cross-modal effects). We visually experience the world, we
auditorily experience the world, we experience the world in an olfactory way, we expe-
rience the world in a gustatory way and so on.
And now we can add that the emotions, too, have their own unique way of present-
ing the world. In emotional experience, the world is presented in an evaluative way;
that is, in having emotions one is evaluatively aware of the world. Emotions phenom-
enologically present the world as possessing evaluative properties, and this evaluative
phenomenological presentation is distinctive in that the ‘evaluative phenomenology’
associated with evaluative presentations is irreducible to any sensory or cognitive
phenomenology that may be associated with emotion. So far, then, I am claiming that
emotions form a general class, like the sensory modalities, and they have in common,
as a general class, the property of presenting the world in an experiential evaluative
way. In section 5 I will consider how just as the sensory divides into more specific ‘sen-
sory ways’ of experiencing the world, to hear, see, taste, and so on, emotion divides into
more specific ‘evaluative ways’ of experiencing the world, to be sad, happy, angry, and
so on.
We can also now make the awareness-of-awareness thesis explicit for emotional
experience by providing the counterpart claims to (b) and (c) for a particular emo-
tional experience as follows:
(b) * In having an experience of sadness, a negative evaluative experience, for exam-
ple, one is aware of, one experiences, the sui generis negative evaluative emo-
tional phenomenological character of the experience of sadness.
So:
(c) * One is accordingly aware of the evaluative phenomenology proprietary to the
experience of sadness, such as psychic deflation or heaviness.
property as a property of an object in the following passage: ‘I found myself transfixed by the intense blue of
the Pacific Ocean . . . what I was focusing on, as it were, were a certain shade and intensity of the colour blue.
I experienced blue as a property of the ocean not as a property of my experience’ (Tye 2002: 448).
Again, I should stress that in saying that a subject is aware of an experience of sad-
ness as an experience of sadness I do not mean to imply that the subject must possess
the concept sadness.
Up until this point, I have focused on presenting a general theory of content, which
captures the different kinds of content that may be present in conscious experience, and
I have begun to elucidate some of the ways in which these different kinds of content are
related. In particular, I have tried to bring out one of the ways in which the intentional-
ity and the phenomenology of a conscious experience are related by arguing that all
conscious awareness constitutively involves an awareness of that very awareness, and
is therefore correctly said to be intentional with respect to itself (this being a condition
of its very possibility). In having an experience the subject of experience is necessarily
aware of the experience itself, and the subject immediately becomes aware of the expe-
rience itself, and this ‘awareness of ’ is to be understood as an intentional phenomenon.
So far, I have mainly described how awareness-of-awareness features in experience,
and now one might ask what reasons we have for accepting the awareness-of-aware-
ness thesis. Part of my argumentative strategy is already employed in the phenom-
enological description of our conscious experience. The idea is that this description
itself would make the awareness-of-awareness thesis seem intuitively plausible. The
hope is that this description captures something essential about perceptual and emo-
tional experience, and in so doing provides a reason for accepting the awareness-
of-awareness claim. In the next section, a more theoretical reason for accepting the
awareness-of-awareness thesis is given. I will argue that the thesis plays an essential
role in accounting for how property attributions are made in typical perceptual and
emotional experiences.20
4. Evaluative Phenomenology and Evaluative

Property Attributions
The first point to make is that phenomenological properties do double duty: in occur-
ring (being instantiated), they not only reveal themselves to the subject as what they
are, experiential properties; they also purport to reveal features of the world. In this
section I focus on the latter feature of phenomenological properties. Insofar as the
phenomenological properties of emotional experiences purport to reveal features of
the world, I am concerned mainly with two questions: in typical emotional experi-
ences, what kinds of properties are attributed to objects and states of affairs, and how
are those properties attributed?
In standard cases of sensory perception, physical objects are presented to a subject
as being coloured, shaped, soft, sweet, loud, and so on. In emotion, objects and states
20
For this second argument for the awareness-of-awareness thesis to be complete, I would have to argue
against other views, which I cannot do here.
of affairs are presented to a subject as being joyful, sad, horrifying, uplifting, annoying,
and so on.21
In what way are states of affairs given as sad, horrifying, uplifting, annoying, tragic?
I will argue for three claims. First, sadness, the joyful, the horrifying, and so on are
at least sometimes properties we attribute to objects and states of affairs, and these
property attributions are experienced as attributions of objective properties of states
of affairs. Second, these property attributions are essentially evaluative property attri-
butions. Third, it is partly in virtue of the evaluative phenomenology associated with
emotions that evaluative properties are attributed to objects and states of affairs.22
To begin to defend these three claims let us return to the example of being sad about
a friend’s death. It seems totally natural for a subject to attribute the property of sad-
ness to the state of affairs of her friend’s death: ‘my friend’s death is sad’. This kind of
attribution of ‘emotion properties’ is fairly common in our language: we feel the behav-
iour of irresponsible stockbrokers contemptible; we find our boss annoying; we feel
that the conflict in Syria is tragic; we feel the war in Bosnia was disgusting, and so on.
These various property attributions, such as sadness, the tragic, the joyful, are experi-
enced as the attribution of objective properties of states of affairs. They are properties
these states of affairs seem to have, independently of us. It may be that a certain state of
affairs has the power to cause me to be sad, but it seems that it is something intrinsic to
the state of affairs itself, its being sad, which has this causal power. That is, when I claim
that the death of my friend is sad, I am saying something specifically about that state of
affairs itself—that the state of affairs is itself objectively sad.
One may feel some discomfort about attributing the property of sadness to a state
of affairs. One may feel that being sad cannot be a property of something that is not
a subject of experience. But I have not yet claimed that sadness actually is a genuine
property of states of affairs, only that it is experienced as one.
Can one go further, and attribute sadness—the property of being sad as a matter of
purported objective fact—to a state of affairs? Well, what are the paradigm cases of
attribution of properties to things that have no experience, things like physical objects
and states of affairs? Shape and weight are paradigm cases of properties of physical
objects. When we attribute shape and weight to an object, for example, we take these to
be genuine, wholly mind-independent properties of that object. Colour, too, is expe-
rienced as a wholly mind-independent property of an object. It seems to us that col-
our—colour-as-we-subjectively-experience-it—is spread out over the object. Nor do
we need to limit ourselves to physical objects and states of affairs, for the same is true
21
We also sometimes attribute these properties to ourselves: I am sad, happy, horrified, and so on. There
is a sense in which attributing sadness to a state of affairs and to myself is inextricably tied together. I am sad
because my friend’s death is sad.
22
My claim here is that evaluative phenomenology is sufficient for making evaluative property attribu-
tions. So I am leaving open the possibility that a subject may be in a depersonalized state and judge that
something is beautiful but have no feeling.
of numbers. We take the natural numbers to have wholly objective, mind-independent

properties—perhaps those defined by the Peano axioms.
The attributions of properties to purported objective phenomena that we make
when we feel emotions are more like the number case than the physical object case,
inasmuch as the former two are not made directly on the basis of the sensory phenom-
enology, as I have defined it, and although the number case and the emotion case may
be different in many other respects.
The attributions of sadness, the tragic, the joyful, the uplifting, and so on all essen-
tially involve evaluative property attributions. At the most basic level, we can divide
these into attributions of negative value and attributions of positive value. Emotions
such as sadness, horror, fear, and rage all represent disvalue, and emotions such as love,
joy, and courage all represent positive value. But things are not always so simple. One
can have a mix of emotions; and there is more to representing the sadness of a friend’s
death than a representation of its disvalue. All I want to claim at the moment, how-
ever, is that a representation of disvalue is actually essential to a representation of sad-
ness, and also, quite generally, that all emotions essentially involve evaluative property
attributions.
So far I have been focusing on the property attributions of ‘sadness’, ‘the tragic’, ‘the
joyful’ that we make in having emotional reactions to states of affairs, but of course
there is also the phenomenology of feeling sad, feeling joyful, and so on. I have already
discussed one of the ways in which the intentionality and the phenomenology of an
emotion are essentially related (in virtue of the awareness-of-awareness feature of
experience). I now want to turn to another, which also, as it turns out, depends on the
awareness-of-awareness feature of experience.
The question I am concerned with is this: what is the relationship between the evalu-
ative phenomenology that is (an essential) part of—and distinctive of—the emotion
of sadness (for example), and the property attribution of sadness to a state of affairs?
To make the answer to this question as clear as possible, it will be helpful to first
discuss perception, and in particular, the case of a visual experience involving colour
experience. Consider the case of having a visual experience of a red ball, and let us focus
on the redness aspect of this experience. We can ask, what is the relationship between
the ‘reddish’ phenomenology, what we might call ‘phenomenal redness’, that is (an
essential) part of—and distinctive of—the visual experience of the red round ball, and
the property attribution of redness, what we might call ‘redness-as-seen’, to the ball?
I will begin with the following two claims:
(1) Generally, a visual experience’s sensory phenomenology is intimately and meta-
physically linked to the kinds of property attributions made in the having of the
experience. In particular, the phenomenology of ‘what it is like’ to experience
reddishness, phenomenal redness, suffices for the attribution of what we take to
be the colour redness of the ball, such as redness as seen.
And:
(2) It is in virtue of the subject’s awareness of the sensory phenomenology of the

visual experience that the phenomenology suffices for being about or focally
directed towards the redness-as-we-see it of the ball. It is partly because of one’s
awareness of experiencing ‘phenomenal redness’ that one attributes the prop-
erty redness to the ball, which is essentially and phenomenologically linked to
‘phenomenal redness’.23
(1) and (2) provide a two-step account of this particular aspect of the relation
between the phenomenology of a visual experience and the intentionality of visual
experience, which concerns the fact that certain sorts of property attributions (attribu-
tions of properties to physical objects in the world) are made in visual experience.
There is a sense in which the general idea behind (1) is fairly uncontroversial in the
philosophy of perception.24 Taken generally, it simply asserts that a perceptual experi-
ence’s phenomenology is intimately and metaphysically linked to the kinds of property
attributions made in having the experience. For example, most representationalists
(in the philosophy of perception) accept the claims that if two perceptual experiences
have the same phenomenological content, then necessarily they share (a kind of)
intentional content, and if two perceptual experiences share (a kind of) intentional
content, then necessarily they have the same phenomenological content.25
(2) is more controversial because it asserts that this link between the phenomenol-
ogy and the intentionality of the experience is made in virtue of the subject’s aware-
ness of the experience itself, and in particular in virtue of the subject’s awareness of
the phenomenology of the experience. Therefore, (2) is appealing to the awareness-of-
awareness thesis to account for the metaphysical link between the phenomenology
and the intentionality of experience.
Let us consider the relationship between (1) and (2) in more detail. In having the
visual experience as of a red round ball we are aware of the ‘reddish’ phenomenology
or the ‘phenomenal redness’. In being aware of the phenomenal redness we are aware
of a property of experience. (This follows from my characterization of the awareness-
of-awareness thesis introduced in section 3.) But we also attribute a colour property to
the ball, what I have described as ‘redness as seen’; this property seems to be spread out
over the ball. According to (1), the property attributed to the object is internally and
essentially linked to the phenomenological property we are aware of. That is, the ‘red-
dish’ phenomenology itself suffices for the attribution of redness to the ball. In attrib-
uting redness to the ball, I attribute the property whose essential intrinsic character I
take to be fully revealed in the phenomenological-qualitative character of experience.
23
I do not think the ‘rednesss’ property we attribute to the ball in virtue of our phenomenal red experi-
ence can be identified with any particular surface reflectance property, because of the possibility of inverted
spectra. I do not have the space to argue for this here.
24
Of course there are well-known philosophers who reject (1). See e.g. Block (1990, 2010).
25
The qualification ‘kind of ’ on ‘intentional content’ is meant to leave open the possibility that a visual
experience can have different kinds of intentional content, only some of which metaphysically co-vary with
phenomenology.
Finally, according to (2), it is only because we are aware of phenomenal redness that we
attribute redness to the ball. (If we were not aware of the phenomenal redness in the
way that we are, we would not then attribute it to the ball.)
But more needs to be said about the internal link between the awareness of ‘phe-
nomenal redness’ and the redness we attribute to the ball. When the subject is aware of
‘phenomenal redness’, she is aware of a purely phenomenological property of experi-
ence, a property that is not in fact a property of non-phenomenological things in the
world. But there is also a straightforward (and fundamental) sense in which the subject
attributes phenomenal redness to the ball. And the redness-as-seen that is attributed
to the ball is experienced as an objective (or mind-independent) property of the ball,
not as a property of experience; it seems to be spread out over the surface of the ball.
Nonetheless, there is clearly a relationship between ‘phenomenal redness’, the property
of experience one is aware of, and the property attributed to the ball. This relationship
cannot be strict identity, because as I have already mentioned, phenomenal redness is
experienced as a property of experience, whereas the redness attributed to the ball is
experienced as an objective property. I propose that the relationship is one of resem-
blance. An aspect of the phenomenal redness resembles an aspect of the redness attrib-
uted to the ball. We feel the experience of phenomenal redness gets it exactly right
about, completely conveys the intrinsic qualitative character of, the objective property.
This is what our belief in the resemblance consists in.
To summarize so far. The awareness-of-awareness thesis is a way of saying what
experience is; it is constitutive of experience that the subject is aware of having experi-
ence. The awareness-of-awareness thesis also accounts for two of the ways in which
the phenomenology and intentionality of experience are related. First, phenomenol-
ogy itself is intentionally structured, because in having an experience the subject is
aware of, hence, trivially, intentionally aware of, having that experience. Second, it is
essentially in virtue of the subject’s awareness of the phenomenological properties of
the experience, in having the experience, that the subject attributes certain properties
to objects. But a further link between the phenomenological properties the subject is
aware of having in having an experience, which are properties of experience, and prop-
erties attributed to objects must be made because the properties attributed to objects
are experienced as objective, mind-independent properties. Resemblance is part of
explaining how the phenomenological property experienced as a property of experi-
ence gets linked to the property attributed to the object—a property experienced as an
objective, mind-independent property.26
I now hope the perceptual case will help make clear how the evaluative phenomenol-
ogy that is (an essential) part of—and distinctive of—the emotion of sadness (for exam-
ple) and the property attribution of sadness to a state of affairs are related. Consider the
following two claims about emotion, which are analogous to the perception case.
26
I am not here attempting to give an account of why we experience certain properties as objective prop-
erties. Perhaps such a discussion will involve appeal to spatial properties. I am taking as a datum that we do
(3) The evaluative phenomenology that is integral to an emotional experience suf-

fices for the evaluative property attributions made in that experience. Thus (for
example) the evaluative phenomenology that is an integral part of feeling sad
about a friend’s death suffices for the experience’s attributing the negative evalu-
ative attribution (and thus the attribution of sadness) to that friend’s death.27
And:
(4) It is in virtue of the subject’s awareness of experiencing the evaluative phenome-
nology that the experience makes the evaluative property attributions it does. It
is in virtue of the subject’s awareness of the evaluative phenomenology essential
to sadness that the experience attributes the negative value involved in sadness
(and thus attributes sadness) to the friend’s death.
I will now explain in more detail why the term ‘evaluative phenomenology’ is par-
ticularly apt for the kind of phenomenology involved in emotional experience. All
emotional experiences have positive affect or negative affect, and these affects are asso-
ciated with (what we take to be) the negative or positive value of objects and states of
affairs. This association is the reason I use the term ‘evaluative phenomenology’ to refer
to a kind of phenomenology distinctive and essential to emotion experience, which is
irreducible to sensory or cognitive phenomenology.
The positive or negative evaluative phenomenology (the positive or negative affect)
of an emotion represents the positive or negative quality of the object or state of affairs
represented. Very roughly, the phenomenology of feeling sad about something, which
involves negative affect, experientially represents that that something is of disvalue. So
feeling sad about a friend’s death experientially represents the disvalue of that death.28
Of course, there is more to representing the sadness of a friend’s death than an expe-
riential representation of its disvalue, but what I am claiming is that in this experi-
ence of sadness, the negative evaluative phenomenology is essential to representing
its disvalue and thus its sadness.29 It is also important to note that negative evaluative
phenomenology, for example, differs from emotion to emotion. The negative evalua-
tive phenomenology associated with sadness is different from the negative evaluative
experience ‘redness as seen’ as an objective property of physical objects, and because of this, something more
than appeal to our experience of phenomenal redness characterized as a property of experience is needed.
27
If one thinks that only subjects of experience attribute properties to things, strictly speaking, and
that experiences never do, one can say that the fact that one attributes sadness to one’s friend’s death, in
reacting emotionally as one does, depends essentially on the fact that one’s emotional experience has the
evaluative-phenomenological component it does.
28
As I mentioned in Section 2, a person who is thoroughly depersonalized can make a conscious judge-
ment that something is of disvalue without having any feeling, and this would still be an experiential rep-
resentation of disvalue, because all consciousness is a phenomenological-experiential phenomenon.
I am arguing, however, that these two kinds of experiential representations of disvalue are very different.
Emotions have their own distinctive way of experientially representing evaluative properties. Perhaps one
can put the point by saying that in emotions, one seems to feel the very nature of value and disvalue.
29
One, for example, has to deploy the concepts of friendship and death.
phenomenology associated with horror, and both of these are different from the nega-
tive evaluative phenomenology associated with rage and so on. I will come back to this
point in Section 5. For now, all I want to claim is that negative and positive evaluative
phenomenology represents disvalue and value.
In a visual experience one visually experiences the redness of the ball; in an emo-
tional experience one emotionally experiences the disvalue of a friend’s death. Visual
experience involves experiences of different colours, such as red and not-red, and
emotional experience involves experience of value and disvalue.
The evaluative phenomenology of experiencing sadness about a friend’s death,
the negative affect one experiences, in the specific mode of sadness, accounts for
one’s attribution of the property of disvalue (and so sadness) to the friend’s death. So,
although one may be able to know intellectually (in some sense) that the death of a
friend is of disvalue, and indeed that such a death is sad without feeling an emotion,
one can experience the disvalue of the friend’s death in this distinctive way only if one
has an emotional experience. Data in Star Trek, who does not have any emotions, may
be capable of saying what is of value and disvalue, and in turn saying what is sad and
what is happy, but he cannot experience sadness or happiness and thus he cannot expe-
rience value or disvalue (in the special way I am indicating). A stronger claim about
Data, which I do not have the space to pursue in this chapter, is that although he can say
what is of value and disvalue, and know what is of value and disvalue, he cannot really
know value and disvalue. It may be similar to someone who is blind from birth who can
say this is red, when told that it is a ripe tomato, and indeed know that it is red, but not
know what red is.
A problem that arose for the case of perceptual experience now also arises for the
case of emotional experience. Recall that according to the awareness-of-awareness
thesis, phenomenological properties are experienced as properties of experi-
ence. But it is also true that sadness, for example, is experienced as an objective,
mind-independent property of states of affairs. So, how can evaluative phenomenol-
ogy, experienced as a property of experience, partly account for the attribution of dis-
value to the friend’s death, experienced as an objective, mind-independent property
of that state of affairs?
Although the solution to this problem in the emotion case is not obviously an appeal
to ‘resemblance’, I think a defence can be made for it here as well. The basic idea is that
in experiencing the negative affect (the negative evaluative phenomenology) that is
part of experiencing sadness, the negative affect is itself experienced as something of
disvalue.30 This disvalue experienced in the experience of negative affect resembles the
disvalue that is attributed to the state of affairs of the friend’s death. It is then partly in
virtue of this resemblance relation that the negative affect experienced as a property of
experience represents the disvalue attributed to the friend’s death. More strongly put,
30
This claim will be slightly modified in section 5, because we will always experience determinates of
disvalue rather than the determinable disvalue.
similar to the case of the experience of redness, we feel the experience of feeling the dis-
value of negative affect gets it exactly right about, completely conveys the character of, the
purported objective property of disvalue.
Of course, there are well-known problems with explicating representation in terms
of resemblance. The first thing to note is that I am certainly not claiming that all rep-
resentation is a matter of resemblance. I am only concerned with a certain class of
properties that are experienced as part of conscious experience. In light of this, many
of the problems with the resemblance theory of representation will not apply here. For
example, I am not claiming that in order for the concept of a number to represent a
number there must be some resemblance relation in place.
One problem that might be raised against my account is that resemblance
is a symmetrical relation. So, for example, if a picture of Napoleon resembles
Napoleon and so represents Napoleon in virtue of this resemblance, then because
Napoleon resembles the picture does the man Napoleon also represent the picture
of Napoleon? Clearly not. Representation can be non-symmetrical, even if resem-
blance is not. My general answer to this worry is that only mental states and experi-
ences can be truly said to represent. So the reason we get representation in virtue of
the resemblance relation for ‘phenomenal redness’, for example, and the apparent
property of the object that resembles ‘phenomenal redness’ is because ‘phenomenal
redness’ is part of a conscious experience. The same reasoning applies to emotional
experiences.
I am also not claiming that resemblance is sufficient for representation, even if we
restrict our attention to conscious experiences. For example, two conscious experi-
ences that have phenomenal redness as a part do not represent one another, despite
resembling one another. In order for a conscious perceptual experience to represent
something, it is necessary that the appropriate causal conditions hold.
5. Fine-grained Evaluative Phenomenology

Our sensory experiences are fine-grained in the sense that they are rich and varied
along various dimensions. Each of our sensory modalities has associated with it a par-
ticular type of phenomenology (putting aside cross-modal effects); for example, vision
has colour-shape phenomenology, audition has sound (timbre [e.g. oboe sound] pitch
[e.g. middle C] loudness [e.g. 52 decibels]) phenomenology, and so on. Furthermore,
within each of these types of phenomenology associated with the sensory modalities
there are further fine-grained variations. For example, within the category of colour
phenomenology, many different colour experiences are possible, such as blue experi-
ences, red experiences. And within each of these particular types of colour experiences,
there are more fine-grained variations, maroon experiences, brick-red experiences,
and so on.
My question now is, how fine-grained is evaluative phenomenology? I have so far
only referred to the very general categories of negative evaluative phenomenology and
positive evaluative phenomenology. But are negative evaluative phenomenology and

positive evaluative phenomenology rich and varied in any way analogous to sensory
phenomenology?
I propose to answer these questions first by considering how fine-grained value and
disvalue are. It seems plausible that value and disvalue are genera, the good and the
bad, under which there are different species of goodness and badness. For instance,
dishonesty, injustice, and murder are different kinds of disvalue, and temperance,
kindness, and fairness are different kinds of value.
In a like manner, it seems plausible that negative evaluative phenomenology and
positive evaluative phenomenology are genera, under which there are several different
species. For example, although sadness, fear, and anger all possess negative evaluative
phenomenology, they also feel different; and although joy and love possess positive
evaluative phenomenology, they feel different. One explanation for these differences
in evaluative phenomenology or these different ways of feeling is that different kinds
of emotion involve the representation of different kinds of evaluative properties.
Very generally speaking, for example, although fear and anger both possess negative
evaluative phenomenology, fear is associated with the disvalue of threat, and anger is
associated with the disvalue of injustice, and so correspondingly the evaluative phe-
nomenology associated with each emotion is different.
So far, then, we have one kind of analogy between sensory phenomenology and
evaluative phenomenology. The different sensory modalities and the properties they
represent have different types of phenomenology associated with them, and the dif-
ferent types of value and disvalue have different types of evaluative phenomenology
associated with them. There may, however, be a further way in which sensory phenom-
enology and evaluative phenomenology are alike. Consider again the broad category
of colour. It is a determinable under which several determinates fall: red, blue, purple,
and so on. However, each of these colours is itself a determinable under which several
determinates fall. For example, there are many shades of red. And up to a certain point,
associated with these different shades of red are different ‘reddish’ phenomenological
experiences.
Now consider the emotion of anger. Is there a kind of fine-grainedness of evalua-
tive phenomenology within the emotion of anger analogous to the fine-grainedness
that exists within our experiences of different shades of red? That is, are there different
kinds of evaluative-phenomenological experiences associated with different instances
of anger? Clearly, there are different degrees of intensity of feeling associated with
anger, ranging from mild annoyance to rage. But can the evaluative phenomenology
associated with anger differ along another dimension besides degree of intensity?
Consider the following two instances of anger: feeling angry about being threat-
ened by a hostile neighbour, and feeling angry about terrorism. Try to imagine that the
degree of intensity of feeling is exactly the same in these two cases. Do these instances
of anger feel different qua anger? My inclination is to say that they do. If they do feel
different, what explains this difference?
One possibility already canvassed previously is that just as the difference between
the negative evaluative phenomenology associated with anger and fear can be traced
to the different kinds of evaluative properties typically represented by these emotions,
perhaps different kinds of evaluative properties represented by different instances of
anger can result in different evaluative-phenomenological experiences. However, in
the examples of anger given previously, both cases involve representing the property of
injustice, but that property may be being combined with other evaluative properties in
such a way that explains the difference in feeling. For example, being threatened by a
neighbour may involve representing the possibility of bodily harm to oneself, whereas
terrorism may involve representing further harms such as coercion and intimidation.
Another way in which these instances of anger differ is that in the first case the emo-
tion is directed at a particular person, and in the second case the emotion is directed at
a general state of affairs, or at least no particular person, or no one one knows. So, there
may be a difference between the way we evaluate particular persons versus the way we
evaluate general states of affairs that manifests itself in differences of feeling.
So far, what I have said about the sense in which evaluative phenomenology may be
fine-grained is very speculative, and I do not have the space to adequately deal with
the complexity of these issues here. I will end by saying that for my own part I do think
evaluative phenomenology is fine-grained along various dimensions and that this can
be traced to the very complex ways in which we experience value and disvalue.
In conclusion, I have argued that emotions essentially have a kind of phenomenol-
ogy, evaluative phenomenology, which is irreducible to sensory phenomenology
or cognitive phenomenology. I have tried to show how evaluative phenomenology
is essential to the property attributions we make to objects and states of affairs—for
example, that our friend’s death is sad, or that the situation is the Middle East is dis-
turbing, and so on.31
References
Bayne, T. (2011). ‘The Sense of Agency’. In The Senses: Classical and Contemporary Philosophical
Perspectives, ed. F. Macpherson (Oxford: Oxford University Press), 355‒74.
Bayne, T. and Montague, M. (2011). Cognitive Phenomenology (Oxford: Oxford University
Press).
Block, N. (1990). ‘Inverted Earth’. In Philosophical Perspectives, 4: Action Theory and Philosophy
of Mind, ed. J. Tomberlin (Atascadero, CA: Ridgeview), 53–80.
Block, N. (2010). ‘Attention and Mental Paint’. Philosophical Issues 20(1): 23‒63.
Brentano, F. (1995 [1874]). Psychology from an Empirical Standpoint. Introduction by P. Simons,
trans. A. C. Rancurello, D. B. Terrell, and L. McAlister (London: Routledge).
Caston, V. (2002). ‘Aristotle on Consciousness’. Mind 111: 751–818.
31
I would like to thank Sabine Roeser, Rachel Singpurwalla, Cain Todd, and Jonathan Vanderhoek for
helpful comments on this chapter.

Dretske, F. (1995). Naturalizing the Mind (Cambridge, MA: MIT Press).
Greenspan, P. (1988). Emotions and Reasons (London: Routledge & Kegan Paul).
Gurwitsch, A. (1966 [1941]). ‘A Non-Egological Conception of Consciousness’. In Gurwitsch,
Studies in Phenomenology and Psychology (Evanston, IL: Northwestern University Press),
287‒300.
Harman, G. (1990). ‘The Intrinsic Quality of Experience’. Philosophical Perspectives 4: 31–52.
Horgan, T. and Tienson, J. (2002). ‘The Intentionality of Phenomenology and the
Phenomenology of Intentionality’. In Philosophy of Mind: Classical and Contemporary
Readings, ed. D. Chalmers (Oxford: Oxford University Press), 520‒33.
Husserl, E. (2001 [1900‒1]). Logical Investigations, 2 vols. Translated by J. N. Findlay with revised
translations by Dermot Moran (London and New York: Routledge).
Kriegel, U. (2009). Subjective Consciousness: A Self-Representational Theory (Oxford and
New York: Oxford University Press).
Nussbaum, M. (1990). Love’s Knowledge (New York: Oxford University Press).
Nussbaum, M. (1994). The Therapy of Desire (Princeton, NJ: Princeton University Press).
Montague, M. (2011). ‘The Phenomenology of Particularity’. In Cognitive Phenomenology, ed.
T. Bayne and M. Montague (Oxford: Oxford University Press), 121‒40.
Montague, M. (forthcoming). ‘The Life of the Mind’. In Phenomenal Qualities: Sense, Perception,
and Consciousness, ed. P. Coates and S. Coleman (Oxford: Oxford University Press).
Prinz, J. (2004). Gut Reactions: A Perceptual Theory of Emotion (New York: Oxford University
Press).
Roberts, R. (2003). Emotions (Cambridge: Cambridge University Press).
Rosenthal, D. (2005). Consciousness and Mind (Oxford: Oxford University Press).
Smith, D. (1989). The Circle of Aquaintance (Dordrecht: Kluwer Academic Publishers).
Smithies, D. (2013). ‘The Nature of Cognitive Phenomenology’. Philosophy Compass 8(8): 744‒54.
Solomon, R. C. (1976). The Passions (Garden City, NY: Doubleday).
Tye, M. (1995). Ten Problems of Consciousness (Cambridge, MA: MIT Press).
Tye, M. (2000). Consciousness, Color, and Content (Cambridge, MA: MIT Press).
Tye, M. (2002). ‘Visual Qualia and Visual Content Revisited’. In Philosophy of Mind: Contemporary
and Classical Readings, ed. D. Chalmers (Oxford: Oxford University Press), 447‒56.
Tye, M. (2009). ‘Representationalist Theories of Consciousness’. In The Oxford Handbook
of Philosophy of Mind, ed. B. McLaughlin, A. Beckerman, and S. Walter (Oxford: Oxford
University Press), 253‒67.
Zahavi, D. (2006). Subjectivity and Selfhood: Investigating the First-Person Perspective
(Cambridge, MA: MIT Press).
4
Emotion, Attention, and the Nature
of Value
Michael Brady
Introduction
Emotions can have epistemic value as a result of their links with attention. In particu-
lar, emotions can facilitate, through the capture and consumption of attention, more
accurate assessment or appraisal of one’s evaluative situation. This account faces a gen-
eralization problem, however, for empirical evidence suggests that the links between
emotion and attention are more complicated than it first appears. In particular, the
evidence indicates that whereas negative emotion focuses attention narrowly, positive
emotion broadens attentional focus, and hence does not capture and consume atten-
tion in the same kind of way. This suggests that the proposal above does not general-
ize to cover negative and positive emotions, and is to that extent suspect. I consider
the empirical evidence in some detail, and show how the different effects of valence
on attentional breadth do not tell against my account. I then present an explanation
of why negative emotion narrows and positive emotion broadens attentional focus,
which both fills an explanatory gap, and tells us something interesting and important
about the nature of the values, and the content of evaluations, that are associated with
emotional experience.
1. Emotion, Attention, and Evaluation

I have argued elsewhere that emotions can have epistemic value as a result of their links
with attention.1 In this section I want to outline the general picture I favour, before—in
the section that follows—raising a serious objection to it.
The idea that attention is an important part of emotional experience is widely
accepted, for there is considerable evidence, from everyday life as well as from psy-
chology and neuroscience, that emotion and attention are closely linked. In particular,
1
Brady (2009, 2010, 2011, 2013).
Emotion, Attention, and the Nature of Value 53
there is widespread agreement that emotions direct and focus our attention on to
objects and events that are potentially significant for us. This commonplace thought
is supported by empirical evidence which indicates that emotions improve our capaci-
ties for detecting and analysing important stimuli. With respect to detection, evidence
indicates that subjects perform better on visual search tasks—that is, the time taken
for subjects to detect a particular target decreases—when the target of their search is
an emotional target, such as a face expressing positive or negative emotion, a snake, or
a spider.2 Another well-known emotional effect on attention is apparent in the ‘atten-
tional blink’ test: when presented with visual stimuli appearing at the same place in
very rapid succession, subjects will often fail to detect the second target; however,
detection of the second target is much more likely when it is an emotional word.3 So
emotions are capable of increasing the speed and capacity with which we detect and
attend to significant events in our environment.
One explanation of why emotion governs attention in this way appeals to the fact
that as human beings we are presented with vast amounts of information about the
state of the world and the state of ourselves, only some of which will be relevant to
our concerns. Given that we have limited mental resources, we face a problem of
efficiently locating or identifying which information in the environment is poten-
tially important: we have a need, that is, to pick out potentially significant stimuli
from the mass of irrelevant stimuli that impinge upon our senses. However, the
cognitive cost of our consciously scanning the environment for significant stimuli
would be prohibitively high. Some theorists think that emotional systems have been
set up—by evolution and by socialization—to solve this problem.4 So on this view,
fear quickly and reflexively directs and focuses a subject’s attention onto a potential
threat, informs the subject of her evaluative situation; that is, that she is threatened,
and mobilizes a subject’s resources to enable her to react appropriately. These include,
importantly, behavioural and motivational resources: fear automatically directs and
focuses attention, and at the same time primes the subject for the appropriate fight-
or-flight response. Although I want to steer clear of making strong claims about the
‘function’ of emotions, it is not implausible to think that emotions have value in part
because they play the role of alerting us, quickly and at little mental cost, to objects
and events that are potentially significant for us. To the extent that our emotions do
not reliably alert us to important objects, it is tempting to think that they do not oper-
ate as well as they might.
2
Dolan (2002: 1191). 3 See Anderson and Phelps (2001).
4
See Vuilleumier et al. (2003: 419): ‘from an adaptive-evolutionary perspective, it can be assumed that
emotion has a privileged role in biasing the allocation of attentional resources toward events with particular
significance for an organism’s motivational state.’ Further: ‘Given a limited processing capacity, the brain
must meet the challenge of detecting and representing only those stimuli most relevant for on-going behav-
iour and survival. It is likely that attentional mechanisms evolved to enable the brain to regulate its sensory
inputs so as to afford such selective perceptual processing and goal oriented action.’
54 Michael Brady
There is, however, another important link between emotion and attention—one
that is much less discussed, but which can be used to support the idea that emo-
tion has a further epistemic role to play; namely; that of promoting a reassessment
or re-evaluation of the ‘evaluative landscape’, and thereby facilitating more accurate
appraisals or representations of value. For our experience, and the empirical evi-
dence, tells us that emotions such as fear and shame do not just automatically and
reflexively direct and focus attention; they can also capture and consume attention.
To say that attention is captured and consumed by emotional objects and events is
to say that such objects and events continue to occupy our attentional resources,
often making it difficult for us to disengage and shift focus elsewhere. Emotions such
as guilt and shame and disappointment are not simply short-term mental events.
Instead, they can stay with us, persisting and dominating our mental thoughts and
reflections.
We saw previously one plausible thesis about the value of reflexive, automatic control
and direction of attention; namely, to alert us very quickly, and at little cognitive cost,
to objects and events that are potentially important. We might now ask whether atten-
tional persistence in emotional experience has value or serves any important need.
Clearly, the value of the persistence of attention does not lie in (its being necessary for)
alerting us to potentially significant objects and events in our environment. Instead, I
want to claim that attentional persistence can enable or facilitate an enhanced repre-
sentation of potentially significant objects and events: by keeping our attention fixed
on some object or event, emotions can enable us to get a better grasp of our evaluative
situation, by allowing us to discover reasons or considerations that bear on the accu-
racy of our initial, reflexive, and automatic emotional appraisal.5
If this is the case, then emotional experience can be regarded as involving two
appraisals or evaluations. The first is a fast-and-frugal reflexive appraisal of potential
importance: when attention is drawn, quickly and automatically, to a potentially sig-
nificant object or event, our emotional response can be understood as involving an
initial appraisal of that object or event. Here it is helpful, and increasingly common,
to employ the language of the ‘perceptual model’ of emotion, and to maintain that this
initial appraisal is a matter of our ‘seeing’ the object or event in an evaluative light.
Thus, when we are afraid our attention is drawn, reflexively and automatically, to a
potential threat, which we thereby see as dangerous. However, the persistence of atten-
tion in emotional experience can facilitate awareness of the features that make it the
case that some object or event has (or lacks) some evaluative property—a property that
it seems to have in the first emotional blush. As a result, the persistence of attentional
focus enables us to assess whether or not something that seems to be a particular way—
dangerous in the case of fear, insulting in the case of anger, and so on—really is that
way. This leads to a second assessment or appraisal of the object—one that the subject
5
See de Sousa (1990: 196): ‘[P]‌aying attention to certain things is a source of reasons.’
forms in light of her discovery and awareness of the relevant features and considera-
tions.6 So emotions involve an initial appraisal akin to a sensory perception of value,
and a second appraisal akin to a judgement which results from an endorsement, or
rejection, of this initial appraisal.
What can be said in favour of this proposal? For one thing, there is a good deal of
empirical evidence that emotions enhance our representational capacities by captur-
ing and consuming attention.7 But the idea has philosophical support as well. There
is, typically, an epistemic need for reassessment and reappraisal, given that the initial
appraisal will be relatively indiscriminate. This is because accuracy and discrimination
is the price to be paid for a fast, frugal, and automatic response to potentially impor-
tant stimuli. As such, fast-and-frugal emotional responses are often rather crude: we
generally respond with fear to looming objects, loud noises, and crawling creatures,
even though most of these will not pose a threat to us. It is therefore important for us to
have a more accurate system of appraisal. Moreover, there is good reason to think that
emotion is important in the provision of such a system. Thus, Thomas Reid, who is one
of the few philosophers to have written on the connection between emotion and atten-
tion, claimed that ‘[i]‌t requires a strong degree of curiosity, or some more important
passion, to give us that interest in an object which is necessary to our giving attention
to it. And, without attention, we can form no true and stable judgement of any object.’8
And: ‘[a]ttention may be given to any object, either of sense or of intellect, in order to
form a distinct notion of it, or to discover its nature, its attributes, or its relations and
so great is the effect of attention, that, without it, it is impossible to acquire or retain a
distinct notion of any object of thought.’9 So for Reid, emotion (or ‘passion’) is neces-
sary for us to pay attention to some object or event, and paying attention is necessary
for us to form an accurate (‘a true and stable’) judgement about that object or event.
Reid would thus be sympathetic to the idea that emotional control of attention has
epistemic value insofar as it makes us aware of the features and considerations that
have a bearing on our evaluative situation, and in so doing facilitates a better grasp of
the potentially important objects and events that we face.10
Now although Reid’s claims about necessity are too strong, since we can intention-
ally direct or fix our attention onto some object in the absence of emotion, never-
theless he is surely correct to stress the importance of emotion in the direction and
control of attention, and the importance of attention to an accurate evaluation of our
circumstances. Although we can intentionally and non-emotionally fix and direct
6
Or, if we are less optimistic, this is an assessment the subject forms in light of her inventing reasons that
support her initial assessment—as when, for instance, someone endorses her feelings of jealousy and judges
that her husband is being unfaithful on the grounds of inventing ‘reasons’ that support how she feels.
7
See, for instance, LeDoux (1996). 8 Reid (1969: 184‒5). 9 Reid (1969: 76–7).
10
The quotation from Reid might suggest that attention is captured by an emotion—such as curiosity—
that is additional to the original emotional response. But on my view it is the same emotional experience
that both directs and captures attention. This is compatible with what Reid says, since the ‘more important
passions’ will include things like fear and anger, love and pride. And we can assume both that such emotions
direct attention, and—with Reid—that such emotions are needed for attentional persistence.
56 Michael Brady
our attention onto some object or event, this is usually very costly in terms of men-
tal resources, in which case there is a significant advantage in having a system which
keeps our attention fixed with little in the way of conscious effort on our part. If con-
siderations of mental economy speak in favour of the automatic and reflexive direction
and focus of attention in emotional experience, then similar considerations speak to
the emotional fixing and consumption of attention.
But it is not simply a matter of cost: an interest in some object that is motivated by
emotion would seem to be more stable and persistent than a non-emotional interest
maintained through sheer strength of will. In the absence of emotion, it is all too easy
to lose interest in some object or event, partly because of the effort involved, but also
partly because a lack of emotion usually means that some object or event does not mat-
ter to us, in which case there seems little point in our continuing to take an interest in it.
If an accurate appraisal of some object or event typically requires more than a fleeting
interest in it, then attentional persistence will typically be needed to facilitate such an
appraisal. Even if the emotional capture and consumption of attention is not strictly
necessary for us to get an accurate picture of our evaluative situation, therefore, the
emotional governance of attention for this end is extremely valuable. As such, the idea
that emotions themselves play an important role in the promotion of accurate evalu-
ation and appraisal, through their effects on attention, therefore enjoys a good deal of
plausibility. This does not mean, of course, that the effects of emotion on attention and
appraisal are always positive. Our evaluative take on the world can, after all, be skewed
in all kinds of ways by our emotions. All I want to claim is that it is unlikely that we will
achieve much in the way of accurate evaluative assessment without our attention being
fixed by emotion. One need not be overly optimistic about the contributions emotions
make to our epistemic life to think this.
2. A Generalization Problem

I have proposed that emotions capture and consume attention, and can thereby
facilitate an enhanced representation of various aspects of our situation. It could be
argued, however, that this account faces a generalization problem, in light of empiri-
cal evidence that some emotions broaden rather than narrow attentional focus. This
suggests that some emotions facilitate a ‘global’ rather than a ‘local’ assessment of our
evaluative situation, which seems to tell against the idea that such emotions facilitate
an enhanced representation of the object or event in question. The thought is that a
broader or more global assessment of one’s evaluative situation will result in neglect
of, rather than focus on, the object or event, such that it is difficult to see why atten-
tion in this instance promotes an enhanced appraisal of the object or event. So the idea
that some emotions involve a broader attentional focus would seem to be incompatible
with the idea that emotions in general have epistemic value insofar as they facilitate a
more accurate representation of emotional objects and events. Now what is interesting
is that the empirical evidence, to be described later, suggests that the different effects
on the focus of attention vary with a difference in the valence of emotional experience.
In short, the evidence indicates that while negative emotions tend to narrow atten-
tional focus, positive emotions tend to broaden such focus.11 The worry for my account,
then, is that it will only seem plausible (if at all) with respect to negative emotional
experience. In this section I will review the empirical evidence, and consider whether
it really does count against my proposal in this way.
The empirical evidence that the valence of emotion has different effects on breadth
of attention is of three (related) kinds. The first is that negative and positive emotions
lead to opposing biases in global‒local visual processing tasks, which indicates oppos-
ing effects on attention. In particular, negative emotions involve a bias for local pro-
cessing (or evaluation), whilst positive emotions involve a bias for global processing
(or evaluation). The second is that negative emotions impair a subject’s ability to detect
peripheral figures, whereas this impairment disappears if the subject experiences posi-
tive emotion or affect. The third is that negative and positive emotions have differing
effects on attention more broadly construed; that is, not simply visual attention. In
particular, evidence indicates that positive emotions tend to promote ways of think-
ing, grouping, and categorizing that are more inclusive, unified, and accommodating,
whereas negative emotions elicit ways of thinking, grouping, and categorizing that are
more exclusive, fragmented, and lacking in unity. Let us consider these three types of
empirical evidence in turn.
(i) Valence Affects Breadth of Visual Attention

There have been a host of studies indicating that differently valenced emotions give rise
to different biases on global‒local visual processing tests. This suggests that differently
valenced emotions have different effects on the breadth of attention. The paradigm
for this kind of experiment was a global‒local visual processing task in Kimchi and
Palmer (1982). Basso et al. (1996), Gaspar and Clore (2002), and Frederickson (1998,
2001; Frederickson and Branigan 2005) have all conducted versions of this experi-
ment. In what follows I will describe the experiment as conducted by Frederickson and
Branigan (2005).
The researchers’ primary hypothesis was that positive emotions such as amusement
and contentment would ‘produce a global bias on a global–local visual processing task,
consistent with a broadened scope of attention’; the researchers also wished to ‘test the
corollary hypothesis that, relative to a neutral state, two distinct negative emotions
(anger and anxiety) [would] . . . produce a local bias on a global–local visual processing
11
It might be objected that some positive emotions seem to narrow attention: love, for instance, tends
to focus attention on the beloved. But this putative counter-example can be accommodated. For one thing,
we might interpret the empirical evidence as telling us that that positive emotions as a class tend to broaden
attention and negative tend to narrow it. For another, the evidence indicates that positive emotions generate
a particular kind of broadening of attention, which is compatible with the idea that love focuses attention of
a different kind on a beloved. I will return to this issue in §3.
58 Michael Brady
Figure 4.1
task’.12 In order to test this, the researchers divided subjects into three groups, and
showed each group a film clip intended to elicit a positive, negative, or neutral emo-
tional response. The former were clips about ‘penguins waddling, swimming, and
jumping’ which primarily elicited enjoyment, and a film about nature showing ‘fields,
streams, and mountains in warm, sunny weather’ which primarily elicited content-
ment. The middle category involved a film featuring ‘a group of young men taunting
and insulting a group of Amish passers-by in the street’ which primarily elicited anger
and disgust, and a film of ‘a prolonged mountain climbing accident’ which primarily
elicited anxiety and fear. A final clip—of ‘an abstract display of coloured sticks piling
up’—served as a neutral control condition, and elicited ‘virtually no emotion’.13
Participants were then assessed using a global-visual processing task involving
the kind of display depicted in Figure 4.1, where a triad of figures was presented that
contained a ‘standard’ figure on top and two ‘comparison’ figures underneath. (These
kinds of figures originally appeared in Kimchi and Palmer 1982.)
Subjects were asked ‘to indicate which of the two comparison figures was more simi-
lar to the standard figure. Judgements could be based either on the global-configural
aspects of the standard figure, or the local elements comprising it.’ In one version of
the test, the standard figure is a square made of square elements (Fig. 4.1). If subjects
chose the comparison figure on the bottom left—the square made of triangular ele-
ments—‘their choice is based on the global configuration of the standard figure’. If,
instead, subjects chose the figure on the bottom right—the triangle made of square
elements—then ‘that choice is based on the local detail elements of the standard figure.
Participants were instructed to give their first, most immediate impression of which
comparison figure looks more like the standard figure.’14
12
Frederickson and Branigan (2005: 318).
13
14
Frederickson and Branigan (2005: 319‒20).
The results were that ‘the two positive emotion film clips—Penguins and Nature—
produced the largest global biases’ when compared with the neutral film and negative
emotion films. As a result, this suggests that ‘people experiencing positive emotions
exhibit broader scopes of attention than do people experiencing no particular emo-
tion or people experiencing negative emotion.’ In other words, it was found that ‘per-
sonality traits associated with negative emotions (anxiety and depression) correlate
with a local bias consistent with a narrowed attentional focus. By contrast, traits asso-
ciated with positive emotions (subjective well-being and optimism) correlate with a
global bias consistent with a broadened attentional focus.’15 Similar results were seen in
global‒local visual focus tests conducted by Basso et al. (1996) and Gaspar and Clore
(2002), although in these cases the findings indicated that positive and negative moods
promote attention to global and local information, respectively.
(ii) Valence Influences Target Detection

Evidence of a different kind of effect that valence has on attention is presented in
Derryberry and Tucker’s (1994) paper ‘Motivating the Focus of Attention’, and
based upon studies by Brandt, Derryberry, and Reed (1992). These studies propose
that negative affect ‘leads to a relative narrowing of attention’, and support this with
evidence from target-detection tests. One experiment indicated that negative feed-
back prior to a search test leads to an impairment in the subject’s ability to detect
peripheral targets. In particular, negative feedback (or ‘failure’) caused an increase
in response times for targets at locations that were far from a central feedback sig-
nal.16 A second study from the same researchers indicated that negative affect led
to increased response times to (and hence an impairment in the subject’s ability
to detect) elements arranged in a global form—such as local T-shaped elements
that are arranged in a larger T-shaped form.17 The authors write that ‘these studies
suggest that failure promotes a relative narrowing of attention, impairing detec-
tion of peripheral visual targets and global forms.’18 There is, in addition, ‘more
extensive evidence that motivational states regulate the breadth of attention’ to be
found in studies on anxiety. ‘Visual attention may become overly focused on cen-
tral targets at the expense of peripheral targets, and object perception may empha-
size local rather than global aspects of form.’ They cite many studies indicating that
stress and anxiety impairs ‘accuracy or speed of response to peripheral targets’. For
instance, ‘Weltman et al. (1971) proposed that there may be either a narrowing of an
attentional “beam” under stress, with the peripheral portion disappearing first, or a
selective enhancement of central vision, leaving less attentional resources to deploy
to the periphery.’19
15
Frederickson and Branigan (2005: 316). 16 Derryberry and Tucker (1994: 179–80).
17
Derryberry and Tucker (1994: 180). 18 Derryberry and Tucker (1994: 180).
19
Derryberry and Tucker (1994: 181).
60 Michael Brady
(iii) Valence Influences Attention More Generally

A third set of experiments indicates the different and more general effects that positive
and negative emotions have on attention. Derryberry and Tucker write:
although anxiety effects have . . . been examined most often in studies of visual atten-
tion, there is also evidence that anxiety constricts attention within the verbal cognitive
domain as well. In the classic studies of Spence and associates (Spence, 1958), anxious
subjects performed well in remembering words that were closely related, but their per-
formance was poorer than that of non-anxious subjects with words that were remotely
related . . . Mikulincer, Kedem, and Paz (1990a) applied Rosch’s analysis of category organi-
zation to examine individual differences in anxiety. Anxious subjects tended to reject
non-prototypical exemplars of categories, the breadth of their categories was reduced,
and they perceived less relatedness between different categories. In subsequent research,
Mikulincer, Paz, and Kedem (1990b) found that anxious subjects categorized objects with
less inclusive and more discrete categories, suggesting to these authors that anxiety may
lead to conceptual fragmentation.20
There is, correspondingly, a significant amount of evidence that positive emotion

and positive mood influences attention in the opposite direction. Isen has, in a number
of studies over the last two decades,21 found that subjects
in whom positive affect has been induced, in any of a variety of simple ways (e.g., watching five
minutes of a comedy film, receiving a small bag of candy . . .) have a broader range of associates,
and more diverse associates, to neutral material . . . Similarly, people in such conditions are able
to categorize material more flexibly, seeing ways in which nontypical members of categories can
fit or be viewed as members of the category.22
As such, subjects who are experiencing positive emotion or affect tend to ‘create
and use categories more inclusively, to group more stimuli together, and to rate
more low-prototypic exemplars as category members than did control subjects’.23
In one such experiment to show the latter, subjects who have been induced to
experience positive affect were more likely to judge ‘fringe examplars’ of a category
such as vehicle—for instance, elevator and camel—as included within that cate-
gory (Isen and Daubman 1984). As such, Isen’s work suggests that positive emotion
or affect produces a flexibility in organization and a capacity to integrate diverse
material.24
There is, as a result, a significant body of empirical research to suggest that nega-
tive and positive emotions have different effects on the breadth of our attentional
focus. Negative emotions tend to promote local bias on visual processing tasks, whilst
positive emotions tend to promote global bias. Negative emotions tend to impair
20
Derryberry and Tucker (1994: 182). 21 See Isen (2000) for a summary and details.
22
Isen (2000: 418). 23 Derryberry and Tucker (1994: 184).
24
Isen (1990: 89). See also Frederickson and Branigan (2005: 316).
detection of peripheral figures through a narrower focus on the target, while positive
emotions do not lead to a similar increase in response times on target detection tasks.
Negative emotions tend to promote thinking and categorization that is exclusive,
narrow, less unified, or more fragmented, while positive emotions tend to promote
thinking that is inclusive, integrative of diverse elements, broader, and more coher-
ent. Insofar as we think that these different forms of categorization reflect differences
in attentional focus, such that thinking that is exclusive results from a narrower focus
while thinking that is integrative results from a broader focus, then we have a third
set of experimental results which indicate that the valence of emotion affects breadth
of attention.
Of course, we ought to be wary of making any grand claims about the relation
between valence and attentional focus on the basis of these specific empirical findings.
Moreover, some recent research suggests that it is intensity, rather than valence, which
predominantly affects attentional focus. This work suggests that high-intensity posi-
tive and negative emotions narrow attentional focus, whereas low-intensity positive
and negative emotions broaden focus.25 We might, in addition, note the work of Paul
Slovic, which indicates that the positive emotion of compassion decreases when we
focus on global suffering.26 Nevertheless, in what follows I will assume that the experi-
ments described previously do provide evidence that valence has an effect on breadth
of attention, and then see whether this causes problems for my thesis. And I want to
argue that, even if we take the evidence in this way, my thesis emerges unscathed.
3. Constitutive and Consequential Attentional

Focus
It strikes me that none of the evidence for valence having an effect on breadth of atten-
tion is telling against my account of the epistemic importance of emotion in facili-
tating a more accurate appraisal of emotional objects and events. Let us begin with
Derryberry and Tucker’s studies, which indicate that negative emotion during a search
test leads to an impairment in the subject’s ability to detect peripheral targets—in
particular, those studies which indicate that negative feedback (or ‘failure’) causes an
increase in response times for targets far from the central feedback signal. As a result,
such studies are thought to indicate that subjects who experience neutral or positive
emotions have broader attentional focus. However, such evidence is not fatal to my
account. The criticism of my proposal suggests that the broadening of attention in pos-
itive emotion would not facilitate an enhanced representation of the emotional object
or event, but would have the opposite effect: namely, positive emotion, by broaden-
ing attentional focus, would lead the subject to neglect the object or event that is the
centre of attentional focus in negative emotional experience. But even if the evidence
25
See Gable and Harmon-Jones (2010). Slovik (2007).
26
62 Michael Brady
in question shows that negative emotion narrows attentional focus, it does nothing to
show that positive emotion causes neglect at the local level, given that response times
for targets closer to the central feedback signal were the same for negative and positive
emotions.27 If so, there is no evidence for positive emotion decreasing attention when it
comes to the emotional object or event, on the assumption that the emotional object or
event is that which plays the role of the central feedback signal in the empirical studies.
A second response casts doubt upon whether any of the three different types of
empirical study cited previously causes a problem for my thesis. For it seems to me that
all of the experimental work done to test the effect of valence on attention fixes on what
I will call consequential attentional focus, rather than constitutive attentional focus.
Let me explain. It is plausible to think that the focus of attention onto some object or
event is constitutive of emotional states. For it seems impossible to think of someone
as being afraid of the upcoming exam without the upcoming exam being the target of
her attention. Similarly, it seems impossible to think of someone as being angry about
the Principal’s pay rise without the Principal’s pay rise being the object or event that
he attends to when angry. Let us therefore call the intentional target of emotion—that
which the emotion is about—the object of constitutive attentional focus.
To say that some emotion involves, as a constituent, attention to some object or event
does not entail, of course, that the subject’s attention is solely focused on that object. My
guilt at my bad behaviour might make me attend not only to what I did, but to ways in
which I can make reparations; my disappointment at the team’s defeat might lead me
to pay attention not only to the loss, but to the possibility of alleviating my feelings with
a trip to the pub. So emotion can also make us pay attention to ‘strategies for coping’
with some emotional object or event. In addition, emotion can draw our attention to
memories of similar events, as when my sorrow at my grandmother’s death leads me
to think about my grandfather’s death. Emotion can motivate attention to imaginative
possibilities, as when my happiness when she accepts my invitation to go on a date
leads me to imagine where we might go for dinner, what I might wear, and so on.
It is not obvious whether paying attention to strategies for coping is constitutive of
emotional experience. The case might be made that fear necessarily involves thinking
of attack or escape; but other emotional experiences do not seem to involve action ten-
dencies or thoughts of coping. Consider, for instance, aesthetic contemplation, which
certainly involves paying attention to an art object, but does not seem to involve much
if anything in the way of action in response to, or in order to cope with, our emo-
tional-evaluative situation. And it certainly seems true that memories and imaginings
generated by emotion are not themselves constitutive of emotion, but are what we
might call the objects of consequential emotional attention. In other words, emotional
experience can generate attentional focus on things other than the object or event that
elicited the emotion (my grandfather’s death, her saying ‘yes’ to my asking for a date),

27
and which are consequential to the attentional focus on the emotional object or event
itself—a focus that, as we have seen, is partly constitutive of the emotion in question.
Now my proposal about the epistemic value of emotion is a proposal about the epis-
temic value generated by attentional focus which is constitutive of emotion. This is
because I argue that emotion can keep attention fixed on the object or event that elic-
ited it, and in so doing can motivate the search for and discovery of reasons that bear on
the accuracy of our initial evaluation of that event. So my proposal is a proposal about
(the value of) constitutive attention. But it seems to me that all of the evidence cited
about the effects of valence on attentional focus is evidence that fixes on consequential
attention—that is, on attention paid to objects and events that are other than those
objects and events that elicit, and are the focal target of, the emotional experience.
Consider, in this light, the experiments that purport to show that positive emotions
lead to a global bias, and negative emotions to a local bias, in visual-processing tasks.
In Frederickson’s version of this experiment (which purports to represent the first
conclusive evidence for such bias after Derryberry’s original 1992 studies, and as such
might be taken as a model for such experiments in general), emotion is first elicited by
getting subjects to watch a film, and then subjects are asked which of the comparison
figures most closely resembles the standard figure. But clearly the attentional focus in
the latter cases is consequential attentional focus, since the object of attention in the
experiment is not the object that generated, and was the intentional target of, the emo-
tional itself. That is, when a subject gains enjoyment from watching Penguins, his con-
stitutive attentional focus is on the film (or the behaviour of the penguins in the film);
when he is asked to judge which of the comparison figures looks most like the standard
figure, his consequential attention focus is on the figure(s). If so, then evidence from
such experiments showing that enjoyment broadens attentional focus says nothing
about whether positive emotion broadens constitutive attentional focus, and thus dis-
tracts the subject from attending to, and reassessing or re-evaluating, the object that
elicits that emotion. For all that the evidence indicates, and as the ‘target-detection’
evidence cited previously suggests, positive and negative emotions do not differ in the
extent to which they motivate focus on and reflection about the emotional object or
event itself. The fact, therefore, that positive emotions lead to a broadening of conse-
quential attentional focus is, from the standpoint of my thesis, simply beside the point.
A similar conclusion is warranted when we turn to the third body of evidence, gener-
ated by the research of Isen and her colleagues. Here too experimental evidence shows
the influence of positive affect on categorization of objects that are not those that elic-
ited the emotion, and hence that are the objects of consequential attentional focus. To
quote Frederickson, ‘individuals induced to feel positive affect more often saw fringe
exemplars of a given category as included within the category . . . report increased pref-
erence for variety and accept a broader array of behavioural options . . . [and display an]
ability to integrate diverse material’.28 But it is not the exemplars or behavioural options
28
64 Michael Brady
which were the objects of constitutive attention during the emotional experience, and
so the fact that positive affect broadens attention so that fringe exemplars are more
often included in a category says nothing about the effects of positive emotion or mood
on attention to the intentional object of the emotion.
By the same token, and as Derryberry and Tucker write:
anxious subjects tended to reject non-prototypical exemplars of categories, the breadth of
their categories was reduced, and they perceived less relatedness between different catego-
ries . . . subsequent research . . . found that anxious subjects categorized objects with less inclusive
and more discrete categories, suggesting to these authors that anxiety may lead to conceptual
fragmentation.29
But again, exemplars and categories are the objects of consequential attentional
focus, rather than constitutive, and so the fact that negative affect causes conceptual
fragmentation says nothing about the effects of negative emotion or mood on atten-
tion to the intentional object of the emotion.
The empirical evidence of differing effects of valence on breadth of attention is,
therefore, clearly fixated on consequential attentional spread or narrowing, rather
than constitutive attentional focus. Such experimental evidence therefore fails to tell
against my proposal, which concerned the epistemic importance of constitutive atten-
tion to objects and events. If so, then the charge that my account of the epistemic value
of emotion suffers from a generalization problem, since it applies only to negative and
not to positive emotions, is unwarranted.30
Nevertheless, we might think that the empirical evidence leaves us with a different
kind of problem; namely, to explain why positive affect broadens, and negative affect
narrows, attentional focus. This is something that psychologists and, to my knowledge,
philosophers, have failed to address. So in the remainder of this chapter I will attempt
to make up for this lack. The explanation will be compatible with two of my central
claims—that emotional and affective experience involves an evaluative seeming or
appearance, and that the persistence of attention in such experience can facilitate an
enhanced representation of that initial assessment of one’s environment. As such,
the explanation is something that I can adopt to enhance my general picture of the
epistemic value and role of emotion. But of more importance, and certainly of more
interest, is the fact that the explanation promises to tell us something significant about
the nature of value and about the content of emotional appraisal and evaluation. As a
result, the empirical evidence has relevance, not because it poses a putative threat to a
29
30
The distinction between constitutive and consequential attentional focus might not be as sharp as the
previous discussion suggests. Sometimes focusing on something other than an emotional target—as when
one looks for escape routes from the fire—is necessary for a correct evaluative assessment of the target as
putting one in danger. So perhaps a more accurate way of making the distinction is to talk of the difference
between constitutive and merely consequential attention, where the latter is drawn to objects or events that
have no bearing on the value or otherwise of object or event that is the target of constitutive attention. My
point is then that the experimental evidence is clearly fixated on merely consequential attentional effects.
particular account of the relation between emotion, attention, and epistemic value, but
because it has important implications for the nature of value and of valuing.
4. Valence and Value

Suppose that positive emotion leads to broadened attention, and that this involves a
tendency towards a global and inclusive perspective rather than a local and restricted
outlook, a tendency ‘to create and use categories more inclusively, to group more
stimuli together, and to rate more low-prototypic exemplars as category members . . .’31
A possible explanation for this is grounded in the content of the evaluation that is
partly constitutive of positive and negative emotional experience. Given the nature of
emotional and affective assessment, the preferences and groupings displayed under
experimental conditions are precisely what we would expect. Or so, at least, I want to
argue.
The experimental evidence detailed in §2 indicates that broadening of attention
occurs with both positive emotion and positive mood. If so, then an explanation for
the effect of valence on attention ought to accommodate this fact, and this gives us
some reason to think that the explanation should appeal to something that emotions
and moods have in common. I want to suggest that the common element is an evalu-
ative element. In particular, I want to suggest that positive emotions and moods share
an evaluation of something as good, and negative emotions and moods share an evalu-
ation of something as bad. Of course, there will be a difference in the specificity of that
which is evaluated as good or bad. This is because emotions are typically distinguished
from moods precisely on the grounds that the former have more specific objects than
the latter. The objects of moods, that is, seem more general: when I am happy the world
seems like a wonderful place; when I am content I regard my life as going well; when
I am anxious my environment seems threatening; when I am depressed my prospects
seem bleak.32 Let us assume, then, that positive emotions and positive moods each
involve an assessment of goodness, and that negative emotions and negative moods
each involve an assessment of badness, and that the general difference between emo-
tions and moods concerns the specificity of the object of these affective states.
How might the idea that emotions and moods share similar evaluative elements help
explain the experimental data? At this point I want to introduce an account of what
intrinsic value is, and hence an account of the content of the evaluations that emotions
and moods share. The account is due to Robert Nozick, and is presented in Chapter 5
of Philosophical Explanations. In this chapter Nozick wishes to investigate the nature of
(intrinsic) value, and raises the intriguing possibility that ‘the basic dimension of value’—
that is, that which ‘underlies and generates our value ranking’—is ‘unity in diversity’

31
I think therefore that it is a mistake to distinguish emotions from moods on the grounds that the for-
32
mer, but not the latter, have intentional objects.

66 Michael Brady
or ‘organic unity’. He illustrates the notion by first appealing to aesthetics, and then pro-
vides examples of other areas where value seems to vary with degree of organic unity. He
writes:
Theories of the arts often extol the virtues of unifying diverse and apparently unrelated (or not so
tightly related) material; the order of the work effects this unification. Unity in a painting can be
established in many ways: by the way forms lead the eye through it and by relationships of forms,
textures, thematic material, color, tones, and so on. A unified painting will be tied together by
various of these modes of relationship. However, it is not merely its degree of unity that deter-
mines the value of a painting . . . The degree of diversity enters, also. The more diverse the mate-
rial that gets unified (to a certain degree), the greater the value.33
Biologists also appeal to the organic unity displayed by organisms, and Nozick
claims that the higher the degree of organic unity an organism has, the more valuable
it is regarded. He writes:
let us assume we can rank organisms roughly in accordance with their degree of organic unity,
so that most plants come below most animals, with higher animals coming above the lower ones.
Sentience and then consciousness add new possibilities of unification over time and at a time,
and self-consciousness, being an ‘I’, is an especially tight mode of unification. Thus the ranking
of organisms in accordance with degree of organic unity matches our value ranking of them,
with people above other animals above plants above rocks.34
Similarly, we value ecological systems in virtue of their ‘intricate relationships, equi-

libria, and complicated patternings’; we value scientific and metaphysical theories
insofar as they represent systems which are inclusive of a great many diverse things
and exclude ‘random or chance facts, so that each fact has a reason or is necessary’.35
So ‘over this great range of things—the arts, organic life and systems, scientific theo-
ries—the dimension degree of organic unity seems to capture our notion of (degree of)
intrinsic value. Organic unity is the common strand to value across different realms’,36
and it is ‘organic unity [that] underlies our judgements of value’.37 He argues, in much
the same way, that disvalue is a matter of ‘disunity, disharmony, strife, and so forth’.38
As a result, Nozick thinks that the basic dimension of intrinsic value is unity in diver-
sity—the unity or integration or inclusion or grouping together of diverse elements in
a coherent whole. By the same token, disvalue is a matter of disunity—the fragmenta-
tion of, or the lack of integration or coherence or harmony between, diverse elements.
33
Nozick (1981: 415‒16). 34 Nozick (1981: 417).
35
Nozick (1981: 417). Later on he writes: ‘some philosophical theories strive to overcome disharmonies,
bifurcations, disunities. The Hegelian system, for example, sets itself to overcome the dualisms of man and
nature, subject and object, freedom and community, finite subjectivity and infinity, and so forth. Why are
these dualisms something to be overcome, why do they dissatisfy us? I suggest that the reason is a value
reason. These dualisms prevent the highest degree of organic unification, and so prevent the highest value’
(p. 421).
36
Nozick (1981: 418). 37 Nozick (1981: 421).
38
Nozick (1981: 420).
Now we might balk at Nozick’s claim that organic unity is the basic dimension of
value, that which underlies all of our judgements and rankings of value, that which all
valuable things have in common. One reason is that some things that display unity and
integration of diverse elements seem bad rather than good: a concentration camp is an
example. Against this, Nozick points out that such a camp would not count as valuable,
since it involves significant amounts of disunity between people, disharmony, frag-
mentation, and strife. Perhaps a better example would be a harmonious, well-func-
tioning cult where members have been brainwashed into unity of belief and behaviour
and where there is no disharmony between the members. Still, such a grouping would
involve a serious lack of integration between the members and reality, and to that
extent would not count as valuable. There are other potential counter-examples. Some
aesthetes value disharmony rather than harmony, such as people who listen to free jazz
or read the novels of William Burroughs; some claim to find beauty in chaos rather
than order; some value simplicity—in an environment, or an idea—over diversity.
Nevertheless, there are responses available here as well: what is valuable about free
jazz is the fact that it promises to disrupt an order that has become stultifying, boring,
no longer diverse. Here, this musical form can be taken to have extrinsic rather than
intrinsic merit, and so does not cast doubt upon Nozick’s account of intrinsic value.
Similar reasoning applies to the idea that there is beauty in chaos, which is perhaps best
understood as the thought that beauty can emerge from chaos. But it is also possible to
regard genres like free jazz as intrinsically valuable along Nozickian lines, for it is pos-
sible to value free jazz because it is a way in which music can be expanded to encom-
pass and include new and diverse elements, structures, styles, and so on. Finally, the
value of simplicity of some idea (for example) is typically thought to vary with explana-
tory power, where this is a function of the diversity of things that the idea applies to.
The thought that we value simplicity is therefore compatible with the thought that we
value unity in diversity.
As the previous discussion suggests, it might be difficult to garner support for (or
opposition to) Nozick’s theory by focusing on our intuitions about organic unity and
value rankings. For it might always seem possible to interpret examples given in sup-
port of his claim about positive value as in fact involving disunity, or, conversely, inter-
pret putative counter-examples to his view as being highly unified after all. A better
kind of reason to favour Nozick’s view is therefore needed. And a better kind of reason
can in fact be given. For the view promises to provide a plausible explanation of why
positive valence broadens, and negative valence narrows, attentional focus. As such,
additional support for Nozick’s thesis is that it provides an explanation of the empiri-
cal data explained in §2—an explanation that has hitherto been lacking. Because of
this, the experimental evidence about emotional effects on attention might tell us
something interesting and important about the nature of value, and the content of the
appraisals that are partly constitutive of emotional experience.
To see this, recall Frederickson’s experiment where subjects were put in a happy
mood after watching Penguins, and then expressed a global bias when selecting the
68 Michael Brady
comparison shape that most resembles the standard shape. If my claim about mood
involving evaluation is correct, then subjects in a happy mood make a general assess-
ment or evaluation of their environment: the world, or their immediate environment,
seems good to them. If Nozick is right that the basic dimension of intrinsic value is the
unity or integration of diverse elements, then subjects who assess their environment as
good or valuable will therefore assess it as an environment in which diverse elements
are unified and integrated. To employ the language of the perceptual model of emo-
tion, people in happy moods will see the world as inclusive and coherent. When faced
with a forced choice situation, such subjects will therefore display a global bias because
such a preference reflects their experience of the world as integrated and coherent. In a
forced choice situation, the subject will thus pick the figure that coheres with or reflects
how the world appears to them in emotional experience.39
A similar conclusion is warranted when we consider Isen’s experiments. Recall that
for Isen and her colleagues, subjects experiencing positive emotion or affect tend to
‘create and use categories more inclusively, to group more stimuli together, and to rate
more low-prototypic exemplars as category members than did control subjects’.40 But
this is precisely what we should expect, if subjects experiencing positive emotion are
assessing the world as inclusive and integrated; for such use, categorization and rating
coheres with or reflects an assessment of the world as inclusive and integrated. That
is: if subjects see the world as coherent and connected, then this explains their ten-
dency to categorize and group objects in that world in a way which reflects how they
see it. The behaviour of subjects in experimental conditions is exactly what we would
expect, given the way that the world appears to subjects experiencing positive emotion
or mood.
By the same token, we have an equally straightforward explanation of the experi-
mental evidence suggesting that negative valence narrows attentional focus. For sub-
jects who are anxious will, on my account, be subjects who evaluate their environment
as bad in some way. If Nozick is correct, then anxious subjects will see the world as
disunified, fragmented, exclusive, and lacking in coherence. But if anxious subjects
see the world in this way, then it seems obvious why anxious subjects tend ‘to reject
non-prototypical exemplars of categories’, why ‘the breadth of their categories was
reduced’, and why ‘they perceived less relatedness between different categories’. In other
words, the truth of Nozick’s account of disvalue helps to explain why ‘anxious subjects
categorized objects with less inclusive and more discrete categories, suggesting . . . that
39
Might negative emotional experiences also involve seeing the world in a coherent and unified way? For
instance, extreme anxiety might result in paranoid fantasies, such as conspiracy theories, in which disparate
elements of experience are linked together in a coherent and unified way. However, it is plausible to think
both that extreme anxiety results from or is correlated with seeing the world as extremely fragmented, as
one in which the paranoid person is cut off from the rest of society, and that the conspiracy theories and
similar fantasies are a way of alleviating such anxiety, precisely by making sense of the world. So the patterns
of evaluation that are distinctive of extreme anxiety and paranoia are compatible with the general line I am
pursuing here.
40
anxiety may lead to conceptual fragmentation’. This is because assessing and catego-
rizing objects and exemplars in these ways reflects the fact that they see the world as
fragmented and lacking in coherence. As a result, the ‘global’ evaluation of our envi-
ronment that is present in positive affect results in a preference for categorizations that
are unified and inclusive, whilst the ‘local’ evaluation of the environment that is pre-
sent in negative affect results in a preference for categorizations that are exclusive and
fragmented. We thus categorize the world and its objects in a way that reflects how we
see the world and its objects. Indeed, this explanation seems so simple and obvious
that it is difficult to see how things could be otherwise.
5. Conclusion
I have argued that empirical evidence showing that valence influences attention in
different ways does not undermine my account of the positive epistemic value that
emotion can have. I have also argued that the empirical evidence tells us something
interesting about the appraisals or evaluations that emotions and moods involve.
For Nozick argues that positive affect involves seeing the world or one’s environment
or one’s life as unified, integrated, and coherent, and that negative affect involves
seeing these things as fragmented, as lacking in unity, harmony, and integration.
And this, as we have seen, provides a simple and straightforward explanation for
why people choose and categorize as they do when experiencing positive and nega-
tive emotions. The empirical evidence thus lends a degree of support for Nozick’s
proposal. Indeed, we might think that we can garner further support for the idea
that value is a matter of organic unity from another aspect of emotional experience
that has to do with valence, attention, and value. This is the idea that a useful cop-
ing strategy when bad things happen is to attempt to integrate such things into the
pattern of one’s life, to think that such things might make sense or have a place or
a meaning in one’s overall existence. Common sense and psychological evidence
show that this is often an effective way of reducing emotional suffering and alleviat-
ing negative moods.41 If Nozick is right, however, it should come as no surprise that
such attempts are effective in this way. Seeing how bad things might fit into a broader
picture will tend to alleviate suffering precisely because this is a matter of bringing
diverse elements into unity, of integrating disparate things into a whole; and see-
ing things in this way—seeing things as good, in other words—can both reflect and
generate positive emotion and positive mood. Adopting an inclusive viewpoint to
bad things is therefore a way of reducing negative emotion, and producing positive
emotion in its stead, precisely because it is a way of seeing the good, rather than the
41
See Frederickson (2003). See also the previous comment in fn. 39 about conspiracy theories and the
alleviation of anxiety.
70 Michael Brady
bad, in one’s situation. But this suggestion and possibility will have to be the topic for
a different occasion.42
References
Anderson, A. K. and Phelps, E. A. (2001). ‘Lesions of the Human Amygdala Impair Enhanced
Perception of Emotionally Salient Events’. Nature 411: 305‒9.
Basso, M. R., Schefft, B. K., Riss, M. D., and Dember, W. M. (1996). ‘Mood and Global–Local
Visual Processing’. Journal of International Neuropsychology 2: 249‒55.
Brady, M. S. (2009). ‘The Irrationality of Recalcitrant Emotions’. Philosophical Studies 145: 413‒30.
Brady, M. S. (2010). ‘Virtue, Emotion and Attention’. Metaphilosophy 41: 115‒31.
Brady, M. S. (2011). ‘Emotions, Perceptions, and Reasons’. In Morality and the Emotions, ed. C.
Bagnoli (Oxford: Oxford University Press), 135‒49.
Brady, M. S. (2013). Emotional Insight: The Epistemic Role of Emotional Experience
(Oxford: Oxford University Press).
Brandt, J., Derryberry, D., and Reed, M. (1992). ‘Failure and Attentional Narrowing’. Manuscript.
de Sousa, R. (1990). The Rationality of Emotions (Cambridge, MA: MIT Press).
Derryberry, D. and Tucker, D. M. (1994). ‘Motivating the Focus of Attention’. In The Heart’s
Eye: Emotional Influences in Perception and Attention, ed. P. M. Niedenthal and S. Kitayama
(San Diego: Academic Press), 167‒96.
Dolan, R. J. (2002). ‘Emotion, Cognition, and Behavior’. Science 298: 1191‒4.
Frederickson, B. L. (1998). ‘What Good Are Positive Emotions?’ Review of General Psychology
2: 300‒19.
Frederickson, B. L. (2001). ‘The Role of Positive Emotions in Positive Psychology: The
Broaden-and-Build Theory of Positive Emotions’. American Psychology 56: 218‒26.
Fredrickson, B. L. (2003). ‘The Value of Positive Emotions’. American Scientist 91; 330‒5.
Frederickson, B. L. and Branigan, C. (2005). ‘Positive Emotions Broaden the Scope of Attention
and Thought‒Action Repertoires’. Cognition and Emotion 19: 313‒32.
Gable, P. and Harmon-Jones, E. (2010). ‘The Blues Broaden, but the Nasty Narrows: Attentional
Consequences of Negative Affects Low and High in Motivational Intensity’. Psychological
Science 21: 211‒15.
Gaspar, K. and Clore, G. (2002). ‘Attending to the Big Picture: Mood and Global versus Local
Processing of Visual Information’. Psychological Science 13: 34‒40.
Isen, A. M. (1990). ‘The Influence of Positive and Negative Affect on Cognitive Organization:
Some Implications for Development’. In Psychological and Biological Approaches to Emotion,
ed. N. Stein, B. Leventhal, and T. Trabasso (Hillsdale, NJ: Erlbaum), 75‒94.
42
An earlier and rather different version of this chapter was first presented at a conference on ‘Emotion,
Self, and Time’, held at the University of Geneva in May 2011. Later versions were presented at a workshop
on ‘Emotions and Attitudes’ at the University of Southampton in March 2012, at the University of Stirling,
at Edinburgh University Philosophy Society, and at the Postgraduate Reading Party at the University of
Glasgow. I would like to thank the organizers of these events, and participants on all of these occasions, for
helpful feedback and questions. Special thanks are due to Sabine Roeser and Cain Todd for their very helpful
comments on a draft of the chapter.
Isen, A. M. (2000). ‘Positive Affect and Decision Making’. In Handbook of Emotions, 2nd edn.,
ed. M. Lewis and J. M. Haviland-Jones (New York: Guilford), 417‒35.
Isen, A. M. and Daubman, K. A. (1984). ‘The Influence of Affect on Categorization’. Journal of
Personality Social Psychology 47: 1206‒17.
Kimchi, R. and Palmer, S. E. (1982). ‘Form and Texture in Hierarchically Constructed Patterns’.
Journal of Experimental Psychology 8: 521‒35.
LeDoux, J. (1996). The Emotional Brain (New York: Simon & Schuster).
Mikulincer, M., Kedem, P., and Paz, D. (1990a). ‘Anxiety and Categorization—1’. Personality and
Individual Differences 11: 805‒14.
Mikulincer, M., Paz, D., and Kedem, P. (1990b). ‘Anxiety and Categorization—2’. Personality and
Individual Differences 11: 815‒21.
Nozick, R. (1981). Philosophical Explanations (Cambridge, MA: Harvard University Press).
Reid, T. (1969). Essays on the Active Powers of the Human Mind, ed. B. Brody (Cambridge,
MA: MIT Press).
Slovik, P. (2007). ‘ “If I look at the mass I will never act”: Psychic Numbing and Genocide’.
Judgment and Decision Making 2: 79‒95.
Spence, K. (1958). ‘A Theory of Emotionally Based Drive and its Relation to Performance in
Simple Learning Situations’. American Psychologist 13: 131‒41.
Vuilleumier, P., Armony, J., and Dolan, R. (2003). ‘Reciprocal Links between Emotion and
Attention’. In Human Brain Function, 2nd edn., ed. R. Frackowiak et al. (San Diego: Academic
Press), 419‒44.
Weltman, G., Smith, J., and Egstrom, G. (1971). ‘Perceptual Narrowing during Simulated
Pressure-Chamber Exposure’. Human Factors 13: 99‒107.
5
Emotions as Unitary States
Jonathan Dancy
My main aim in this chapter is to consider whether it is possible to think of an emotion

as a unitary state, and, if so, how to go about it. I will be focusing on what are known as
‘moral emotions’.
The prevalent tendency in recent work on the emotions—though there are notable
exceptions—is to view an emotion as having various parts, or elements, or ingredi-
ents, or aspects. This is because most people think that emotions ‘involve’ belief, atti-
tude, feeling, and motivation. If asked how these things are related to each other in the
emotion, their answers tend to be less than persuasive, and this leaves the emotions as
compilations or collections of disparate elements, when it would be much more satis-
factory to be able to think of them as unitary states. It is far from clear, however, what
this desirable property of unitariness actually amounts to. I reject simple appeals to
causation, and consider various alternative suggestions.
1.
I take as my first stalking horse Thomas Reid’s discussion of approbation and disap-
probation in his Essay on the Active Powers of Man.1 Reid’s views are instructive despite
the fact that it is very hard to extract a consistent position from the various passages
in which he addresses this topic. Approbation and disapprobation—or approval and
disapproval—are not emotions, for Reid, but they do involve emotion. Approbation
occurs when there is moral judgement, recognition of meritorious behaviour, but it
involves, or ‘includes’, further states, one of which is an affection towards the agent, and
the other is a feeling.
1
References to Reid’s Essay here are to the version in W. Hamilton’s edition The Works of Thomas Reid,
DD, 6th edn. (Edinburgh: McLachlan and Stewart, 1863).
Emotions as Unitary States 73
But we approve of good actions, and disapprove of bad; and this approbation and disapproba-
tion . . . appears to include, not only a moral judgment of the action, but some affection, favour-
able or unfavourable, towards the agent, and some feeling in ourselves.2
But the feeling is not a ‘mere’ feeling:

Let me now consider how I am affected when I see a man exerting himself nobly in a good cause.
I am conscious that the effect of his conduct on my mind is complex, though it may be called by
one name. I look up to his virtue, I approve, I admire it. In doing so, I have pleasure indeed, or an
agreeable feeling; this is granted. But I find myself interested in his success and in his fame. This
is affection; it is love and esteem, which is more than mere feeling. The man is the object of this
esteem: but in mere feeling there is no object.3
This tells us that the feeling is pleasure in the case of approbation, and pain in the case
of disapprobation. (Myself, I find this very unconvincing, and not only because Reid
seems to think that pleasure has no intentional object.) But there is also a passage in
which Reid calls that feeling an emotion:
We are next to consider the agreeable or uneasy feelings in the breast of the spectator or judge,
which naturally accompany moral approbation.
There is no affection that is not accompanied with some agreeable or uneasy emotion. It has
often been observed, that all the benevolent affections give pleasure, and the contrary ones pain,
in one degree or another.4
We see here that Reid is happy simply to switch from talk of agreeable and uneasy feel-
ings to talk of agreeable and uneasy emotions; and he understands this agreeableness
and disagreeableness simply in terms of giving pleasure and giving pain.
So there are these three elements in approbation and disapprobation. For Reid, the
most interesting element is the affection, which in the case of approbation he also calls
‘esteem’. This esteem is one of the ‘benevolent affections’.5
We feel a sympathy with every noble and worthy character that is represented to us. We rejoice in
his prosperity, we are afflicted in his distress . . . .
This sympathy is the necessary effect of our judgment of his conduct; and of the real approba-
tion and esteem due to it; for real sympathy is always the effect of some benevolent affection,
such as esteem, love, pity or humanity.6
Here Reid seems to say both that the sympathy felt is an effect of the approbation as a
whole, that is, of the co-presence of judgement, affection/esteem, and feeling, and that
it is an effect just of the esteem. Either way, the esteem, being a benevolent affection, is
a desire for the good of the person esteemed.
2
Essay 3.7; p. 592.1. 3 Essay 5.7; p. 672.2
4
Essay 3.7; p.593.1.
5
Others are parental affection, gratitude to a benefactor, pity and compassion towards the distressed,
friendship, love between the sexes, and public spirit.
6
Essay 3.7; p.593.2.
74 Jonathan Dancy
So Reid offers us a tripartite account of approbation as a combination of judge-

ment, benevolent affection, and feeling of pleasure. And he sometimes talks discon-
certingly about the ingredients of a benevolent affection, which include an agreeable
feeling. But he also addresses the question of how these various things are connected
together:
I am likewise conscious that this agreeable feeling in me, and this esteem of him, depend entirely
upon the judgment I form of his conduct. I judge that this conduct merits esteem; and, while
I thus judge, I cannot but esteem him, and contemplate his conduct with pleasure. Persuade
me that he was bribed, or that he acted from some mercenary or bad motive, immediately my
esteem and my agreeable feeling vanish.
In the approbation of a good action, therefore, there is feeling indeed, but there is also esteem
of the agent; and both the feeling and the esteem depend upon the judgment we form of his
conduct.7
So there is a relation of asymmetrical dependence involved, with the affection and feel-
ing depending on the judgement. Is it a causal dependence? It may seem that this is
what Reid intends, when he writes that ‘while I thus judge, I cannot but esteem him,
and contemplate his conduct with pleasure’. And the language of effect, if not of cause,
is present when he writes that ‘sympathy is the necessary effect of our judgment’. But I
am not convinced that he intends the dependence to be causal. ‘I judge that his conduct
merits esteem; and, while I thus judge, I cannot but esteem him.’ This need not be read
as saying that the judgement causes the esteem. It could mean, instead, that when I
judge that esteem is merited, I naturally respond in the way I judge to be merited. One
might think of this as a normative effect. (Cf. the remarks about one thing naturally
accompanying another, quoted previously.)
2.
I now turn to my second stalking horse, which is a recent account that is explicitly
modelled on Reid’s; namely, that of Sabine Roeser. There is an initial difficulty in put-
ting these two accounts together, because Reid’s account is an account of moral appro-
bation, which he does not conceive as an emotion, while Roeser’s is an account of what
she calls a cognitive moral emotion. Such an emotion is a complex state (CME3) which
consists of:
1. A moral judgement.
2. Positive or negative affection for the relevant agent.
3. An agreeable or disagreeable feeling.8
Essay 5.7; p.673.1.

7
See Roeser (2011: 149). Here I have allowed myself to ‘improve’ Roeser’s text, with her consent, so as to
8
make it more consistent with her original intentions.

We then ask what account is offered of how these ‘features’ or ‘aspects’ are connected.
And the answer to this question is that Reid’s idea that the affection and the feeling are
dependent on the judgement is reversed, being replaced by the claim that ‘Without
the affective aspects we would not be able to fully appreciate the moral value of a situa-
tion.’9 If I understand this right, it amounts to the claim that the judgement is depend-
ent on the affection and feeling. But this is a new sort of dependence, since it seems
to mean, not that the judgement would not be made, but that it would not be true or
correct if the relevant affections and feelings (the ‘affective aspects’) are not in place.
(A weaker and perhaps more plausible version would claim only that the judgement
would be less likely to be true.)
There are other possible readings of the claim that in the absence of the affective ele-
ments we would not be able to fully appreciate the moral value of the situation. One
reading stresses a distinction between full appreciation of value and correct apprecia-
tion of value; correct judgement, on this distinction, would not require affection and
feeling, but full appreciation would. But I am not going to pursue this possible reading
because it looks too much like a persuasive definition: Roeser would just have decided
not to count as ‘full’ appreciation any appreciation that is not accompanied by the rel-
evant affection and feeling, without there being any sense that there is an independent
property of ‘fullness’ that is lacked by a judgement that is not so accompanied. We
have, after all, been told nothing about what would be lacking.
So I return to the idea that the affection and feeling are required for correct apprecia-
tion, where appreciation is not to be distinguished from judgement. I am encouraged
in this reading by Roeser’s remark a few pages later that ‘a missing emotion prevents
a fully fledged judgment’.10 One might, I think, be forgiven for not having a very clear
idea what it is for a judgement to be fully fledged, but we learn more on the next page,
where we read that ‘we need emotions to have access to objective moral truths’. Even
this way of putting the point is at the best misleading, since an emotion, for Roeser,
is a combination of judgement, affection, and feeling, and such things might just be
our access to moral truth rather than something we need in order to get such access;
the point should surely be that without affection and feeling, our judgement will not
capture (objective) moral truth. So it is not that the emotions help our judgement, but
that those judgements that are aspects of cognitive moral emotions are more likely to
be correct.
Even if that is so, however, the judgement and the feeling, as components of that
emotion, may still be separate states. And even if it is true that the presence of these
features helps judgement, there are many other features whose absence would make
our moral judgements less reliable, such as a reasonable understanding of the situa-
tion, experience of similar problems in the past, and sound assessment of the probable
consequences of the various available actions. And there is little temptation to think
of these additional features as further aspects of a unitary state of any sort. That our
9
Roeser (2011: 149). 10 Roeser (2011: 153).
76 Jonathan Dancy
judgement is less reliable without them does nothing to persuade us that there is a uni-
tary state consisting of moral judgement, understanding, experience, and assessment
of probable consequences. All we could say is that these things are all co-present and
related to each other in their various ways. And, for all we have seen so far, the same is
true of the affective elements. That they serve to improve one’s chances of hitting the
objective moral truth does nothing yet to draw us to think of them as part of a complex
but unitary state of which judgement is another part.
A further difficulty lies in the fact that we can think of plenty of moral judgements
that are less likely to be true because of the presence of negative affection for the agent
and a disagreeable feeling of some sort. But there is no suggestion here, I hope, that
moral judgement is always more secure when accompanied by affections and feel-
ings. The idea has to be that a cognitive moral emotion only occurs when we have a
judgement that would be undermined by the absence of the affective aspects, because
it would be less likely to be true in their absence. There may be other moral emotions
(just as there may be other judgements), but a cognitive moral emotion is one that is
well on the way to being true, because of the way it is put together.
If this is the view, we need a reason to believe that it does two things:
1. It will not be vulnerable to the thought that there are many features whose pres-
ence raises the probability that our moral judgement be true but which are not
aspects of a unitary whole of which that judgement is another aspect.
2. It reveals a sense in which a cognitive moral emotion is a unitary state.
It seems to me that the default position here is that judging is one thing and feeling
another. After all, we are allowing that a judgement can perfectly well occur in the
absence of affection and feeling. But the suggestion is that judgement is dependent
upon the co-presence of affection and feeling, not for its existence, but for its (chance
of) truth. Yet even if one allows that this is so, it seems not to be an answer to the ques-
tion of how the various aspects of a cognitive moral emotion are connected; it offers
us the wrong sort of glue. Instead of offering us something like a necessary connection
between the parts, as did Reid, we are told that some parts are conditions, not for the
existence, but for the success of another part. I do not see that this should generate any
sense that the emotion, understood as the sum of those parts, is a unitary state. All that
we have is what we have already seen, that there are various conditions whose presence
makes correct judgement more likely, such as relevant past experience and knowledge
of the context of the action. And these things are not therefore going to be promoted to
being part of an occurrent mental state of a certain sort. So it seems to me that, for all
we have yet seen, the unitariness of these cognitive moral emotions is fictitious.
So far I have been concentrating on Roeser’s claim that ‘Without the affective aspects
we would not be able to fully appreciate the moral value of a situation.’ This was the
claim that reversed Reid’s account of the linkage between judgement, affection, and
feeling. All I have said about it is that we have not yet been given any reason to believe
it true that it would succeed in generating a good sense in which these cognitive moral
emotions are unitary states. But Roeser also says that in a cognitive moral emotion
‘feeling and judging cannot be separated’ (149). This is a much more interesting thesis,
and it is obviously one which, if defensible, would generate a form of unitariness. What
could be more unitary than a complex whose parts cannot be separated? Still, the new
claim looks much harder to defend than its predecessor. For an argument that objec-
tive truth is unachievable (or harder to achieve) in the absence of feeling and attitude
need not go so far as to identify the feeling and the judgement (if, indeed, this is what
is intended). Still, one might be able to move in the reverse direction, arguing that a
sort of combined state of feeling and judgement has epistemic advantages. So this new
remark of Roeser’s raises the prospect of a very different reason for saying that ‘with-
out the affective aspects we would not be able to fully appreciate the moral value of a
situation’, which is that without the affective aspects we would not be doing this sort of
appreciating in the first place, because the affective aspects are part of such an appre-
ciation. This is a more a metaphysical point than an epistemological one.
What we would then see is that the affective aspects are part of the cognitive moral
emotion, as the original formulation of the previous position reveals, and also part of
what originally looked like another aspect of that emotion. This makes possible the
view that, properly understood, cognitive moral emotions are value judgements, as
Roeser claims on p. 149. At first sight one would take this claim to be a confusion, since
it appears to identify the emotion with one of its parts. One could surely be pardoned
for thinking that it is no more true that cognitive moral emotions are judgements than
it is that they are feelings, or affections, and both are as false as that a family is a mother.
But on the approach that is emerging, there is no confusion involved, since the judg-
ing, originally represented as merely one constitutive element in a cognitive moral
emotion, somehow sucks the other two elements up into it. In that sense, there is noth-
ing else for the emotion to be. But we have not yet seen how any of this can be true.
On p. 150 we read that ‘moral emotions let us see salient features’, and that ‘emotions
shape the way we see the world’. But in putting things this way, Roeser seems to have
forgotten that, according to CME, a cognitive moral emotion is (perhaps among other
things) a way of seeing the world, and cannot be said to shape itself, or to enable itself.
The point Roeser should have made here, I suppose, is that it is not emotions but feel-
ings and affections that shape or enable our moral judgements. And we are still left with
two possible reasons why this is so. The first is that the affective aspects increase our
sensitivity to relevant distinctions. The second is that the affective aspects are part of
the judgement, part of the way in which we see the world.
3.
I therefore turn to the work of some other thinkers to whom Roeser appeals to see if
they can be of any help. The main prospect is Margaret Little, who writes (in a passage
that Roeser quotes):
78 Jonathan Dancy
In order to ‘see’ the moral landscape clearly, in order to discern it fully and properly, one must
have certain desires and emotions. Caring, being outraged, being motivated to act—all these are
part of discerning moral features clearly. The ideal epistemic agent herself would have appropri-
ate affect, for it is needed if one is to discern all that there is to see.11
There are various things to note about this passage. The first is its liberal use of adverbs;
‘fully’, ‘properly’, ‘clearly’. The second is the alternative suggestion that without affect
one will miss some relevant facts, since one will not ‘discern all that there is to see’. But
though these issues affect one’s reading of Little, it is not yet clear that our decision on
the point will help us understand Roeser. For Little takes it that having appropriate
affect, namely the right desires and emotions, is part of judging. And it looks as if this
is an entirely different picture from Roeser’s. Roeser has it that the judgement is part of
the emotion; Little has it that the emotion is part of the judgement.
But this may not be quite correct. As emerged in the previous section, Roeser’s more
extreme view was that the judging somehow sucks up the affective aspects, in such a
way that the cognitive moral emotion is that judging (sometimes called ‘a full appre-
ciation’). Little’s view that the desires and emotions are part of clear discernment need
not be taken to be so different from Roeser’s, since all that would be at issue is whether
to think of this sort of judgement itself as an emotion. Roeser would be willing to do
this and Little would not, but it does not seem very much to matter which way one goes
on the point.
However this may be, Little’s claims waver here. There are four things she might be
saying, and it is hard to tell which is the one she is really after. These are: (1) having emo-
tions is a necessary condition for discerning certain moral facts, (2) having emotions is
a necessary condition for discerning certain moral facts clearly, fully, and properly, (3)
having emotions is a part of what it is to discern some moral features, and (4) having
emotions is part of what it is to discern any moral feature. I take it that her true view is
the third of these, which is what makes her interesting to Roeser. But this is very much
a conjecture on my part. Little’s intriguing essay, though it makes various claims about
the emotions, tends in its later stages towards the claim that moral judgement involves
motivation, since it is the discerning of practical relevance. The relation between this
claim about motivation and some claim that moral judgement involves the emotions is
not at all clear to me. A motivated person is no doubt affected by the way she takes the
world to be, but this sort of affectedness need not involve the things that theorists of
the emotions tend to refer to as ‘affect’. And Little distinguishes between emotions and
desires, in a way which seems to allot motivational considerations to desires, leaving
the emotions as playing some other ‘affective’ role.
Roeser also appeals to the work of Patricia Greenspan, in defence of a later expression
of her position, that ‘emotions cannot be resolved into their components’. Greenspan
defines an emotion as an affective response to an intentional object.12 A good example
11
Little (1995: 127). 12 Greenspan (1988).
of such a thing would be a case where I am angry that you forgot your keys. Now it is
true that here we do see an analysis which is genuinely invulnerable to worries about
unitariness, and one can see that it therefore makes a very attractive model for Roeser.
But in fact Greenspan’s views are no help to Roeser, because the various ‘aspects’ of a
cognitive moral emotion do not stand in the same relation as my anger does to your
having forgotten your keys. In fact the sort of unitariness Greenspan offers is the same
as the unitariness of belief and that of doubt (which are both responses to intentional
objects), and this sort of unitariness is to be seen in one of Roeser’s aspects: the judge-
ment. But it does not also serve to link the three aspects of cognitive moral emotions
together in such a way as to show that they ‘cannot be resolved into their components’.
Things are different when we turn to Roeser’s final appeal, to John McDowell.
McDowell, she says, tells us that perceptions of saliences cannot be decomposed into
pure awareness and appetitive state.13 This is because the sort of perception at issue
already includes everything that the supposed independent appetitive state could
offer. So there is no need for a further Humean desire, once a perception of saliences is
in place.
Is McDowell’s picture one under which we have a complex that has components but
which cannot be ‘resolved’ into those components? Somehow I doubt it. McDowell’s
‘perception of saliences’ is a one-component account; his argument is that when we
understand properly the nature of that component, we see that the addition to it of an
independent desire would achieve nothing. This is not the same as saying that there
is a need for the desire but that the two cannot be disentangled. So it emerges that
McDowell, in offering a one-component account, faces no issues about unitariness.
Roeser’s CME3, however, is not in such a happy state, since it departs from McDowell
by saying that we do need to add something about affections and agreeable or disagree-
able feelings to what we say about judgement. It is a three-component account, not
a one-component account. Roeser’s idea seems to be that though these components
are distinct in analysis, they are not ‘resoluble’ into the elements revealed by analysis,
because the affection and the agreeable/disagreeable feeling somehow permeate the
content or nature of the belief or judgement. But this picture remains hard to make
sense of, and we are offered no further help in that direction.
4.
At this point I should stress that the views we have considered so far (except in the
case of Greenspan) are not accounts of emotion in general. Reid was discussing appro-
bation. Roeser was talking about her ‘cognitive moral emotions’, without making any
general claim about any other sorts of emotion there might be. Little was interested in
the relation between emotion and moral judgement. And I am certainly not intending
13
See J. McDowell, ‘Are Moral Requirements Hypothetical Imperatives?’ and ‘Virtue and Reason’, both
reprinted in McDowell (1998).
80 Jonathan Dancy
in this chapter to offer any general remarks about the emotions. What I have been after
is the sort of unitariness (if any there be) enjoyed by the various states we have consid-
ered (approbation, cognitive moral emotions, and a judgement of which an emotion
is a part). I confess that I have considerable sympathy for the rather sceptical attitude
of Jesse Prinz, though, as will now emerge, I do not think he manages to make the
point very effectively. Prinz claims that all ‘component theories’ of the emotions (that
is, theories that do not identify an emotion with any single state) face two problems—
problems that are especially hard for those versions of component theory that ‘either
claim that every instance of an emotion contains all of the kinds of components I have
been discussing, or . . . claim that each emotion must contain at least some of these’.14
These two problems are the Problem of Parts and the Problem of Plenty:
By including everything, one can lose sight of how the different parts hang together. Privileging a
single part is a way of drawing attention to the feature that is most fundamental for understand-
ing emotions. An encompassing account that fails to do this suffers from what can be termed the
Problem of Plenty.
The Problem of Plenty is the counterpoint to the Problem of Parts. The Problem of Parts asks:
What components of an emotion episode are really essential to its being an instance of some par-
ticular emotion? The tempting answer is that all parts are essential. The Problem of Plenty then
asks: If all parts are essential, how do they hang together into a coherent whole?15
The general idea is that if a component theory requires all the parts to be present if
there is to be an emotion, it fails to tell us how those parts ‘hang together’. If it requires
only some of the parts, it faces the question which parts are really essential, and why.
There is a suggestion here that the only way of addressing the question of how all the
parts hang together is to nominate one of the parts as the most fundamental; doing
this enables us to display a sort of coherence that consists in the way all the other parts
cluster round the chosen fundamental part.
It is clear that all component theories need to tell us how the components they iden-
tify ‘hang together’, on pain of failing to show how those components, when all present
at once, constitute ‘a coherent whole’. It is also clear that if a component theory allows
that various different configurations of the components are each of them sufficient for
the occurrence of an emotion, it has to say something about how this can be. And if it
says that some components are more central, or important, or something like that, it
has to say what sort of centrality is at issue and why some components have it when
others do not. These points are undeniable. But there is nothing here to show that there
cannot be answers to them. And the challenge itself is blunted by Prinz’s appeal to an
unanalysed idea of a ‘coherent whole’. What does that amount to? What sort of coher-
ence is it that an emotion, conceived as constituted by the co-presence of various com-
ponents, needs to display? And why is it that the only acceptable way of displaying
coherence is to nominate some one part as fundamental? Without an answer to this
14
Prinz (2004: 18). 15 Prinz (2004: 18).
question, it is hard to feel that the Problem of Parts and the Problem of Plenty have any
real bite.
I do not mean to suggest that there are not serious issues here. But it would be nice
to have a more precise sense of what it is that component theories are supposed to lack.
To go back to the beginning: what would be the difficulty for a theory of emotion that
listed several individually necessary and jointly sufficient conditions for emotion? It
is no answer to this to say that we just do not think that such an approach has much
chance of success. Such a theory could claim that a set of individually necessary and
jointly sufficient conditions is as unified as anyone could possibly want, so that it has
no trouble with the Problem of Plenty, and its answer to the Problem of Parts is just that
the different components are individually necessary and jointly sufficient. Why is this
a bad answer? Prinz addresses this question later with a different example:
Suppose one wants to provide a theory of conscious visual states. What, one might ask, is a con-
scious red experience? . . . [O]‌ne might say that conscious red experiences have several parts.
There is a feeling, a thought, an action tendency, an attention controller, and a memory trig-
ger . . . When asked to point out which one is the red experience, one might point to the whole set
of entities. Red experiences, one might say, have many components.
This complexity would be gratuitous. It would be better to say that a conscious red experience
is a unitary mental entity that has several functions, properties and effects.16
I agree that the tendency to write down everything one wants to say about emotion,
whether as a list of individually necessary and jointly sufficient conditions or as some-
thing more nuanced, is not very promising. But why not? Gratuitousness is not the
problem. Someone who thinks that each component is necessary cannot be accused
of gratuitousness. The real thrust here is the simple claim that it is better to think of a
conscious red experience as a unitary mental entity that has several functions, proper-
ties, and effects. But why is that better? Would not such an account itself face its own
analogue of the Problem of Parts and the Problem of Plenty? If it has several functions
and several properties, do we just offer a list of these, or are we seeking some sort of
coherence? If so, what sort? Once we have said that the relevant ‘mental entity’ has sev-
eral properties, functions, and effects, how is the unitariness of that entity to be estab-
lished, or, as it were, re-established? As far as I can see, the problem of unitariness still
remains. And the evidence for this is that it is perfectly open to Roeser, for example, to
maintain that a cognitive moral emotion is a unitary state that has three ‘aspects’, and
the notion of a function is flexible enough for each such aspect to count as a function.
What would be gratuitous here is not the claim that the relevant state is complex, but
the bald claim that it is unitary, with no further support. (This is intended to remind
you of my argument in the previous section that Roeser offered no effective support for
her repeated claim that a cognitive moral emotion cannot be resolved into its parts.)
Prinz (2004: 241, 242).

16
82 Jonathan Dancy
But have we really seen that non-unitariness is a problem? I think that, to do this,
what is needed is a sort of caricature.
Let us then invent a new concept of a mental state that is constituted by:
1. a geographical judgement about some location;
2. a feeling of heat or of cold;
3. an attitude toward whatever is at that location.
And we are going to call this a ‘frisson’.17 Now I hope that everyone shares my sense
that the notion of a frisson is a fiction. And it would remain a fiction even if one
added a requirement that the three aspects be causally connected, so that the mere
co-instantiation of the three aspects, even if they be all happening to the same person,
would not be sufficient for a frisson. But what is wrong with this notion of a frisson?
There clearly could be such a thing, or rather such a complex, whether causally con-
nected or otherwise. I suggest, as a tempting hypothesis, that there is no reason why
anyone who has one of the ‘elements’ of a frisson should have either of the others. The
addition of the second two elements in a frisson to the first does seem to be, in this
sense, gratuitious. So the notion of a frisson lacks something, and what it lacks could
be considered a sort of unitariness, which I will call R-unitariness. An R-unitary men-
tal state is one that, though complex, is held together by relations of reason—that is,
there is some reason why someone who has one ‘element’ of such a state should have
the others. The link involved in R-unitariness is a normative link.
Consider, by contrast, Reid’s understanding of approbation as a judgement that
makes a certain affection necessary, and a feeling of agreeableness. This sort of talk
about necessary connections between the different elements seems to me to be too
strong. I do not suppose that someone who makes a moral judgement cannot avoid
having the affection and the feeling. It is more that there is a relation of appropriate-
ness at issue. I will say that in such a case, the individual components are R-unified to
each other, and the combination of these distinct features is R-unitary. R-unification
is a relation that holds between the components, and R-unitariness is a property of a
complex state whose components are R-unified.
So this relation of R-unification is far short of any relation of necessary connection.
It is true that a complex of components so related that if one lacks any one component
one necessarily lacks the others, or at least one of the others, is beautifully unified, but
I agree with those who think that this model is too demanding, and also unnecessary.
The same can be said of a complex of components so unified that one cannot change
one without thereby changing another, or all the others. It does not seem at all promis-
ing to try to establish the emotions as unitary states of such exalted kinds. R-unitariness
is one possible form of less exalted unitariness. A complex can be R-unitary without
any suggestion that the components are not separable, or that the whole cannot be
‘resolved’ into its parts.
17
‘Frisson’ is roughly the French for ‘shiver’.
Returning now to Roeser, we should also note that this sort of unitariness is consist-
ent with the idea that one of the elements is somehow primary. In her picture we might
say that the judgement is primary, because the consideration that is a reason to be
pleased and to adopt a certain attitude is so because it is—in the first place, as it were—
a reason to judge that a right action has been done. So though there is unitariness here,
it is not even-handed. Some elements are more central than others, and there is (or at
least can be) an explanation of why this is so.
Are Roeser’s cognitive moral emotions then to be understood as R-unitary in my
sense? I suggest not. For it is not as if someone who judges an action right has, there-
fore, a reason to feel one way rather than another, or to take a certain sort of attitude to
the relevant agent. It is not the judging that grounds the reason: it is the thing judged,
that something wrong is being done. We do not want to say that one can give oneself
reason to feel outraged by judging that a wrong has been committed, for if no wrong
has in fact been committed there is no reason for outrage. The truth seems rather to be
that while the fact that a wrong has been committed gives one reason to feel outraged,
there is a distinct and derivative normative relation between the wrongness-judgement
and the feeling of outrage, but it is one that should be run in terms of rationality rather
than in terms of reasons.18 This would constitute a second sort of unitariness, Rat-
unitariness, where the driving connection is not one of reasons but of rationality, and it
is perhaps one to which Roeser could appeal.
However things may be on that front, I suggest that the notion of R-unitariness pro-
vides something of an answer to Prinz’s charges against component theories, with-
out committing us to obscure metaphysical connections. It also generates a sense in
which a full appreciation is different from a true appreciation. A full appreciation is an
appreciation that is appropriately accompanied by motivational and affective changes.
Fullness, in this sense, is a normative notion.
5.
I now want to present an account of certain emotions—moral emotions, if you like—
which is another instance of this notion of R-unitariness.
It has been standard in the tradition of ethical intuitionism to think of moral intui-
tions as self-evident judgements. But recent work on intuition in general,19 which does
not restrict itself to the moral intuitions but also considers intellectual and philosophi-
cal intuitions, tends to think of an intuition not as a judgement but as an intellectual
‘seeming’. The model for a seeming is the perceptual case, most often our old friend the
Müller-Lyer illusion. A perceptual seeming of this sort need not be a belief; it is better
18
The contrast between the normativity of reasons and that of rationality is a tricky topic which is not well
controlled in the literature. I make my own attempt in Dancy (2009).
19
I am thinking here especially of work by George Bealer, e.g. his (1996), and John Bengson (e.g. his ‘The
Intellectual Given’, forthcoming).
84 Jonathan Dancy
understood as a presentation of a supposed fact, where presentations are distinguished

by the way in which they, as it were, thrust their content at you. (Beliefs do not do
this, and so are not presentations, but representations.) Such presentations can still be
rejected; we can come to believe that things are not the way they are presented to us as
being, though they do not cease to be presented that way when we decide that the pres-
entation is fallacious. On this model, seemings are presentations, but presentations are
not yet beliefs. This applies to intellectual seemings, to philosophical seemings more
narrowly—and to moral seemings of the sort that the classic moral intuitionists appar-
ently had in mind.
Now some moral seemings are indeed just special cases of intellectual seemings.
These moral presentations, like other presentations, are not yet beliefs or judgements,
though of course most of the time we accept that things are the way they are presented
to us, in moral cases as elsewhere. Nor are they accompanied by motivational changes.
When it strikes us that Caesar was wrong to cross the Rubicon, we are not ourselves
necessarily motivated not to do such a thing (or, as Gibbard might say, planning not
to do it). There is absolutely no chance of my getting command of a Roman army on
the borders of Roman Italy. And it can, I suppose, strike one that (that is, it can be pre-
sented to one that) some course of action is morally permissible, though moral seem-
ings of this sort are not necessarily accompanied by motivation of any sort. So there is
no reason to think of these seemings as other than intellectual.
If these moral seemings are to be understood on the model of the Müller-Lyer case,
it should be possible for one to have such presentations while resolutely believing that
they are false. For a possible example of this, consider a case in which an action of
some distant historical figure presents itself to us as wrong, in a way that is consistent
with our deciding that actually it was not wrong though it still presents itself to us as
wrong, or, more plausibly, that it was not as wrong as it presents itself to us as being.
Consider, for example, the Ancient Greek habit of exposing unwanted babies on the
mountainside. It may be that, despite my strong and persisting sense of outrage about
this, I come, by understanding the circumstances in which this practice grew up, and
the beliefs which held the practice in place—beliefs which I of course take to be false—
I cease to condemn the practice as I originally did, or as strongly as I originally did,
though I cannot shake off my original sense of outrage. But most moral presentations
will not involve this sort of internal conflict.
Still, in addition to such intellectual moral seemings, we might also entertain the
possibility of practical seemings which, again, are not judgements, but which can be
endorsed or rejected by judgement. A practical seeming, as I conceive it, is a presenta-
tion of a practical reason. His behaviour might present itself to us as a reason to protest.
So some practical seemings will be presentations of a moral reason to the person
concerned; they are practical because the reason is presented as a reason for us to act.
This is what distinguishes them from intellectual moral seemings, if such there be,
where the reason is not presented as a reason for action but simply as a reason for moral
judgement. In a practical seeming, then, some consideration or other is presented as
a practical reason. It is not that the consideration is presented as being the case. We
may merely believe that things are so, but this is enough to allow their being so to be
presented as a reason for a certain sort of response. I think of this sort of presentation
as motivational; that is, the person experiencing the presentation is motivated thereby.
But this does not yet mean that he believes the consideration to be a reason. Its pres-
entation as a reason can be rejected, and will be if the agent comes to believe that the
consideration is not the reason it presents itself as being. Such motivation can thus be
quelled. But in order to be quelled it must first exist.
The consideration that is presented as a reason for action also serves as a reason to
believe that one has some reason to act. And it also serves as a reason to be motivated.
And, finally, it can serve as a reason to feel one way or another (if motivatedness does
not already introduce feeling). There is a sort of rational unity here, but it is asym-
metrical, since it is its role as a (potential) reason for action that lies at the centre of the
complex. So this is a case of R-unitariness, with one element that is primary, and which
is not judgement but prior to judgement.20
6.
One might ask, of the picture I have just given, whether affect is a necessary condition
of the presentation of a practical reason. My answer would be that if motivation is taken
to be affect, it is; if by ‘affect’ we intend something distinct from motivation, the answer
might be that such affect is appropriate, so that the combination of the presentation of
the reason with that affect is R-unitary, but that one could have that presentation with-
out the affect. The question might then be raised whether an affect-less presentation of
a reason is properly to be understood as a moral emotion. But it is not clear to me that
the notion of an emotion is in good enough shape for this question to be very pressing.
By this I mean that I do not really feel that the emotions form a natural psychological
kind. There may be no single ‘theory of emotion’. And it may be that different emotions
differ from each other in different ways, and that the class of emotions is pretty loosely
unified. So I am not worried about offering a picture of one ‘sort’ of emotion, a moral
emotion that is a presentation of a practical reason, while allowing that other emotions
may not be presentations of reasons at all.
Still, this general insouciance about classificatory issues does not extend to an insou-
ciance about particular accounts of emotions, or accounts of particular emotions. I do
think that accounts of emotion need to display the unitariness of the emotions they
treat.
D’Arms and Jacobson often speak of emotions as being a ‘syndrome’, for instance in
this account of fear:
20
I treat the main suggestions of this section in much greater detail in Dancy (2014).
86 Jonathan Dancy
The fear system, for instance, can plausibly be described as monitoring the environment for
threats to the organism, even if (as neuroscientist Joseph LeDoux claims) there are distinct path-
ways into the syndrome known as fear: a syndrome of directed attention, physiological changes,
affect, and motivation that can be functionally understood as constituting a kind of appraisal of
the circumstances.21
Their appeal to the notion of a ‘syndrome’ worries me. Etymologically, a ‘syndrome’

is just some things that ‘run together’, and the notion of ‘running together’ is not a
very precise one. In what sense do the physiological changes that are (sometimes)
involved in fear run together with the motivation that is (sometimes, again) involved?
Perhaps they are effects of a common cause. But one would expect the use of the term
‘syndrome’ not to be associated with thoughts of a group of effects with a common
cause, because syndromes are normally supposed not to instantiate this kind of focus.
Perhaps the most that can be said is that one often finds these several things together.
But this seems to me distinctly disappointing.
7.
The idea of a syndrome is also in play in Peter Goldie’s conception of grief as a process.22
But in that context it seems more promising, and Goldie does suggest that what he calls
a ‘narrative account’ is appropriate for other emotions as well. This is the Abstract of
the relevant paper:
Grief is . . . a complex pattern of activity and passivity, inner and outer, which unfolds over time,
and the unfolding pattern is explanatorily prior to what is the case at any particular time. The
pattern of a particular grieving is best understood and explained through a narrative account
and not merely through a causal account, for narrative accounts in such cases have powerful
explanatory, revelatory and expressive powers which causal accounts lack. I believe that this
view of grief can be generalised to other kinds of emotions.
Goldie comments that one main difference between his view and others is that they
privilege some element in the mix (some mental state, usually, but at least something
that can occur) as the emotion, while he privileges the process itself, which consists of
mental states and events among other things. And for this reason Prinz’s Problem of
Parts and Problem of Plenty do not get a grip on Goldie’s picture, since nothing is there
privileged in the sort of way necessary for those Problems to arise. (In fact I would say
that the different approaches use different conceptions of privileging.)
So if we ask why someone is behaving in that unusual way, the answer ‘she is griev-
ing’ is explanatory not because it alludes to some present mental event or state, but
because it locates this behaviour as part of a complex and extended process, the pro-
cess of grieving, and the explanation works not by revealing a causal chain of distinct
D’Arms and Jacobson (2003: 138).

21
Goldie (2011). Quotations from Goldie that follow will be to this paper.
22
events, each causally linked to its predecessors and successors, but by showing the
behaviour as part of a coherent pattern. That coherence is what a narrative explanation
is in the business of revealing.
On this account, the notion of grief is to be understood not by thinking about the
differences between the affectual states of grievers and non-grievers, but by thinking
about the various ways in which those who are grieving behave and are affected. Grief
is understood by appeal to grieving and not vice versa. As Goldie puts it, the identity
of the process is not determined by what is going on at any particular time; that this
is grief is not made so by the nature of what is occurring at this moment, but by the
nature of the extended process of which it is a part. By analogy, we might say that I am
winning the match by scoring this point, but still that I am winning the match is not
made so by what is happening just at this moment, but by what has taken place over an
extended period.23
Goldie does not at all wish to deny that there are feelings of grief at a moment, and
that they may be distinctive; his claim is that those feelings are not themselves the grief.
The ‘of ’ here is not the same ‘of ’ as the ‘of ’s in ‘feelings of warmth’ or ‘feelings of inad-
equacy’. Grief is not what is felt, any more than someone who has feelings of anger is
feeling her own anger. The ‘of ’ here is more like ‘characteristic of ’. So we can speak
of feelings of grief without committing ourselves to supposing that grief can be felt,
or that grief is a feeling. And we are in no danger of privileging such feelings in our
account of the emotion of grief.
This obviously raises the question of what ties the various aspects of grieving
together. We do not want just to say that they are all happening at once, or all popping
up occasionally over some extended period. There is to be a pattern to all this, and
talk of patterns brings with it the notion of a sort of coherence. There is an immediate
worry here that the notion of coherence is a name for the problem, not part of the solu-
tion. But I do not think one should be discouraged too easily. For the sort of coherence
we are seeking here is supposedly the coherence of a narrative, and though that notion
may be hard to explicate, it is at least not itself suspect.
What, then, is the difference between a narrative account and a (merely) causal
account? Goldie picks out two differences. The first is that a narrative can somehow
contain, or present, different perspectives, such as that of the agent at the time, that of
the narrator, and that of the agent later. The second is that narratives can contain ‘general
events’; for instance, the first sentence of Proust’s A La Recherche du Temps Perdu: ‘For
a long time I would go to bed early’. But I worry about these two points that they seem
to concern narration rather than what is narrated, as if the relevant relations, the ones
on which the coherence of the narrative depends, are to be thought of as elements of a
narration, but incapable of being instantiated in the course of events narrated. Can gen-
erality be a feature of the events narrated? Surely the events narrated remain stubbornly
23
Goldie quotes C. S. Lewis A Grief Observed, p. 50: ‘I thought I could describe a state; make a map of sor-
row. Sorrow, however, turns out to be not a state but a process. It needs not a map but a history.’
88 Jonathan Dancy
particular. And the sense in which a narrative can contain the perspectives of the agent
both at the time and also later is surely different from any sense in which the narrative
contains the perspective of the narrator; and it is not straightforwardly obvious that
the agent’s perspective is incapable of playing a causal role. In general, if we want to
give an account of grief, we surely do not want to suppose that the coherence of griev-
ing as a process is constituted by relations that are incapable of instantiation out there
where grief and grieving are to be found, and only visible in the story we tell about those
events. For this reason I find it alarming that Goldie writes ‘Of course I wish thoroughly
to resist any assimilation of a narrative with what is narrated.’24 For it seems to me that if
we are to sustain a narrative account of grief, and thus reveal the narrative coherence of
the process of grieving, we must either undermine the distinction between a narrative
and what is narrated, or, accepting that distinction, concentrate purely on what is nar-
rated—which Goldie’s own account fails to do.
What we are trying to capture here is that there is a pattern in the events—in the
observable behaviour as well as in the inner turmoil (such as an inability to concen-
trate, which is not itself an event), and maybe even in physical changes that are not
behaviour, such as premature ageing, skin disorders, or digestion problems. Perhaps
these last are part of the process too; nothing in our account should determine other-
wise. But the sort of pattern, or of coherence, that we find here must somehow be there
for narration to express.
With this qualification, I find Goldie’s approach to grief very plausible. My final
question is whether the sort of coherence he finds in grieving as a pattern is a form of
R-unitariness. If some aspects of the pattern are merely physical, then as far as they go
the answer to this question must be ‘no’. There is no reason for those grieving to age
prematurely (if they do). This would have to be a mere effect, whose occurrence does
not contribute to the coherence of the process. But there is a reason for those who
have lost a loved one to grieve, and grieving includes thinking of the person one has
lost, thinking of shared experiences, feeling sad (or worse) when so doing, having sud-
den emotional rushes, looking over old letters—all of these, we would say, are not just
understandable but perfectly appropriate in the circumstances. This notion of appro-
priateness (otherwise known as fittingness) is the one we need in this context, and it is,
of course, familiar to those who write about reasons.25
References
Bealer, G. (1996). ‘On the Possibility of Philosophical Knowledge’. Noûs 30: 1–34.
Bengson, J. (2014). ‘The Intellectual Given’. Mind (forthcoming).
Goldie (2011: 131, fn.38).

24
I am grateful to the editors of this volume for their persistent demands that I explain myself, which have
25
much improved this chapter.

Dancy, J. (2009). ‘Reasons and Rationality’. In Spheres of Reason, ed. J. Skorupski, S. Robertson,
and J. Timmerman (Oxford: Oxford University Press), 93‒112.
Dancy, J. (2014). ‘Intuition and Emotion’. Ethics 124: 1‒26.
D’Arms, J. and Jacobson, D. (2003). ‘The Significance of Recalcitrant Emotion (or,
Antiquasijudgmentalism)’. In Philosophy and the Emotions, ed. E. Hatzimoysis (Philosophy,
suppl. vol. 52): 127–45.
Goldie, P. (2011). ‘Grief: A Narrative Account’. Ratio 24: 119–37.
Greenspan, P. (1988). Emotions and Reasons: An Enquiry into Emotional Justification (New York
Routledge, Chapman and Hall).
Lewis, C. S. (1961). A Grief Observed (London: Faber and Faber); published under a pseudonym
‘N. W. Clerk’.
Little, M. (1995). ‘Seeing and Caring: The Role of Affect in Feminist Moral Epistemology’.
Hypatia 10: 117–37.
McDowell, J. (1998). ‘Are Moral Requirements Hypothetical Imperatives?’ and ‘Virtue and
Reason’, both reprinted in his Mind, Value, Reality (Cambridge, MA: Harvard University
Press).
Press).
Reid, T. (1785). Essay on the Active Powers of Man. In The Works of Thomas Reid, DD, 6th edn., ed.
W. Hamilton (Edinburgh: McLachlan and Stewart, 1863).
Roeser, S. (2011). Moral Emotions and Intuitions (Basingstoke: Palgrave Macmillan).
6
Relatively Fitting Emotions and
Apparently Objective Values
Cain Todd
1. Essential Contestability and Value Relativism

Sentimentalism holds that for some class of values or evaluative concepts, these can be
analysed in terms of some class of human responses, such as emotions, sentiments, or
attitudes.1 This position can be viewed as one variety of a more general type of analysis
of value or evaluative concepts; namely, a ‘fitting attitude’ analysis. A fitting attitude
analysis of value (FA) holds that for something to be valuable is for it to be the fitting
object of a pro-attitude, where fittingness is to be understood as some kind of norma-
tive constraint governing the correct way of responding to some object or situation
(X); that is, the way that reflects the real value of X:
(FA) For something (object, situation, and so on) X to possess value Φ is for it to be fitting to have
some particular attitude A towards X.
FA inherits the attractions of buck-passing accounts of value, promising to avoid

the metaphysical and epistemological difficulties involved in identifying them.
Moreover, it is supposed to be neutral between realist and anti-realist accounts of
value, for the relevant judgements of fittingness can, it seems, be construed cognitively
or non-cognitively.2
Unfortunately, however, FA also inherits the much-discussed ‘wrong kind of reason’
problem (henceforth WKR) that afflicts buck-passing accounts. This is the difficulty of
providing a non-circular method of distinguishing the reasons that bear on the value
of X from those that do not; that is, of distinguishing the attitudes that are (genuinely)
fitting—that reveal the real value of X—from those that are not. The problem is thus to
specify the right kind of reason(s), to specify the right notion of fittingness. The use of
1
The difference in the target analysis between value and evaluative concepts is not always made clear, but
it is important, as I shall discuss later.
2
For good general overviews of these issues see Lang (2008); Rabinowicz and Rönnow-Rasmussen
(2004); D’Arms and Jacobsen (2000a, 2000b, 2006); Reisner (2009).
Relatively Fitting Emotions and Apparently Objective Values 91
examples in setting up this problem, and hence in establishing the ultimate plausibility
of FA, is important. Here are two typical cases from the literature:
1. ‘Imagine that an evil demon will inflict a severe pain on me unless I prefer this
saucer of mud; that makes the saucer well worth preferring. But it would not
be plausible to claim that the saucer of mud’s existence is, in itself, valuable.’
(Quoted in Rabinowicz and Rönnow-Rasmussen 2004: 401)
2. ‘Imagine that you have a rich and generous but touchy friend. If he suspects you
of envying his possessions, he will curtail his largesse. That is a good reason not
to envy him . . . but surely it does not speak to whether his possessions are envi-
able. Another reason you might think it inappropriate to envy him would be
based on moral qualms about being pained at a friend’s good fortune, but this
too seems irrelevant to the ascription of the [evaluative] property [that is, being
enviable].’ (D’Arms and Jacobsen 2000b: 731)
The first type of example describes a case where we apparently have a reason for a
response that does not, however, bear on the value of X—the required response is in
some way fitting, but not in the right way. It might be fitting from a merely prudential,
moral, or aesthetic point of view, for example. The second type of example describes
a case where we have a reason to withhold a response that does, however, bear on the
value of X—a response that is fitting.
A number of (more or less unsuccessful) solutions have been proposed to solve
WKR, but I will not be concerned to explore any of these here. Instead, I want to point
out something problematic about the very way in which WKR is supposed to follow
from FA via the kinds of examples posed previously. Such examples stem from an ini-
tial stipulation about the real value of X that is appealed to in order to explain how
actual responses can fail to match the fitting responses that reveal X’s true value. As
D’Arms and Jacobsen put it, for example: ‘the trouble stems from the apparent fact that
what sentiments or attitudes one endorses feeling about something is a different ques-
tion from whether it is truly funny, enviable, regrettable, and so forth’ (2006: 23).
This way of setting up the analysis is puzzling, however, given the purported aim
of FA to analyse the value of X in terms of which attitudes are fitting. For how can the
analysis to do this job properly—that is, explicate the notion of ‘fittingness’ without
circularity—if we have already assumed the value of X? What counts as fitting in the
examples is predetermined by the value of X, whereas it should apparently be the other
way around: the value of X should be determined by what counts as fitting. Our intui-
tions about the value of X should not be governing what attitudes count as fitting, given
the aim of FA to analyse value in terms of just this notion. In short, given the apparent
aim of FA, the initial assumptions about value which drive WKR seem unwarranted.
It is thus difficult to see what is being illuminated once the value of X is no longer
being informatively analysed in terms of fittingness, especially where the value of X is
precisely what is in dispute. Note that the problem here is not the same problem as the
potentially vicious circularity inherent in the attempt to identify the relevant attitudes
92 Cain Todd
independently of the evaluative concepts that FA employ them to analyse, although

that is a grave problem (see D’Arms 2005 for discussion). Rather, I wish to emphasize
that the understanding of FA that is supposed to lead to WKR is problematic. More
specifically, it seems to undermine the ability of FA to deal plausibly with evaluative
concepts that are in dispute. Specifically, FA seems unable to account for essentially
contestable evaluative concepts.
Evaluative concepts are essentially contestable where there is room for dispute over
their application without one party to the dispute simply being guilty of conceptual
confusion about the meaning and extension of the concept. That is, disagreements
about value often appear to be legitimate disagreements not only about the extension
of some concept, but also about its constitution. People may disagree not only over
whether particular acts are wrong, courageous, or shameful, but also about what fea-
tures an act needs in order to be justifiably attributed such values.3
FA cannot, if formulated in the way that is supposed to lead to WKR, account in a
non-question-begging way for essentially contestable evaluative concepts, because in
such cases it just is the value of X that is in dispute and hence what needs to be settled.
The relevant value cannot just be presupposed before FA has accomplished its analys-
ing work. If all or most evaluative concepts turn out to be essentially contestable, the
limited scope and informativeness of such an understanding of FA will prove to be far
less attractive than its proponents would wish. For, presumably, the fortunes of any
plausible FA should not be constrained to wax or wane according to how many of our
evaluative concepts turn out to be essentially contestable.
More importantly, from our point of view, it seems that if we do not already assume
a value of X where such a value is essentially contestable, there may be any number of
emotions or attitudes that are fitting, in the sense that we can provide reasons for them.
An urgent question for the proponent of FA is thus this: if, as a matter of psychological
fact, someone were able to admire the demon in virtue of the prudential reason con-
fronting her—perhaps there could be people who are able to induce aesthetic or moral
admiration in themselves at will, or perhaps there are agents who get their aesthetic
or moral thrills from acceding to just this kind of demand—what on the FA analysis
would prohibit the value of ‘admirable’ from accruing to the demon? Nothing about
FA seems able to rule this out without either begging the question about the (poten-
tially disputed) value at the heart of the analysis, or admitting that the demon really
is admirable relative to the person for whom there are reasons for having this attitude
and who can, as a matter of psychological fact, respond in the relevant way.4 That is,
3
As D’Arms notes, ‘if either party is guilty of error, that error must be located somewhere other than
in a failure of conceptual competence or nonevaluative knowledge’ (see Wiggins 1987; D’Arms 2005: 12–13;
D’Arms and Jacobsen 2006: 16, 198).
4
Of course, if there are relevant constraints operative on what is psychologically possible this may allow
a way to formulate fittingness for certain cases. However, such cases are likely to be limited and hence of lim-
ited use in escaping the problem identified here. For relevant discussion see Persson (2004).
there is reason to admire the demon, and so in that sense the attitude of admiration
will be fitting.
Thus, one potential implication of these considerations appears to be a form of rela-
tivism about the fittingness relation: for any object X and the application to X of any
essentially contestable evaluative concept Φ, there may be any number of incompat-
ible attitudinal responses that are fitting, in the sense that there exist reasons for those
responses. Of course, avoiding this kind of relativism about fittingness just is the prob-
lem of finding independent, non-circular grounds for distinguishing the right kind
of reason, the fitting attitude, that reflects the true value of X. Unless this can be done,
however, it is difficult to see how FA can by itself avoid the menace of relativism.
The chief problem here is that deciding which reasons are the wrong or right kind is
simply a natural part of the difficulty involved in resolving certain evaluative disputes
about the applicability of such concepts in the first place. That is, the application of
essentially contestable evaluative concepts seems to be in large part, if not wholly, an
evaluative matter.5 After all, what counts as admirable, shameful, or enviable, for exam-
ple, is surely in part a question of what to admire, what to be ashamed of, what to envy.
What is genuinely enviable other than what one takes to be worthy of envy? Insofar as
‘enviable’ is a normative, evaluative concept, the question about its applicability, and
hence the fittingness of envy in some context, just does amount to deciding what one
should be envious of.
So, if one succeeds in changing one’s emotion itself or changing what one endorses
as worthy of envy, then nothing about FA prevents the purported real value of the cir-
cumstances changing accordingly. If the value of X is held to amount to whatever one
judges to be the fitting emotional response to X, then this value will be a result of what-
ever responses one endorses as fitting. The decision about what is fitting will itself be an
evaluative decision, in the sense that judgements about what one ought to feel—about
what response(s) to endorse—are a matter of weighing up reasons. Yet now the judge-
ment of fittingness, being an evaluative judgement, will be constrained not just by the
non-evaluative features of the situation, but will also depend in part on one’s (higher-
order) values and preferences. What counts as a ‘wrong kind of reason’, that is, will be
in part relative to one’s own further values, attitudes, beliefs, emotions, and interests.
This is particularly the case for emotional responses, as I shall attempt to demonstrate
in what follows.
However, at this point one might object that these conclusions are too quick. After
all, it might be argued, we cannot adequately analyse ‘fittingness’ without some initial
grasp of the value of X and we can only begin, therefore, with some stipulation about
5
How large? I think this is a very difficult question to answer and will depend partly on what sort of
account we give of the use of thick concepts in FA. This can be seen more readily in the example of the envi-
able friend, since the issue of psychological possibility is more salient in this relatively normal context than
that in which an evil demon commands us to adopt a response for no reason related to the normal eliciting
or appropriateness conditions of the response.
94 Cain Todd
this value. FA is designed to capture the intuitive sense that we ought to respond, say,
with admiration only to those things that are really admirable, whilst recognizing that
the admirable can only be fully understood with reference to the attitude of admira-
tion. Perhaps, then, FA should not be thought of as a reductive analysis of value, but
merely as a type of account that aims to shed some light on the connection between
values and attitudes; namely, that there is some kind of dependence between values
and human responses.6 More specifically, one might maintain that such analyses aim
to capture something important about the content of our evaluative concepts rather
than to illuminate the values to which such concepts supposedly correspond. A revised
formulation of FA makes this new orientation clear:
(FA*) To apply an evaluative concept Φ to an object/situation X is to think it fitting to have some
particular attitude A towards X.7
This, indeed, is the version of FA that D’Arms (2005) proposes in outlining his
‘rational sentimentalism’. The plausibility of this account lies in carefully circumscrib-
ing the types of evaluative concepts and their accompanying responses that can be gen-
uinely explained by such an analysis. The ‘natural emotions’, to which D’Arms’ account
appeals, indeed appear to possess all of the requisite characteristics required to provide
a plausible sentimentalism that avoids WKR, accommodates essential contestability,
and illuminates the fittingness connection between human responses and evaluative
judgements. In order to assess the prospects of sentimentalism, therefore, it is impor-
tant to spend time examining whether this theory can live up to its promise.
2. Rational Sentimentalism

D’Arms (2005: 2) argues that ‘certain evaluative concepts are regulative concepts for
paired emotion types: concepts whose primary function is to guide or regulate specific
kinds of emotional response by appeal to reasons of a particular sort.’ The relevant con-
cepts are those that have an especially close connection to specific human emotions,
such as ‘shameful’, ‘fearsome’, ‘enviable’, ‘disgusting’, ‘funny’, ‘pitiful’, and so on. For the
application of these concepts, the judgement that an emotion fits plays a role in ‘regu-
lating our tendencies to feel and act’. So, the point of using these evaluative concepts is
to endorse (but also to motivate) feeling and acting in appropriate ways.8
Importantly, what counts as appropriate here depends not just on features of the cir-
cumstance—for instance, we ought to be afraid only if the situation really is dangerous—
but also on the nature of the emotions themselves; in particular on their characteristic
6
See Wiggins (1987).
7
‘The central idea of contemporary sentimentalism is that to judge that one of these evaluative response-
dependent concepts applies is not to feel F, nor be disposed to feel it, but to favor feeling F, or to think there is
reason to feel it, in response to X’ (D’Arms 2005: 3).
8
To reflect the different usages found in the literature, I will from now on use ‘fitting’ and ‘appropriate’
interchangeably.
intentional objects. The ‘constraints on what reasons are relevant to whether something
is shameful, fearsome, enviable, and so on, are therefore partly determined by features
of these emotions themselves . . . It is because it is not part of the nature of envy to pre-
sent its object as undeserved that the fact that your friend deserves his success is irrel-
evant to whether it is enviable’ (D’Arms 2005: 4). This condition is, clearly, central to
the position’s success in avoiding WKR, by constraining the appropriateness conditions
governing emotional responses from, as it were, two sides: the world, and the nature of
the response.
To support this view, D’Arms invokes what he calls a ‘regulative role argument’,
which draws on the various features of emotions—for instance, their motivational
capacities, their valence, their amenability to reason and yet relative independence of
judgement—that ‘both incline us and give us reason to attempt to regulate them: that
is, to reflect upon, confer about, and develop (more or less articulate) standards con-
cerning when to have them, and to take such steps as we can to feel emotions in accord-
ance with our conclusions’ (D’Arms 2005: 6). In short, his argument claims that the
most plausible explanation of the function of these evaluative concepts is to regulate
our emotional reactions to things in terms of their fittingness towards those things.
We use these concepts precisely because our emotions play an essential role in our
practical lives, and thus we need a way of discussing, reflecting, and acting on their
appropriateness.
Furthermore, D’Arms argues that his account best accommodates the essential con-
testability of our evaluative concepts. Specifically, it does so while nonetheless preserv-
ing the univocity of our evaluative predicates. He contends that, in cases of dispute,
we can refer to a common emotional response—a response that provides a common
element in the intension of our evaluative thoughts, and which we can identify inde-
pendently of the evaluative property that is its object; thus crucially without the cir-
cularity that menaces other versions of sentimentalism. This identification, he holds,
is properly an empirical matter, and in this light he points to the class of natural emo-
tions, which are ‘heritable suites of cognitive, affective, motivational, and behavioural
changes that are part of the normal human repertoire in every culture because of our
shared evolutionary history’ (D’Arms 2005: 16).
Granting the existence and identification of such a class, and leaving aside the prob-
lematic issue of whether such (non-‘cognitively-sharpened’) emotions include those
properly paired with his list of evaluative concepts, I want to question whether D’Arms
can provide the plausible account of fittingness that he claims to, and whether as such his
rational sentimentalism can accommodate essential contestability in the way envisaged.
Recall that, in the previous section, we concluded that once we have essential con-
testability combined with the idea that the judgement of fittingness is itself an evalu-
ative judgement, FA looks prone to some sort of relativism concerning the fittingness
relation. In order for his sentimentalist picture to avoid it—and in the process to pro-
vide an illuminating notion of fittingness—D’Arms must assume that, at least for his
limited range of natural emotions, the world constrains their appropriateness tightly
96 Cain Todd
enough to give some kind of objective grounding for what counts as the appropriate
response. In other words, the assumption must be that emotions possess something
like objective fittingness conditions. One major piece of support for this assumption is
an appeal to emotional phenomenology:
It’s in the nature of these experiences to present themselves as sensitivities to something outside
them. And what they present themselves as sensitivities to is a fairly restricted feature of the situ-
ation: a socially significant personal inadequacy, or a threat to one’s safety, for instance. A little
introspection makes it obvious, I think, that feelings of shame, fear, and so on just aren’t about
the advisability, or the moral permissibility, of feeling precisely that way. They are about a feature
of the circumstance in virtue of which this is a fitting way to respond. In fearing the wild animal,
I am struck by the things that make it fearsome (its size and ferocity, say), not by things that
might make it wise or virtuous to be afraid . . . Assessments of fittingness are attempts to make
sense of or criticize our emotions using standards that speak to the distinctive concerns we take
them to embody. It is therefore important to have a vocabulary that expresses such assessments,
in particular, as a vehicle for rational interpretation of ourselves and one another. (D’Arms 2005:
10–11)
So, the fact that any specific emotion presents itself to us as being sensitive to some fea-
ture of the circumstance that thereby justifies that particular response explains why we
have a developed vocabulary of evaluative concepts that express judgements of fitting-
ness. This justifying feature (or features) must answer to the specific concerns intrinsic
to the emotion and not to extraneous concerns that are merely prudential, or moral,
for example. In other words, this is the familiar idea that emotions have ‘core rela-
tional themes’: they are identified and individuated by appeal to the intentional objects
that are intrinsically connected to them—fear is only made genuinely fitting by being a
response to danger, anger to offence, sadness to loss, and so on. And these intentional
objects are constituted by or supervene in some way on certain non-evaluative fea-
tures; for instance, the lion’s big teeth [fear], my nemesis’s remark [anger], the death of
my goldfish [sadness].9 Hence, the fittingness relation has an objective grounding in
the connection between emotions and the objects at which they are properly (by their
very nature) directed.
D’Arms’ sentimentalist framework rests in large part on the claim that this fitting
responsiveness of emotions is evident in their phenomenology, and the view of evalu-
ative concepts he proposes is seductive to the extent that this is indeed how our emo-
tional phenomenology often appears to us. Unfortunately, however, the observations
used to support such a view, outlined in the previous quote, are insufficient to establish
this claim. It is not controversial that fear is made fitting by the dangerous, or anger
made appropriate by the offensive, but the appeal to these trivially true constraints on
emotions in terms of their intentional objects does nothing to illuminate the fittingness
relation. Nor, however, and more importantly, do non-evaluative features form simple
9
See Prinz (2004) and Deonna and Teroni (2012) for discussion of core-relational themes.
neutral, objective constraints on emotional responses in the way D’Arms assumes and
which his view requires.10 This is because it is the way in which the non-evaluative fea-
tures of a situation are evaluatively construed by a subject that partly underpins the
judgement of the appropriateness of that response, as we are about to see.
The general problem here is that the view of emotional phenomenology on which
D’Arms’ general picture of emotional appropriateness rests is misguided.11 Whether
X is, say, presented as dangerous or offensive, admirable or enviable, in our emotional
experience will be a function not merely of X’s objective non-evaluative features, but
of how we construe those non-evaluative features in evaluative terms, and also how
we assess the fittingness of our responses to them. Emotional appropriateness in turn,
I contend, is partly a function of the extent to which this construing figures (or fails
to figure) in our emotional phenomenology. As such, in emotional experience eval-
uative features are not necessarily, and not always, presented as objective properties
of the world independent of those experiences.12 In order to demonstrate this, let me
sketch what I take to be a more plausible picture of emotional appropriateness and
phenomenology.
3. Emotional Phenomenology and Apparent

Objectivity
It is uncontroversial that emotions are in some way dependent upon what I shall call
for the sake of convenience, ‘Subjective Evaluative Conditions’ (SEC). SEC comprise
that compilation of a particular subject’s (or of some particular group of subjects’) vari-
ous motivations, interests, beliefs, goals, cares, values, character traits, imaginative and
attentive capacities, other psychological characteristics, and even physical constitu-
tion, that forms the background conditions of a subject’s emotional constitution and
dispositions. One could say, in other words, that the evaluative states of affairs that are
the emotions’ intentional objects supervene not just on objective, non-evaluative fea-
tures of the world but also on a subject’s particular SEC.13
So, my sadness at the death of my goldfish or my team’s loss depends on my having
cared for it (that is, my goldfish’s life, my team’s success;); my anger at my boss’s offensive
remark (indeed, perhaps even my construing it as offensive in the first place) depends,
for instance, on my paranoia or sense of fairness, and/or perhaps on my sensitivity to
and interest in preserving my dignity or reputation; my fear of the dog may depend on
10
Indeed, this is to some extent evident from D’Arms own stipulation that, as noted previously, ‘con-
straints on what reasons are relevant to whether something is shameful, fearsome, enviable, and so on, are
therefore partly determined by features of these emotions themselves’ (D’Arms 2005: 4).
11
A view, it might be added, that he arguably shares with the majority of emotion theorists.
12
In this respect my account also departs from Montague’s view outlined in her chapter in this volume. In
most other respects, however, I am sympathetic to her picture of emotional phenomenology as sui generis.
13
For further discussion and similar views see e.g. Goldie (2000); Helm (2000); Roberts (2003).
98 Cain Todd
the (potentially false) belief, or even on a vivid imagining, that it is dangerous; my guilt
at having lied depends on my valuing the moral principle ‘never tell lies’, and so on.
These observations should seem relatively obvious, yet they pose a grave problem
for the quite commonplace view of emotions as straightforwardly representing eval-
uative features and in virtue of these possessing objective fittingness conditions.14 If
the intentional objects that emotions purport to represent are thought of as evalua-
tive properties, such properties would have to supervene not merely on some set of
non-evaluative features in the world—such as ‘danger’ supervening on the dog’s sharp
teeth—but also on the relevant SEC. The dog may well have sharp teeth and be snarl-
ing in a threatening manner, but: if I know how to calm even very angry dogs; if I am
unmoved by the prospect of, or maybe even desire physical harm; if I am somehow
physically immune to the pain of dog bites; or if I manage to somehow reconstrue
the threatening manner as non-dangerous—perhaps I imagine vividly that the dog is
more afraid of me, or perhaps I force myself to attend not to the sharp teeth but to the
wagging tail—I will not fear the dog even if you do.
Moreover, if this is right, then given potential differing SEC between subjects (or
some wider socio-cultural group), two very different, even contradictory emotional
responses might be equally fitting in virtue of the very same non-evaluative features.
My goldfish’s death is not a loss for you and hence not something you will or ought to
feel sad about. Perhaps you even find the death of goldfish amusing. Which is the eval-
uative property really represented? Is the dog really fearsome if you, but not I, ought
(given our respective SEC) to fear it? Even if sadness is always properly directed at
‘loss’, or fear at the ‘dangerous’, this tells us nothing in itself about what features of the
world will or ought to be construed as a loss or a danger. This is particularly evident
given that it is not merely one’s actual emotional responses, but also what one takes to
be the appropriate responses, that are relative to SEC.
Keeping in mind the conclusions drawn in the first section of this chapter, the failure
of FA to elucidate the ‘ought’ in virtue of which we are supposed to recognize which
responses are the right ones—especially problematic in cases of evaluative disagree-
ment—coupled with emotions’ dependence on SEC, militates against conceiving of
emotional fittingness as objective and in favour of holding it to be relativistic.15
Nonetheless, in order to make sense of our ordinary discourse about emotions we
must continue to give some kind of account of the appearance of emotional error, and
an explanation of why, in common-sense thought and talk, we supposedly take emo-
tions to be subject to something like objective conditions of fittingness.16 But, given the
picture I have just outlined, how can we account for this?
14
See Deonna and Teroni (2012) for further discussion.
15
Although she talks in terms of ‘emotional circularity’, for an interesting view that seems to be largely
compatible with the points I am making here, see the chapter by Zagzebski in this volume.
16
It might be objected that strategy that follows leads to an error theory of our understanding of our
emotional lives. However, on the one hand, I do not find it obvious that, in our everyday experiences of the
vast array of emotional responses and their eliciting scenarios, we think of all of our emotional responses as
First it should be acknowledged that, given our own particular SEC, we may be in
error about how the non-evaluative facts presented to us will impact upon us emo-
tionally. For instance, it might be the case that the dog I am confronted with is, despite
by mistaken beliefs to the contrary, a danger to me. Relative to my own SEC, therefore,
the dog is in this sense ‘really’ dangerous and so I ‘ought’ to fear it. I may not feel any
grief at the death of a loved one, and conclude that I ought not feel sad about it; yet a
resurfacing of this grief at a later period, or its manifestation in other behaviour may
indicate that in fact it really was a loss for me and I may come to acknowledge that
I ought to feel sad. Note that in such cases, however, what I take to be the ‘real’ evalua-
tive property presented, and what I think ‘ought’ to be the case, are inevitably relative
to SEC.
Second, we should deny that emotional phenomenology is quite as straightfor-
ward as D’Arms’ picture assumes, and instead provide an alternative explanation
of objective fittingness in terms of what I shall call the appearance of objectivity in
our emotional experiences.17 The claim I wish to defend is that although emotions
do not have objective fittingness conditions, it can appear to us that they do, where
this appearance of objectivity is subject to degree and best thought of in terms of
a spectrum, ranging from the least objective-seeming to most objective-seeming
responses. How objective-seeming a particular emotional response appears to be is
a function of a complex range of various ‘SEC-sensitive’ factors, themselves subject
to degree.
Briefly, the idea is that the further towards the more objective-seeming end of the
spectrum a particular response is, the more we will be inclined to think of its fitting-
ness conditions as independent of SEC; more specifically, the less aware we will be of
the role played by SEC in our emotional response. Conversely, the further towards
the opposite (that is, ‘relative-seeming’) end of the spectrum a response lies, the more
inclined we will be to think of its fittingness conditions as dependent upon SEC and
hence relative in nature. Here, the role of SEC in our response will be more apparent
to us.
What is it for objectivity and the relevant SEC-sensitive factors to be ‘apparent’? This
is a difficult issue that I cannot do full justice to here, but which I hope will become
clearer as the discussion proceeds. The notion of apparent is to be understood in a very
encompassing way, as referring to whatever is present in the overall phenomenology
of the emotion state, encompassing elements that one is experientially aware of in the
intentional, representational content of the response, and in the phenomenal charac-
ter of the response, as well as elements that may only be evident on reflection, includ-
ing cognitive awareness of the nature of the relationship between response and object.
possessing objective truth conditions. On the other hand, insofar as we do take our responses to be objective,
I think the explanation I offer gives a more accurate picture of supposed common sense.
17
For insightful discussions of emotional phenomenology that, it seems, are not incompatible with what
I say here, see Gunther (2004) and Montague (2009).
100 Cain Todd
One of the primary SEC-sensitive factors determining where on the apparent objec-
tivity spectrum any particular emotional response will be located is the ‘apparent var-
iability’ of the response. Two main factors underpin the apparent variability of our
responses. First, how constrained by the relevant non-evaluative features of the object
or state of affairs our emotional responses and evaluations seem to us to be. Second,
how variable the emotional responses themselves appear to be, between individuals or
groups of individuals.
Articulating a general account of how flexible, how apparently variable our emo-
tional responses will seem to us to be is extremely difficult. For it will have to appeal
to as many different factors as comprise SEC and will be as complex as specifying pre-
cisely the nature of the constraints governing the myriad ways in which non-evalua-
tive features of the world can be evaluatively construed. Apparent variability will thus
be a function of a number of complex factors including: certain general psycho-physi-
cal constraints on emotional responses; certain socio-cultural evaluative conventions,
norms, and standards; and individual physiological and psychological capacities,
including for example how reflective, attentive, and imaginative we happen to be.
Thus, what seems an objectively valid emotional response to one person may seem
highly relative to another. Determining whether an emotional construal is, to use a
nice phrase of Robert Roberts’, ‘in our emotional repertoire’ is a complicated matter
depending on any number of factors encompassed by SEC. Indeed, in order to flesh
out this picture we can compare it to Robert Roberts’ (2003) theory of emotions as
‘concern-based construals’. His notion of concern is closely related to my rather
broader notion of SEC, and the nature of construals suggests the amenability of emo-
tions to some kind of control, if not in terms of directly changing one’s emotion states
themselves, then indirectly by changing the way in which the relevant object or situa-
tion is construed.
Construals are articulated in terms of the familiar phenomenon of aspect-perception
or ‘seeing-as’. They are essentially experiential states—and hence not reducible to mere
beliefs or judgements—involving a way that things appear to the subject, and because
they involve a relationship in which one thing is seen in terms of something else, they
are interpretive or constructive in a way that mere sense perceptions are not. They are,
as Roberts acknowledges, a ‘hard-to-specify structure of percept, concept, image, and
thought’ (Roberts 2003: 77).
In this light it is worth quoting what Roberts has to say about the issue of volun-
tary control, which sometimes emotions are subject to and sometimes not; which is
a question of degree; which involves the subject’s desires and concerns; and which
involves changing patterns of attention regarding the organization of the features
of the relevant object or situation. As such, there may be certain constraints, nat-
ural or conventional, on how any given object or state of affairs can be construed
emotionally.
A person at whom I am inclined to be angry may be regarded, quite at will, in various ways: as
the scoundrel who did such-and-such to me, as the son of my dear friend so-and-so, as a
person who, after all, has had a pretty rough time of it in life, and so forth. If these construals
are all in my repertoire, and in addition are not too implausible with respect to the present
object [e.g. seeing the young-old woman as an odd-shaped pizza], then the emotions that
correspond to them, of anger, affection, and pity, are also more or less subject to my will . . . In
some situations an emotion may be so compelling that we are . . . virtually helpless in the face
of it. The therapist or friend, by suggesting and fostering other possibilities of construal, may
be able to liberate us from it by contributing to our emotional repertoire. Or she may not.
(Roberts 2003: 81)
These observations strike me as absolutely right and suggest again that, given the
relativity and relative flexibility to which construals are subject, it is more accurate
to think of the conditions grounding the fittingness of emotions in relative rather
than objective terms.18 Emotions can be as amenable or impervious to direct vol-
untary control, and apparent voluntary control, as ways of construing situations
are. Indeed, these two factors, construal and response, will generally be inextricably
intertwined.19
We are now, I think, in a position to offer some general and inevitably rather
speculative observations about the conditions governing apparent variability. For
certain emotion-types and/or certain episodic emotional responses, SEC play, and
may seem to us to play, a minimal determining role. As a matter of contingent fact,
certain pan-cultural, hardwired emotions—D’Arms’ class of natural emotions—
when elicited by environmental factors that are closely (perhaps phylogenetically)
connected to them, may seem to be not very amenable to any type of voluntary con-
trol, and to admit of little or no variation in terms of the relevant object‒response
relation. We may thus be inclined to think of such cases as involving either some-
thing like response-dependent properties or straightforward objective values and
hence as having something like simple objective fittingness conditions. The rele-
vant response group or value will then be relatively clear and easily identified. This
kind of view, apparently, is what underpins D’Arms reliance on the class of natural
emotions.
For example, construing the death of loved ones as a loss to which sadness is the
warranted response may be so hardwired and universal, so immune to our ability to
control it, that any failure to feel sadness, or any tendency to have another emotional
response, will normally be explained in a way consistent with response-dependent
accounts of evaluative properties, such as in terms of certain abnormal conditions.
Abnormal conditions in such cases will just be any conditions that are appealed to in
18
Although Roberts himself seems to hold onto the notion of objective truth conditions (2003: 147).
19
We might also add that what counts as apparently voluntary will depend in part either on how flexibly
one can construe the relevant features of the object or situation, or on how much control one actually has
directly over one’s own emotional states. And there are naturally cases where even your beliefs about the
situation are unable entirely to change your actual emotional feelings. This is what happens in cases of irra-
tional and phobic emotional responses.
102 Cain Todd
order to explain departures from the norm, and the norm just will be taken to be some-
thing like ‘all’ or ‘most’ human beings.
At the other end of the spectrum, comprising the least objective-seeming, most
relative-seeming responses, it is nigh on impossible to give any equivalent gen-
eral statement. For here we can expect to find culturally or individually variable
emotional responses shaped and bound up with complex socio-cultural contexts,
and with all the variations to which individual emotional reactions can be subject.
In other words, we can expect to find just those emotions (types and tokens) that
will be more prone to seem to us to be relative to SEC in terms of their apparent
variability.
Inevitably the picture just sketched is far from simple and allows of many pos-
sible variations and complications. At the objective-seeming end, for example,
there may well be room for SEC to play, and appear to some subjects to play some
role in their responses, and there may be some variability across groups of sub-
jects in their emotional responses to the relevant states of affairs—a variability
that may or may not be apparent. Similarly, some of the emotion‒object relations
at the relative-seeming end may seem to some subjects to be hardwired and not
amenable to voluntary control or variation, depending perhaps on which values
are at issue, and on various facts about our particular upbringing and psychology,
and so on.
It is crucial to stress, however, that even on this multifaceted picture there is room
for emotional error. There is no guarantee that what appears to us to be the case
regarding the objectivity of the fittingness conditions of emotional responses really
is the case. It might seem to me that I cannot help but think of my team’s loss as
anything but sad, but I may be mistaken about my own psychological capacity to
change my emotional outlook or how I construe the situation. Moreover, I may be
mistaken about my own values and interests, and hence to that extent mistaken,
in any given instance, about what I would take to be the appropriate emotional
response.
In sum, the relative fittingness of emotions will be more or less reflected in emotional
phenomenology, but no matter how apparently objective our emotional responses
appear to us to be, this does not suffice to grant emotions as such objective conditions
of fittingness.
Conclusion
What conclusions should we draw from this for the viability of D’Arms’ rational senti-
mentalism, and more generally for fitting attitude analyses of evaluative concepts that
appeal to the emotions? In some respects, it leaves D’Arms’ own account intact. For
certain emotional responses at the objective-seeming end of the apparent objectivity
spectrum—where we can expect to find the natural emotions—it may well be possible
to give a plausible rational sentimentalist picture of the regulative role of the relevant
evaluative concepts, as well as a coherent account of essentially contestable evaluative
disputes that avoids WKR and preserves the univocity of the relevant predicates. At the
level of phenomenology, that is, concerning how we use our concepts, how we reflect
on our responses, and how we think of the disputes in which we engage, rational sen-
timentalism may well offer some insight into the apparently tight connection between
certain concept‒response pairings.
Yet the overall picture on which rational sentimentalism rests—of the objectivity
of emotions’ fittingness conditions and their objective phenomenology—is, I have
argued, inaccurate. To that extent, a sentimentalist position centring on the emotions
cannot escape the kind of relativism about fittingness that I have outlined, however
objective particular instances of concept‒response pairs might strike us on reflection.
Furthermore, it should be clear from my account just why an emotion-centred sen-
timentalism about evaluative concepts leads to ‘fittingness-relativism’ and why WKR
and essential contestability are just reflections of the various complex considerations
that may be brought to bear on emotional fittingness. Indeed I think my account offers
a more plausible explanation of essential contestability and of the variegated nature of
evaluative disputes.
Consider, for example, an evaluative dispute between subjects whose emotional
phenomenology is located at the relative-seeming end of the apparent objectivity
spectrum. In such cases, arguably, the very essential contestability of the predicate at
issue will be more or less apparent to both disputants, leading them to reject objec-
tivity about the domain in question. But then this might well be enough to end the
disagreement, to lead them to acknowledge that there is no genuine disagreement after
all. Alternatively, it might reorient the disagreement, leading the disputants to focus,
at a higher level of reflection for instance, on the commensurable and incommensu-
rable merits of different types of emotional responses, their evaluative sources, and
the values that are their objects. In any case, given the complex and degree-dependent
structure of the spectrum of apparent objectivity, we should expect that such disputes
themselves will be subject to enormous variation, and this I suspect is what we would
find if we were to examine the empirical phenomena more closely.20
20
This chapter has been through many incarnations, and there are many people to thank for various parts
of it. The initial research for this chapter was carried out while on research leave at the Swiss Centre for
Affective Sciences (CISA) at the University of Geneva, with support from the Thumos Research Group on
emotion there. I would like to thank CISA and Thumos for the welcoming and stimulating research environ-
ment they provided, and for the generous funding they awarded me while on leave. In particular, I would
like to thank Fabrice Teroni, Julien Deonna, Olivier Massin, Kevin Mulligan, and other members of Thumos
for their feedback on earlier versions of this chapter. The latter stages of research for this chapter were
funded by a generous grant awarded by the Swiss National Science Foundation for the project ‘Imagination,
Emotion, and Value’ at the University of Fribourg. I would like to thank all the members of the department
there, particularly Gianfranco Soldati and Martine Nida-Rümelin, for insightful discussions on many of
the issues discussed here. Finally, I would like to thank the audience at The Hague where I presented part of
this chapter, and especially the specific and extremely helpful comments I received later from Peter Goldie,
Simon Blackburn, Michael Brady, and Sabine Roeser.
104 Cain Todd
References
D’Arms, J. (2005). ‘Two Arguments for Sentimentalism’. Philosophical Issues 15: 1–21.
D’Arms, J. and Jacobson. D. (2000a). ‘The Moralistic Fallacy: On the “Appropriateness” of
Emotions’. Philosophy and Phenomenological Research 61: 65–90.
D’Arms, J. and Jacobson, D. (2000b). ‘Sentiment and Value’. Ethics 110: 722–48.
D’Arms, J. Jacobson, D. (2006). ‘Sensibility Theory and Projectivism’. In The Oxford Handbook of
Ethical Theory, ed. D. Copp (Oxford: Oxford University Press), 186‒218.
Deonna, J. and Teroni, F. (2012). The Emotions: A Philosophical Introduction (London: Routledge).
Goldie, P. (2000). The Emotions: A Philosophical Exploration (Oxford: Oxford University Press).
Gunther, Y. (2004). ‘The Phenomenology and Intentionality of Emotion’. Philosophical Studies
117: 43–55.
Helm, B. (2000). ‘Emotional Reason: How to Deliberate about Value’. American Philosophical
Quarterly 37: 1–22.
Lang, G. (2008). ‘The Right Solution to the Wrong Kind of Reason Problem’. Utilitas 20: 472–89.
Montague, M. (2009). ‘The Logic, Intentionality, and Phenomenology of Emotion’. Philosophical
Studies 145: 171–92.
Persson, I. (2004). ‘Primary and Secondary Reasons’. Ethics 114: 391–423.
Press).
Rabinowicz, W. and Rönnow-Rasmussen, T. (2004). ‘The Strike of the Demon: On Fitting
Pro-Attitudes and Value’. Ethics 114: 391–423.
Reisner, A. (2009). ‘Abandoning the Buck Passing Analysis of Final Value’. Ethical Theory and
Moral Practice 12: 379–95.
University Press).
Wiggins, D. (1987). ‘A Sensible Subjectivism?’ In Needs, Values, Truth: Essays in the Philosophy of
Value (Oxford: Blackwell), 185‒214.
PA RT I I
Emotion, Evaluation, and
Justification
7
Emotion, Evaluative Perception,
and Epistemic Justification
Adam C. Pelser
Emotions often give rise to beliefs. The tendency of emotional experience to issue in
belief, even if it does not always do so, reveals a human proclivity to trust emotions, at
least implicitly, as justifying evidence or reasons for belief.1 As psychologists Gerald
Clore and Karen Gaspar explain, “beliefs are adjusted to be compatible with internal
evidence in the form of feelings, just as they are adjusted to be compatible with external
evidence from perceptual experience . . . Evidence from the sensations of feeling may
be treated like sensory evidence from the external environment” (2000: 25). This is
just one of many similarities between emotion and sense perception that have led a
growing number of philosophers and psychologists to defend a family of perceptual
theories of emotion, according to which emotion is a kind of perception (see, e.g., de
Sousa 1987; Frijda and Mesquita 2000; Nussbaum 2001; Roberts 2003; Zagzebski 2003;
Prinz 2004).
The fact that people so regularly trust emotions by forming beliefs on the basis of
emotional experience raises the question of whether such trust is ever epistemically
justified. Put slightly differently, this phenomenon raises the question of whether
emotion, which is often a source of belief, is ever a source of justified belief (or epis-
temic justification). One widely held view (see, e.g., Goldie 2004; Brady 2012) is that
unlike sense perceptual experiences emotions are not capable of justifying beliefs
except perhaps in the uninteresting way that they might provide justification for
beliefs an agent forms about her emotional experiences (for example, that she is
angry, or that grief feels like this). Michael Brady (2012: 147–8) takes this apparent
dissimilarity between emotions and sense perceptions to be evidence against per-
ceptual theories of emotion.
Linda Zagzebski offers an insightful discussion of trust in emotions in her chapter in this volume.
1
108 Adam C. Pelser
Against this popular view I argue here that emotions can justify the beliefs to which
they give rise. Put a bit differently, emotion is a basic source of epistemic justification.2
Call this the justificatory thesis of emotion (or the justificatory thesis for short). The jus-
tificatory thesis is not the claim that an agent can form a justified belief on the basis
of an inference from her awareness of her emotional states, but rather the claim that
emotions themselves can confer justification on beliefs formed non-inferentially out
of (or on the basis of) emotional experience, as, for example, sense perceptions can
confer justification on sense perceptual beliefs. The justificatory thesis has been sug-
gested and supported in limited ways in recent literature (see, e.g., Zagzebski 2003;
Döring 2003; Wedgewood 2007: 225–47; Elgin 2008; Roberts 2010; Roeser 2011), but
no focused, sustained defense of the thesis has yet been offered.3 Such is the task of this
chapter.
1. The Perceptual Character of Emotion and the

Epistemology of Justification
I shall begin by clarifying the justificatory thesis in light of the perceptual character of
emotion and situating it with respect to competing views on the epistemology of justi-
fication. Perceptual theorists of emotion have identified several significant structural
and functional parallels between emotions and sense perceptions, including the fact that
emotions, like sense perceptions, are intentional mental states that represent their objects
to us in organized, conceptually rich, propositionally structured, and often belief-induc-
ing ways. As Martha Nussbaum observes: “Emotions are not about their objects merely
in the sense of being pointed at them and let go, the way an arrow is released toward
its target. Their aboutness is more internal, and embodies a way of seeing” (2001: 27).
Emotions, like sense perceptions, have representational (propositional) content. That
they do is evidenced by their phenomenology (in fear the object of our fear really seems
dangerous) as well as by the observations that emotions give rise to beliefs and that chang-
ing our beliefs about the object of an emotion can result in a change of emotion.
It might be argued that some emotions, such as depression, general anxiety, objectless
delight, and so on, do not seem to take objects or be about anything at all. These mental
states are indeed similar to emotions in some ways and often predispose their subjects
to certain emotional experiences. They are, nevertheless, better understood as moods
than as emotions, since they are more akin structurally and phenomenologically to
such non-episodic, non-intentional states as cantankerousness, cheerfulness, and light-
heartedness than they are to paradigmatic emotions such as anger, sorrow, and joy.
2
See Sosa (2007: 106) for a helpful discussion of the nature of epistemic sources.
3
After I wrote the initial draft of this chapter, Linda Zagzebski (2012: ch. 4) and Robert Roberts (2013: ch.
3) published complementary defenses of the justificatory thesis of emotions.
Emotion, Evaluative Perception, and Epistemic Justification 109
I shall accordingly reserve my use of the term ‘emotion’ for specific mental episodes that
take objects.4
In development of a perceptual analysis of emotion, Robert Roberts (2003) has
argued that emotions are “concern-based construals” of their objects that are akin to
visual construals of gestalt figures like the famous duck–rabbit. That it is possible to
switch between construing the duck–rabbit as a duck and as a rabbit without a change
in sensation (the mere visual appearance of the lines does not change) and without
the generation of any new beliefs reveals that there is another mental state operative
in perception besides belief (or judgment) and sensation. This is the mental state
that I, following Roberts, call construal (Roberts 2003: 69–83; Pelser 2010: 369–71).
As Roberts observes, emotions essentially involve a distinctively perceptual kind of
construal, though the construals involved in emotion are not essentially caused by sen-
sory experience the way sense perceptual construals are. He explains: “The person who
feels triumphant is not merely judging that he is triumphant, but is ‘perceiving’ himself
as such, yet without this experience being constituted of any sensory experience of
himself ” (2003: 67).5
In addition to being non-sensory in the way described, emotions are distinct from
paradigmatic sense perceptions in that they are essentially evaluative perceptions of
their objects. Whereas sense perception involves the construal (conceptualization)
of objects in terms of their visible, audible, tactile, olfactory, and gustatory proper-
ties, emotional perception involves construal of objects in terms of their value or dis-
value, be it moral, aesthetic, and so on. Moreover, emotions are typically experiences
of particular value-laden objects or states of affairs (for instance, that act of injustice),
as opposed to general evaluative propositions (for example, general moral rules, such
as that murder is always unjust). As Linda Zagzebski (2003) has argued, emotions
are also typically perceptions or construals of their objects as possessing some ‘thick’
evaluative property or another, such as injustice or sublimity, as opposed to ‘thin-
ner’ properties like rightness or wrongness. To experience gratitude, for example, is
to construe someone as having generously bestowed some underserved benefit upon
one. Likewise, to be afraid is to construe something in one’s environment as a threat
(that is, as dangerous) to one’s safety or to the safety of something (or someone) about
which (whom) one cares. Therefore, given the plausible assumption that emotions,
like sense perceptions, will typically give rise to beliefs that share their propositional
contents, the beliefs that emotions will be capable of justifying (if they are capable
4
Roberts (2003: 64) and Lamb (1987: 107–9) both argue for the emotion/mood distinction along the
lines of intentionality. Lamb is willing to allow that some objectless moods can also be emotions, but this
seems arbitrary. Roberts is more consistent in treating all such moods as distinct from emotions due to their
non-episodic, non-intentional character.
5
Whiting (2012) argues that the fact that emotions are not tied to a particular sensory faculty (“per-
ceptual modality”) undermines perceptual theories of emotions. I disagree, though I do not have space to
respond here.
110 Adam C. Pelser
of justifying beliefs at all) will typically be ‘thick’ evaluative beliefs. So, for example, an
instance of fear might justify the following belief: That object is a threat (is dangerous)
to me or mine. Likewise, gratitude might justify the belief: She has generously bestowed
an undeserved benefit on me and is worthy of thanks. In each case, the (potentially) justi-
fied belief is simply an assent to the evaluative propositional content contained in the
emotion.6
In defending the thesis that emotions can justify the beliefs to which they give rise,
I am assuming that they also can fail to do so, as common experience reveals that they
often do. The kind of perceptual theory I am interested in defending is one that takes
emotions to be accurate or inaccurate, apt or inapt responses to values that exist in
the world independently of our emotional responses to them. It is thus important to
distinguish my view from sentimentalist views such as that defended by Jesse Prinz
(2004, 2007), which is also a kind of perceptual theory of emotion. For Prinz, emo-
tions are not direct perceptions of objective values, but rather perceptions of bodily
changes that reveal what we value. As such, emotions project subjective, response-
dependent values onto the world. According to Prinz, the only way that an agent’s
emotions can ‘misperceive’ value is by being inconsistent with the agent’s own set-
tled values as determined by her “sentiments” or long-term emotion-dispositions
(2007: 84, 104). On his view the justificatory thesis comes cheap because the evalu-
ative judgments to which emotions give rise are not beliefs about values that exist
independently of their being ‘perceived’ by emotions, but are rather about “response-
dependent properties” that are constructed by the agent’s sentiments themselves.
Prinz acknowledges as much: “If I make a judgment that something is wrong, and that
judgment is made under epistemic conditions in which I have accurately assessed my
long-term memory and discovered a sentiment of disapprobation toward that thing,
then my judgment is warranted because wrong refers to that toward which I have
such a sentiment . . . Warrant is cheap if constructive sentimentalism is true” (2007:
236). By contrast with Prinz’s constructive sentimentalism, the perceptual theory of
emotion under consideration here, informed as it is by axiological realism, does not
trivially imply the justificatory thesis.
Having situated the justificatory thesis with respect to a perceptual theory of emo-
tion, it is now important to clarify the relevant concept of justification. The justifica-
tion in view in the justificatory thesis is, first and foremost, epistemic justification, as
opposed to pragmatic or moral justification. Epistemic justification here refers to a
normative property enjoyed by agents with respect to their beliefs and, perhaps, with
respect to other epistemic attitudes (such as disbelief, withholding, and so on). It has
been argued that there is no single concept of justification shared by epistemologists
(see, e.g., Alston 2005). Nevertheless, in keeping with one popular conception of justifi-
cation I shall assume that justification is necessary for knowledge, though not sufficient
6
Here and throughout I rely on Roberts’ (2003) insightful and impressively thorough discussion of the
propositional contents characteristic of various emotion types.
to render a true belief an instance of knowledge, and that justification involves believ-
ing on the basis of good reasons or evidence. Two points of qualification are in order
here. First, a belief can be based on good reasons or evidence even if one does not
form the belief on the basis of an inference from one’s reasons or evidence. This is
important for our purposes because the justificatory thesis is concerned with cases
in which emotions give rise to justified beliefs non-inferentially. Second, the claim
that justification involves believing on the basis of good reasons does not entail that
the agent be aware (potentially or actually) of the reasons for her beliefs, as argued
by proponents of what has been called variously awareness internalism, accessibi-
lism (see Conee and Feldman 2004), or simply internalism (see Bergmann 2006). As
Michael Bergmann (2006: 3–4) argues, something like the minimal conception of
justification outlined here seems to be shared by at least some internalists and exter-
nalists alike.
Though it is arguably neutral with respect to internalism and externalism, the jus-
tificatory thesis, as stated previously, is not neutral with respect to foundationalist
and coherentist theories of justification. Coherentists will object to the idea that epis-
temic justification comes from belief-producing faculties (sources of belief), claim-
ing rather that such justification is grounded in the coherence a belief enjoys with
others of the agent’s beliefs. In conjunction with the perceptual theory of emotions
sketched previously, however, the justificatory thesis could perhaps be modified to fit
coherentist theories that allow a justificatory role for perceptual experiences in addi-
tion to beliefs (cf., Kvanvig and Riggs 1992). Therefore, while I will defend the thesis
that emotion is a basic (foundational) source of justification, much of what I say can
be taken as support for the related claim that emotions can justify beliefs in roughly
the same way sense perceptual experiences do—an insight that is compatible with at
least some versions of coherentism.
In contrast to the epistemic justification described here I shall follow Patricia
Greenspan (1988) in reserving the term emotional justification for the justification, if
any, that emotions themselves enjoy (more on this in section 4). Having clarified the
justificatory thesis in light of the perceptual character of emotion, and having situated
it with respect to competing views on the epistemology of justification, we are now in a
position to consider the best arguments for and against the thesis.
2. Emotion as a Source of Justified

Evaluative Beliefs
Perhaps the strongest argument in favor of the justificatory thesis is that it plausibly
explains the justification we enjoy with respect to those of our evaluative beliefs that we
seem to form on the basis of emotional experience. We justifiedly believe (and many
of us know), for example, that slavery is unjust, that sunsets are beautiful, that integrity
is admirable, that children are precious, that the Jewish Holocaust of the 1940s and the
112 Adam C. Pelser
Rwandan genocide of the 1990s were abominable, and so on. Whence the justifica-
tion for these beliefs? One plausible answer is that the justification for at least some
evaluative beliefs comes from direct experiences of the value properties instantiated
in the states of affairs the beliefs are about. If there are such experiences, it is plausible
that they can function as epistemic justifiers in precisely the way that sense perceptual
experiences do. As Graham Oddie argues,
the visual experience of a bright red rose—that is to say, the rose’s appearing bright red to me—
gives me a reason to believe that the rose really is bright red . . . If there are genuine experiences
of value, they could stand to values as ordinary perceptual experiences stand to the objects of
perceptual experience. An experience of the goodness of P, say, would be the state of P’s seeming
(appearing, presenting itself as) good, where this seeming is an experiential, non-doxastic take on
the value of P. If there is such a state as the experience of the goodness of P, then, by analogy with
the perceptual case, it would give me a reason to believe that P is good. (2005: 40)
In defense of his “evaluative outlook conception of desire,” Talbot Brewer argues that
we do have the kind of direct experiences of value Oddie envisions and that, far from
being mere ‘thin’ pro or con (thumbs-up or thumbs-down) attitudes, our experiences
of value employ rich and varied evaluative concepts: “We welcome things, and are
pleased by things, under the guise of the just, the kind, the funny, the elegant, the philo-
sophically illuminating, the ironic, the loyal, the friendly, the dignified, the gracious,
the human . . . the list could be extended at length” (2009: 140–1). Oddie’s and Brewer’s
observations are, so far, consistent with viewing emotions as belief-justifying percep-
tual experiences of value. Yet they both posit desires, not emotions, as the most plausi-
ble candidates for our experiences of value.
However, there is good reason to think that desires, even if they involve perceptions
of value, are not the best candidates for at least some of our evaluative experiences
and corresponding justified beliefs. Indeed, our capacity for emotion seems to enable
experiences of certain ‘thick’ evaluative properties that we cannot directly experience
otherwise. To experience something as sublime, for example, just is to be awed by it; to
experience something as funny just is to be amused by it; to experience something as
unjust just is to be indignant toward it; to experience someone as generous just is to be
grateful (that is, to feel gratitude) toward her, and so on.
Of course, one might judge that an object instantiates one of these properties without
the aid of occurrent emotions, and one might desire something or someone because
she possesses (or, ‘under the aspect of ’) one of these properties, but it does not follow
that one can directly experience (construe) objects as possessing these thick evaluative
properties without the aid of occurrent emotions. While I might desire to hear a funny
joke or believe that a joke is funny without feeling amusement, until I am amused by
the joke I have not yet directly experienced the funniness of the joke. Likewise, while I
might believe that slavery is unjust and I might be averse to slavery because of my belief
that it is unjust, until I have been angered by or felt indignation toward an instance of
slavery, I have not directly ‘seen’ the injustice of it. As evidence of the emotional nature
of many of our experiences of value, it is telling that in his attempt to demonstrate the
ways in which desires involve complex and varied experiences of value, Brewer (2009)
consistently appeals to examples, not of desires, but of paradigmatic emotional experi-
ences including fear, delight, musical emotions, sympathy, jealousy, admiration, grief,
gratitude, and wonder.7 Indeed, attempting to account for the richness and complexity
of our evaluative experiences apart from emotions is, to appropriate a phrase Brewer
employs in another context, “a recipe for inarticulacy.”
Given that many of our experiences of value seem to be essentially emotional,
understanding emotions as having epistemic justificatory force (that is, as conferring
epistemic justification) can help us make sense of the justification for many of our eval-
uative beliefs. How do we come to believe justifiedly (and, we might add, to know) that
the Holocaust was an abominable injustice or that sunsets are beautiful? One plausible
answer is that we directly experience the injustice of the Holocaust and the beauty of
sunsets through our emotions (in particular, through indignation or moral horror and
through a kind of aesthetic admiration or awe, respectively) and that our emotional
perceptions justify our beliefs.
In fact, without the aid of emotions we might lack the evaluative understanding
necessary for informedly, let alone justifiedly, believing at least some evaluative prop-
ositions. Consider, for example, the following story of the autistic scientist, Temple
Grandin, recounted by Oliver Sacks (1995: 293):
As we drove into the park, the landscape opened out into an immense mountain plateau, with
limitless views in every direction. We pulled off the road and gazed toward the Rockies—snow-
capped, outlined against the horizon, luminously clear even though they were nearly a hundred
miles away. I asked Temple if she did not feel a sense of their sublimity. “They’re pretty, yes.
Sublime, I don’t know.” When I pressed her, she said that she was puzzled by such words and had
spent much time with a dictionary, trying to understand them. She had looked up “sublime,”
“mysterious,” “numinous,” and “awe,” but they all seemed to be defined in terms of one another.
“The mountains are pretty,” she repeated, “but they don’t give me a special feeling, the feeling you
seem to enjoy.” After living for three and a half years in Fort Collins, she said, this was only the
second time she had been to them.
Sacks explains that Grandin exhibits a “poverty of emotional or aesthetic response

to most visual scenes: she can describe them with great accuracy but they do not seem
to correspond to or evoke any strongly felt states of mind” (1995: 286). Grandin’s lim-
ited emotional range seems, by her own account, to have resulted in a limited concep-
tual repertoire. Not having ever directly experienced sublimity through the emotion of
awe, or some related emotion, Grandin lacks the concept of sublimity; or, if you prefer,
she lacks our concept of sublimity. That is, even if Grandin has, with the help of the
dictionary and conversational testimony, formed a concept of sublimity that refers to
7
Brewer’s examples can be found on the following pages: fear (26), delight (60, 180), the musical emo-
tions (104), sympathy (111, 174), jealousy (113), admiration (156–7), grief (176), gratitude (180), and wonder
(160–1).
114 Adam C. Pelser
the same property as our emotion-generated concept, the descriptive content of her
concept—her understanding of sublimity—is diminished due to the fact that it is not
informed by any direct experiential acquaintance with the property.8 Grandin’s case
thus might be thought of as an emotional parallel of Frank Jackson’s famous Mary the
Scientist case (see Jackson 1982). Grandin’s autobiographical insight suggests that she is
unable to understand what sublimity is like on the basis of testimony, without the aid
of direct emotional experience of sublimity, thus limiting her ability to form the belief
that the Rockies are sublime, since even forming the belief seems to involve employing
a concept of sublimity that has been informed by direct acquaintance with the property.
Emotional experience thus might be necessary for the formation of some evaluative
concepts,9 let alone the justification of evaluative beliefs employing those concepts. The
justificatory thesis of emotion, together with the perceptual theory of emotion sketched
previously, has the resources to provide a compelling explanation of this phenomenon.
So far we have seen that the justificatory thesis of emotion, taken together with a
perceptual theory of emotion, provides a plausible explanation for the implicit trust
we place in our emotions when we form emotion-based beliefs, as well as for the fact
that at least some of our emotion-based beliefs seem to be justified. The reason some
emotion-based beliefs are justified is, plausibly, that emotions are direct experiences
(perceptions) of thick, particular values (as those values are instantiated in the objects
of emotions) and that, as such, emotions function as justifying reasons for correspond-
ing evaluative beliefs. Despite its explanatory power, however, the justificatory thesis
remains unconvincing to many for two primary reasons. First, even if emotions give
rise to true beliefs some of the time, emotion seems far too unreliable or untrustworthy
on the whole to count as a source of justified beliefs—call this the unreliability objec-
tion. Second, even if emotional experiences sometimes give rise to justified beliefs, it
seems that they only do so when the emotions themselves are justified; hence, what-
ever reasons or evidence justifies the emotion must also be the source of the justifica-
tion for the belief—call this the common source objection since it posits a common
(non-emotional) source of justification for emotions and their corresponding beliefs.10
In section 3 I shall take up the unreliability objection, leaving discussion of the com-
mon source objection for section 4.
3. The Unreliability Objection

Recent psychological studies on emotions and cognition have lent support to the unre-
liability objection. The psychologist Paul Slovic (2000) and his team have, for example,
demonstrated the prevalence of an “affect heuristic,” according to which many of their
experimental subjects’ judgments concerning the risks of various technologies were
8
On the referential/descriptive content distinction, see Buras (2008).
9
Cf. Pettigrove and Tanaka’s (2013: 14–15) discussion of “historically affective concepts.”
10
I am grateful to Daniel McKaughan for suggesting the phrase “common source objection.”
unduly influenced by whether they experienced positive or negative emotion (affect)

toward the expected benefits of the technologies. Such research, together with vast
anecdotal evidence of the unreliability of emotions in tracking the truth, poses a strong
(but not insurmountable) objection to the justificatory thesis of emotion.
In his introduction to a discussion of the standard sources of knowledge, which are
also widely taken to be the standard sources of epistemic justification, Matthias Steup
(2007) excludes emotions from the list, explaining that
Beliefs arise in people for a wide variety of causes. Among them, we must list psychological fac-
tors such as desires, emotional needs, prejudice, and biases of various kinds. Obviously, when
beliefs originate in sources like these, they don’t qualify as knowledge even if true. For true
beliefs to count as knowledge, it is necessary that they originate in sources that we have good
reason to consider reliable. These are perception, introspection, memory, reason, and testimony.
Steup, of course, is correct that a belief formed solely to satisfy an emotional need
is not a justified belief and, hence, not an instance of knowledge. Likewise, belief out-
puts of something like Slovic’s affect heuristic likely will be unjustified as well. Yet, not
all routes from emotional experience to belief are so epistemically suspect. A virtuous
person’s coming to believe that the Holocaust was a terrible injustice on the basis of the
indignation and moral horror evoked by pictures of concentration camps is neither a
case of belief formation for the sake of satisfying an emotional need nor a case of epis-
temically irresponsible reliance on an unreliable heuristic mechanism.
The first thing to note in reply to the unreliability objection, then, is that even though
a source of belief might be unreliable in some cases or for some individuals, this does
not entail that the source in question is never capable of generating justified beliefs.
As William Alston explains concerning sense perception—a source that is generally
taken to be very reliable (at least in favorable environments)—“for most of our beliefs,
including perceptual beliefs, what we typically identify as a justifier provides only
defeasible, prima facie justification” (1999: 223).
According to internalist and evidentialist, perhaps as well as some externalist, theo-
ries of justification there might be and often are overriding factors that prevent the
prima facie justification of a particular source of belief from becoming ultima facie
justification without rendering the source impotent to justify in every case. On such
accounts a particular emotional experience might fail to provide justification for
a belief formed on the basis of that experience in case the subject has evidence that
counts against trusting her emotional experience or conflicts with the propositional
content of the experience. The mere fact that the prima facie justification conferred
by emotional experience is often defeated, however, does not entail that the prima
facie justification provided by emotion never results in ultima facie justification.11
11
Ralph Wedgewood (2007: 243) offers a version of this conflicting evidence account of defeaters in his
defense of the claim that justification can be conferred by moral intuitions, which he takes to involve emo-
tions in important ways.
116 Adam C. Pelser
Alternatively, according to certain externalist theories of justification, an emotion’s

being unreliable or malfunctioning in a particular case might mean that the emotion
does not even grant prima facie justification.12 However, since this is also true of the
standard sources of justification on such externalist views, this does not threaten the
justificatory thesis, as long as emotions are not always unreliable or malfunctioning
sources of belief.
Yet, the worry might remain that emotions seem so highly unreliable that they never
actually confer (ultima facie) justification. The second thing to note in response to the
unreliability objection, then, is that some of the emotion-dispositions of some agents
reliably track the relevant features of their objects. As Robert Roberts and Jay Wood
(2004) argue, virtually the whole range of virtues, both intellectual and moral, are
in part dispositions to have the right emotions toward the right objects at the right
time and to the right degree. Insofar as some people possess some virtues, therefore,
the unreliability objection does not undermine the justificatory thesis of emotion. Of
course, situationist moral psychologists (e.g., Doris 1998 and Harman 1999) have chal-
lenged the assumption that people actually have robust virtues, but the situationist
challenge has, at best, shown that possession of moral virtues is rare (cf., Kamtekar
2004; Miller 2013), not that no one has them at all. Moreover, it remains an open
empirical question whether those of us who are less than virtuous have specific emo-
tion dispositions that generally function well enough to justify the beliefs to which
they give rise.
One final thing to note in response to the unreliability objection is that while many
people have a tendency to trust unreliable emotions, so too do many people have a
tendency to trust unreliable sense perceptions. Consider, for example, the amateur or
poorly trained wine-taster who, lacking the humility to admit her own lack of skill,
regularly forms false beliefs about the quality of wines on the basis of unreliable olfac-
tory and gustatory sensitivity. As with sense perception, introspection, memory, rea-
son, and testimony, the fact that some (perhaps most) people have a tendency to trust
the outputs of unreliable emotion dispositions does not threaten the ability of emo-
tions to function as a basic source of justified beliefs.
4. The Common Source Objection

Another common reason for rejecting the justificatory thesis of emotion is that when-
ever we have a justified belief that arises out of or in conjunction with an emotional expe-
rience (that is, an emotion-based belief), it seems we can find non-emotional reasons
for the belief (such as ordinary sense perceptual, memorial, or testimonial reasons) that
are sufficient to justify the belief without any aid from the emotion in question. Brady
12
Although he does not address emotions specifically, Bergmann (2006), for example, seems commit-
ted to the general claim that improperly functioning sources of justification do not provide prima facie
justification.
(2012) supports this objection with three primary observations. First, emotions them-
selves are reasons-responsive; that is, we often speak of the reasons that justify (or fail to
justify) having a particular emotion. Second, we do not typically trust emotions as suf-
ficient reasons to believe, but rather as reasons to look for non-emotional reasons to con-
firm “our initial emotional appraisal” (Brady 2012: 139–40). Third, Brady observes that
we rarely, if ever, cite emotions as our reasons for believing as we do; rather, when asked
why we believe that p, where our belief that p is (apparently) an emotion-based belief, we
tend to describe the non-evaluative features (sometimes referred to as the “natural prop-
erties”) of the situational objects of our emotions.13
4.1 Emotional Justification

Consider first Brady’s observation that emotions are reasons-responsive, that they
can be justified or unjustified. We do, indeed, often say things like “Your fear of that
harmless spider is irrational,” or “Her anger at her cheating husband is justified.” But
what is the basis for such judgments? In her seminal book on this question, Patricia
Greenspan (1988) argues that emotional justification involves both evidential or epis-
temic reasons for experiencing the emotion (“backward-looking” “appropriateness”)
and practical reasons for experiencing the emotion (“forward-looking” “applicabil-
ity”), though she acknowledges the priority of the epistemic reasons by referring to the
epistemic (“backward-looking”) “appropriateness” of emotions as “the justification of
the emotions themselves” (1988: 8). Setting aside considerations of whatever practical
justification there might be for experiencing certain emotions, emotional justification
seems to involve two primary components: an emotion is justified for an agent S just
in case (1) S’s beliefs about the object of the emotion on which the emotion is based are
(epistemically) justified, and (2) the emotion is an appropriate (that is, fitting or accu-
rate) response to the situation as S (justifiedly) believes it to be.
To put the first condition negatively, if one’s beliefs about an object on which one’s
emotion is based are themselves epistemically unjustified then one’s emotion is not
emotionally justified. For example, if George believes on the basis of scant evidence
or a mere hunch that his wife is having an affair, his emotion of jealousy that is based
on those beliefs is unjustified. This is not to claim that the beliefs basic to the emotion
must be true in order for the emotion to be justified, but only that those beliefs must be
justified. What this reveals is that emotion, like testimony and unlike sense perception,
is a dependent source of justification, if it is a basic source of justification at all. In order
to enjoy justification for a belief p formed on the basis of testimony, one must first have
some non-testimonial (typically perceptual) justification for the belief that the testifier
testified p. With dependent basic sources of justification, the evidence provided by the
source cannot be any stronger than the evidence for the beliefs on which the source
depends. So, if I do not have good reason to believe that you testified (or are testifying)
13
Peter Goldie (2004) also defends the common source objection, but as Brady develops the objection in
more detail I will focus my response on his articulation of the objection.
118 Adam C. Pelser
some proposition p, then, a fortiori, I do not have good reason to believe p on the basis
of your testimony. Likewise, if the beliefs basic to my emotion are not justified then, a
fortiori, any beliefs formed on the basis of my emotion will not be justified.
That the beliefs basic to an emotion are justified, however, is not a sufficient con-
dition for the justification of the emotion. The emotion must also be an evaluatively
appropriate or fitting response to the object as it is justifiedly believed to be. This is
not to say that there is only one justified emotion for every value-laden state of affairs.
A father whose daughter dies in a car crash caused by a drunk driver might be justi-
fied in his grief over the loss of his daughter as well as his anger at the drunk driver,
the varying emotions being justified by the evaluative complexity of the situation. By
contrast, we would not say that the father is justified in feeling grateful, since gratitude
would be an inappropriate and unfitting emotional response—it would get the value of
the situation wrong.
Notice, here, that neither condition for emotional justification amounts to epistemic
justification for the evaluative belief(s) to which an emotion might give rise. In order to
be justified in grieving over his daughter’s death, for example, a father must have good
reason (perceptual, testimonial, and so on) to believe she is dead (the belief basic to the
grief), but he need not have non-emotion-based epistemic justification for believing
her death to be an irrevocable loss of someone of great value (the belief consequent to
the grief). Imagine, for example, an estranged father believing (at least disposition-
ally), as a result of having his heart hardened by many years of estrangement from his
daughter, that her death would not constitute a significant loss of someone valuable.
Perhaps he even loses whatever justification he might have previously enjoyed for the
belief that his daughter is valuable and that her death would be a great loss as a result
of the testimonial evidence of a ‘support group’ of similarly hard-hearted, absentee
fathers. Now imagine that same father coming to believe on the basis of his (emotion-
ally) justified grief in response to news of his daughter’s death that she is indeed pre-
cious and valuable and that her death is a great loss. In such a case, assuming his grief
arises out of a trustworthy emotion-disposition, the emotion might justify the conse-
quent belief without relying on antecedent justification for that belief. That emotions
of loss often seem to function as such ‘new’ evidence or reasons for belief is reflected in
the well-known phrase, “You don’t know what you’ve got until it’s gone.”
While brief and merely suggestive, the foregoing analysis is sufficient for the present
purpose of defending the justificatory thesis against the common source objection.
As the analysis reveals, emotional justification does depend on epistemic justification
for some of the beliefs relevant to emotional experience, but only for the beliefs basic
to emotions, not for the evaluative beliefs consequent to emotions, which I have been
calling emotion-based beliefs.
One might object at this point to the assumption that the beliefs in question are
emotion-based in the first place. To use an example offered by Brady, why not think
that one can form a belief that a dog is dangerous immediately and non-inferentially
on the basis of perceptual awareness that the dog is approaching, snarling, and off its
leash? Even granting that such non-emotional experiences of value are possible, most
people’s experiences of danger (and a father’s experience of the significant loss of his
precious daughter) are not so emotionless. Normal experiences of danger seem to be
not only accompanied or mediated, but also constituted by emotions of fright.
Of course, someone with a diminished range of emotional responses to visual stimuli
like Temple Grandin might be able, with the help of testimony, to learn to ‘perceive’ the
Rockies as sublime without the aid of a corresponding emotion. Yet this is no reason to
think that the evaluative beliefs of others, whose beliefs are formed on the basis of very
qualitatively different (emotional) perceptions of the sublimity of the Rockies, must be
justified solely by the features of their experience that they share with those who have
no direct experience of or acquaintance with sublimity. Consider the way in which a
color-blind person might learn to ‘see’ his socks with the white stripes as his black socks
so as not to get them mixed up with his brown socks which have no stripe. The fact
that the color-blind person can learn to have such an experience of the striped socks
as black does not entail that it is only the evidence available to him, which involves no
direct experience of the blackness of the socks, that is responsible for the justification
enjoyed by those who believe the socks to be black on the basis of direct perceptual
awareness of or acquaintance with their blackness. What justifies the normal sighted
person’s perceptual belief that the socks are black is her perception of their blackness (or,
if you prefer, what justifies her belief is the blackness of the socks, which she perceives14).
We have seen that emotions seem to be at least one way, if not the most natural way,
that we experience certain values. Why, then, should we think that just because there
is an emotionless way to experience those values that it is the emotionless experience,
or certain aspects thereof, which must account for the justification of evaluative beliefs
in those instances when the belief is formed on the basis of an emotional experience? I
answer that we should not.
4.2 Are Emotions Just Reasons to Look for Reasons?

Let us now briefly consider Brady’s second observation in support of the common
source objection; namely, that we typically treat emotions, not as sufficient reasons to
believe, but as reasons to look for other reasons. While we often withhold believing on
the basis of emotional experience until we have found additional corroborating evi-
dence, this is by no means a universal phenomenon. As Clore and Gaspar (2000) note,
there is good empirical evidence that we often treat our emotional experiences as suf-
ficient reasons for corresponding beliefs. In fact, as discussed previously, the popular
view that emotions are not capable of functioning as sources of epistemic justification
(or knowledge) is often motivated by the observation that we so often do form beliefs on
the basis of our implicit trust in unreliable emotional experiences. In addition to not-
ing that we often treat emotions as sufficient reasons to believe, it is important to note
14
Miller (2008) argues that it is not cognitive mental states, but rather facts about the world that serve as
subjective epistemic reasons.
120 Adam C. Pelser
that we do not always treat the standard sources of justification as sufficient reasons to
believe. Indeed, we often have visual and auditory experiences that compel us to look or
listen more closely before forming any beliefs about the objects of our sensory experi-
ences. The fact that we occasionally treat sense perceptual experiences not as sufficient
reasons to believe, but rather as reasons to look for more supporting evidence, does not
undermine the ability of sense perceptual experiences to justify beliefs. Neither does
the fact that we sometimes treat emotions as reasons to look for more supporting evi-
dence for our evaluative beliefs undermine the justificatory thesis of emotion.
4.3 Why Do We Not Cite Our Emotions As Our Justification?

What, then, should we say in response to Brady’s third observation in support of the
common source objection—that is, that we do not cite emotions as our reasons for
believing as we do even when we seem to have formed our beliefs on the basis of an
emotional experience? To see that this observation fails to undermine the justificatory
thesis, consider the difference between having a reason or justification and being able
to give a reason or justification in conversation. It occasionally happens when we base
our beliefs on perceptual evidence or reasons that our reasons for believing as we do
are not shareable with others. Were I to try to convince someone that I had good reason
for believing that it is raining outside I might explain that I see the rain when I look out
the window. Alternatively, I might just point at the rain and expect my interlocutor to
have the same perceptual experience I am having. I am especially likely to choose the
pointing strategy over the citation of my reasons when I am attempting to convince
a skeptical interlocutor that it is raining outside or that my perceptual experience is
good reason to believe that it is raining. Since I cannot actually present my interlocutor
with my evidence—that is, my perceptual experience of the rain—if he is unwilling to
trust my testimony that I have had such a justifying experience, the best I can do is to
point and hope that he will gain similar perceptual evidence of his own.
Given our society’s widespread skepticism about the trustworthiness of emotions
and the prevalence of emotional disagreement, it should not be surprising that most
cases in which we are asked to give reasons for an emotion-based evaluative belief are
cases in which citing our emotional experience alone will not be convincing to our
interlocutors. Recognizing this, we often avoid appealing to emotional experience,
choosing instead to describe (often vividly) the non-evaluative features of its situ-
ational object in the hopes that our interlocutor will experience a similar emotional
reaction to the object as described, thereby gaining emotional evidence of her own
for the evaluative belief in question. This is analogous to pointing in sense percep-
tion cases. Incidentally, much moral debate, both in the general population and among
professional philosophers, involves precisely this rhetorical strategy.
Moreover, in light of the way that evaluative properties supervene on non-evaluative
properties, by describing the non-evaluative properties of a situation to an interlocu-
tor we can, as it were, place the value-laden state of affairs before her mind in hopes
that she will have a similar emotional reaction and, as a result, form the evaluative
belief in question. So, when someone asks why we find some situation unjust, instead
of citing our indignation toward the situation, we describe the non-evaluative details
of the situation and then expect our interlocutor to ‘see’ (construe) it as unjust through
her own response of indignation or some similar emotion. Of course, it sometimes
happens that our interlocutor has a very different emotional reaction or no emotional
reaction at all to the situation as we present it. Despite such emotional disagreement,
the object might still have evaluative properties that really are there to be experienced
by those whose sensitivity to the non-evaluative as well as the evaluative features of
situations enables accurate emotional (perceptual) experience (cf., Watkins and Jolley
2002; Deonna and Teroni 2012: 112–15). Indeed, neither the fact of emotional disagree-
ment nor the fact that we do not typically appeal directly to our emotional states when
asked to give the reasons for our evaluative beliefs shows that emotions are not capable
of being justifying reasons for beliefs.
Conclusion
We often trust our emotions by forming emotion-based beliefs. The justificatory the-
sis of emotion has it that at least some emotion-based beliefs are directly and non-
inferentially justified by emotions themselves; that is, that emotion is a basic source
of epistemic justification. We have seen that the justificatory thesis, especially when
understood in light of the perceptual character of emotions, provides a plausible
explanation for the justification we enjoy with respect to many of our thick evalua-
tive beliefs. Moreover, the justificatory thesis of emotion is capable of withstanding
the most challenging objections to the view. While the observation that emotions can
themselves be justified and thus that emotion is a dependent source of justification
reveals a significant epistemological difference between emotions and sense percep-
tions, this observation does not threaten the justificatory thesis or the (non-sensory)
perceptual theory of emotion sketched previously. Likewise, while empirical research
might yet reveal that emotions are less reliable than sense perceptions in producing
true beliefs, their general unreliability would not undermine their ability to justify
some beliefs for some agents. The fact that we often form beliefs through emotional
experience unreliably, and despite the presence of defeating evidence, should not lead
us to reject emotion as a source of justified belief, but rather should encourage us to
seek out ways to improve our emotional sensitivities to the value in the world, thereby
improving both our moral character and our stock of justified evaluative beliefs.15
15
For helpful comments on previous drafts, I am grateful to Jonathan Kvanvig, Stephen Evans, Christian
Miller, Daniel McKaughan, Win-Chiat Lee, and Adam Kadlac, as well as audiences at the 2009 Central
Division meeting of the American Philosophical Association and the 2011 Moral Emotions and Intuitions
conference held in The Hague. I am especially grateful to Robert Roberts and Daniel Johnson for many help-
ful conversations on this material and for providing insightful and challenging comments on multiple drafts
of this chapter. Support for this work was funded in part by the Wake Forest Philosophy Department Thomas
Jack Lynch Memorial Fund and by the Character Project grant from the John Templeton Foundation. The
122 Adam C. Pelser
References
Alston, W. (1999). “Perceptual Knowledge.” In The Blackwell Guide to Epistemology, ed. J. Greco
and E. Sosa (Malden, MA: Blackwell), 223‒42.
Alston, W. (2005). Beyond “Justification”: Dimensions of Epistemic Evaluation.(Ithaca,
NY: Cornell University Press).
Bergmann, M. (2006). Justification without Awareness: A Defense of Epistemic Externalism
Brady, M. (2012). “Emotions, Perceptions, and Reasons.” In Morality and the Emotions, ed. C.
Bagnoli (Oxford: Oxford University Press), 135–49.
Brewer, T. (2009). The Retrieval of Ethics (Oxford: Oxford University Press).
Buras, T. (2008). “Three Grades of Immediate Perception: Thomas Reid’s Distinctions.”
Philosophy and Phenomenological Research 76: 603–32.
Clore, G. R, and Gaspar, K. (2000). “Some Affective Influences on Belief.” In Emotions and
Beliefs: How Feelings Influence Thoughts, ed. N. H. Frijda, A. S. R. Manstead, and S. Bemet
(Cambridge: Cambridge University Press), 10–44.
Conee, E. and Feldman, R. (2004). Evidentialism (Oxford: Oxford University Press).
Deonna, J. A. and Teroni, F. (2012). The Emotions: A Philosophical Introduction
Döring, S. (2003). “Explaining Action by Emotion.” The Philosophical Quarterly 53: 214‒30.
Doris, J. M. (1998). “Persons, Situations, and Virtue Ethics.” Noûs 32: 504–30.
Elgin, C. Z. (2008). “Emotion and Understanding.” In Epistemology and Emotions, ed. G. Brün,
U. Dogûoglu, and D. Kuenzle (Burlington, VT: Ashgate), 33–50.
Frijda, N. H. and Mesquita, B. (2000). “Beliefs through Emotions.” In Emotions and
Beliefs: How Feelings Influence Thoughts, ed. N. H. Frijda, A. S. R. Manstead, and S. Bemet
(Cambridge: Cambridge University Press), 45–77.
Goldie, P. (2004). “Emotion, Feeling, and Knowledge of the World.” In Thinking about Feeling,
ed. R. C. Solomon (Oxford: Oxford University Press), 91–106.
Greenspan, P. (1988). Emotions and Reasons: An Inquiry into Emotional Justification
Harman, G. (1999). “Moral Philosophy Meets Social Psychology: Virtue Ethics and the
Fundamental Attribution Error.” Proceedings of the Aristotelian Society 99: 315–31.
Jackson, F. (1982). “Epiphenomenal Qualia.” Philosophical Quarterly 32: 127–36.
Kamtekar, R. (2004). “Situationism and Virtue Ethics on the Content of our Character.” Ethics
114: 458–91.
Kvanvig, J. and Riggs, W. D. (1992). “Can a Coherence Theory Appeal to Appearance States?”
Philosophical Studies 67: 197–217.
Lamb, R. (1987). “Objectless Emotions.” Philosophy and Phenomenological Research 48: 107–17.
Miller, C. (2008). “Motivation in Agents.” Noûs 42: 222–66.
Miller, C. (2013). Moral Character: An Empirical Theory (Oxford: Oxford University Press).
Nussbaum, M. (2001). Upheavals of Thought (Cambridge: Cambridge University Press).
opinions expressed in this chapter are my own and do not reflect the views of the Character Project or the
Templeton Foundation.
Oddie, G. (2005). Value, Reality, and Desire (Oxford: Oxford University Press).

Pelser, A. (2010). “Belief in Reid’s Theory of Perception.” History of Philosophy Quarterly
27: 359–78.
Pettigrove, G. and Tanaka, S. (2013). “Anger and Moral Judgment.” Australasian Journal of
Philosophy <http://dx.doi.org/10.1080/00048402.2013.795990>.
Prinz, J. (2004). Gut Reactions: A Perceptual Theory of Emotions (Oxford: Oxford University
Press).
Prinz, J. (2007). The Emotional Construction of Morals (Oxford: Oxford University Press).
Roberts, R. C. (2003). Emotions: An Essay in Aid of Moral Psychology (Cambridge: Cambridge
University Press).
Roberts, R. C. (2010). “Emotions and Judgments about Risk.” In Emotions and Risky Technologies,
ed. S. Roeser (Dordrecht: Springer), 107–27.
Roberts, R. C. (2013). Emotions in the Moral Life (Cambridge: Cambridge University Press).
Roberts, R. C. and Wood, W. J. (2004). “Proper Function, Emotion, and Virtues of the Intellect.”
Faith and Philosophy 21: 3–24.
Roeser, S. (2011). Moral Emotions and Intuitions (New York: Palgrave Macmillan).
Sacks, O. (1995). An Anthropologist on Mars (New York: Vintage Books).
Slovic, P., Finucane, M. L., Alhakami, A., and JohnsonS. M. (2000). “The Affect Heuristic in
Judgments of Risks and Benefits.” Journal of Behavioral Decision Making 13: 1–17.
Sosa, E. (2007). A Virtue Epistemology: Apt Belief and Reflective Knowledge, vol. 1 (Oxford: Oxford
University Press).
Steup, M. (2007). “Epistemology.” In The Stanford Encyclopedia of Philosophy (Winter 2007
Edition), ed. E.N. Zalta <http://plato.stanford.edu/archives/win2007/entries/epistemology>.
Watkins, M. and Jolley, K. D. (2002). “Pollyanna Realism: Moral Perception and Moral
Properties.” Australasian Journal of Philosophy 80: 75–85.
Wedgewood, R. (2007). The Nature of Normativity (Oxford: Oxford University Press).
Whiting, D. (2012). “Are Emotions Perceptual Experiences of Value?” Ratio 25: 93–107.
Zagzebski, L. (2003). “Emotion and Moral Judgment.” Philosophy and Phenomenological
Research 66: 104–24.
Zagzebski, L. (2012). Epistemic Authority: A Theory of Trust, Authority, and Autonomy in Belief
8
Why Recalcitrant Emotions Are
Not Irrational
Sabine A. Döring
It is fairly uncontroversial among contemporary philosophers that emotions play a

rational role, and not merely a causal role. However, what is uncontroversial is only
that the emotions play a rational role, not how this role is to be explained. At the core of
the debate is the phenomenon of recalcitrant emotions. A ‘recalcitrant’ emotion is one
which persists despite the subject’s ‘better’ judgement; that is, despite his judgement,
reached by deliberation, about what is right, all things considered. David Hume and,
before him, Michel de Montaigne provide a prime example: your fear of falling may
persist although you judge that you are safe.1 This kind of conflict between emotion
and judgement need not be a pathological case of acrophobia, but occurs as an ordi-
nary experience of ordinary people. It is, as Bennett Helm put it, ‘readily intelligible
and happens all too often’.2
Helm thinks, therefore, that any theory about the emotions which conceives them
as (similar to) judgements is bound to fail. Recalcitrant emotions, Helm says, are irra-
tional but not so irrational as recalcitrant judgements would be. At the same time,
Helm dismisses the view that emotions are (or resemble) perceptions as unsatisfying,
since he has it that recalcitrant perceptions are not irrational.
My aim here is to defend a perceptual model of emotion against the Helm objec-
tion, yet not by accounting for it but by showing that this objection is based on a wrong
intuition. Recalcitrant emotions are not irrational. More precisely speaking, the subject
is not irrational in experiencing a recalcitrant emotion. Just like recalcitrant percep-
tions, recalcitrant emotions involve a cognitive conflict between experiential state and
1
In his Treatise Hume invites us to consider ‘the case of a man, who being hung out from a high tower in a
cage of iron cannot forbear trembling, when he surveys the precipice below him, tho’ he knows himself to be
perfectly secure from falling, by his experience of the solidity of the iron, which supports him’ (Hume 1978
[1739/40]: 148); see also Montaigne (2007 [1588]: Book 2, chap. 12).
2
Helm (2001: 42).
Why Recalcitrant Emotions Are Not Irrational 125
better judgement, and yet the subject is not irrational because he does not contradict
himself. The wrong intuition that recalcitrant emotions involve irrationality, whereas
recalcitrant perceptions do not, occurs because, in addition to cognitive conflict, recal-
citrant emotions typically lead to practical conflict by motivating the subject to act in
ways that interfere with the reasoned pursuit of his goals.
I.
Many have argued that, because of the emotions’ possible recalcitrance, their rational
role cannot be explained by a judgementalist model. According to judgementalism,
an emotion is, or conceptually involves, a certain evaluative judgement: to fear x is, or
involves, to judge that x is fearsome. Emotional recalcitrance would then be explained
by attributing to the subject two conflicting and even contradictory judgements: the
judgement that x is fearsome and, at the same time, the judgement that x is harmless
(and thus not fearsome). To make this explanation more plausible, judgementalists
resort to the claim that the judgement which supposedly defines the emotion is some-
how unconsciously held and thus not acknowledged by the subject. However, their
only reason for claiming this seems to be the assumption that judgementalism is true,
while they would need an additional, special reason for violating the ‘principle of char-
ity’.3 After all, the subject is attributed a rational dysfunction in the form of an ‘unac-
knowledged judgement in conflict with those he acknowledges’, as Patricia Greenspan
was the first to point out.4
By now, judgementalism is mostly abandoned in favour either of perceptual or of
‘quasi-judgementalist’ models.5 Perceptual models conceive emotions as, or by anal-
ogy with, perception: to fear x is to perceive x as fearsome. Accordingly, recalcitrant
emotions are seen to be similar to recalcitrant perceptions. Standard examples are per-
ceptual illusions such as the bent stick in water or the famous Müller-Lyer illusion.
Just as, despite their better judgement, the protagonists in Hume’s and Montaigne’s
examples cannot help but be afraid of falling, the perceiver of the Müller-Lyer illusion
cannot rid himself of seeing the two lines as being of different lengths, even when care-
ful study has convinced him that they are the same length.6
Against this assimilation of recalcitrant emotions to perceptual illusions, Helm
has objected that it fails to capture the irrationality involved in recalcitrant emotions.
Although Helm agrees that judgementalism cannot do justice to the phenomenon of
emotional recalcitrance, according to him, perceptualism cannot do so either. For,
in Helm’s view, there is an important difference between recalcitrant emotions and
3
Roughly speaking, this principle requires that we ought to avoid attributing irrationality to speakers
when a rational interpretation of their statements is available. I am applying this methodological principle to
the interpretation of a person’s judgements.
4
Greenspan (1988: 18). 5 See, e.g., Brady (2009).
6
See the discussion in Crane (1992: 151f.).
126 Sabine A. Döring
perceptual illusions consisting in that recalcitrant emotions are irrational, whereas

perceptual illusions are not. ‘It is not at all irrational to have a stick half-submerged
in water look bent even after one has judged that it is straight’, Helm says.7 But it is
irrational by Helm’s lights to feel fear of some x after one has found out that x is in fact
harmless.
The ‘Helm objection’, as I will call it, presents a particularly serious challenge to the
perceptualist, since he typically uses the supposed analogy between recalcitrant emo-
tions and perceptual illusions as an essential argument for his account of emotion.
While some are, because of the Helm objection, driven to a quasi- or neo-judgemen-
talist account of emotion, Christine Tappolet has recently defended perceptualism.8
She accepts the Helm objection, but argues, pointing to numerous other analogies
between emotions and sensory perceptions, that this objection does not threaten per-
ceptualism: on her view, it is only because our emotional systems are more plastic than
our sensory systems that emotions are subject to rationality requirements, whereas
sensory perceptions are not. Thanks to our emotions’ plasticity, Tappolet maintains,
‘we can hope to lose our emotional illusions’, but we ‘cannot hope to learn to see the
lines of the Müller-Lyer illusion as having the same length’.9 This is not to say that we
can do much about a recalcitrant emotion while experiencing that emotion. Tappolet’s
claim rather is that we can influence our emotional dispositions over time so as to make
them fit our better judgements. Therefore we are, according to Tappolet, rationally
required to do so—‘can’ implies ‘ought’, in this case—whereas it would be pointless to
pose a similar rationality requirement on our sensory perceptions of which Tappolet
denies that they could be improved in the same way.
Like Tappolet, I shall here defend a perceptual view of emotion. But unlike Tappolet,
I will reject the Helm objection instead of trying to account for it. The Helm objection
is grounded on an intuition. As Michael Brady has written: ‘We have an intuitive sense
that there is something wrong, from the standpoint of rationality, when fear persists in
the face of a subject’s judgement that she is in little or no danger’.10 Yet intuitions may
be wrong, and I think that the one underlying the Helm objection is in fact wrong. To
show this, let us first further investigate the nature of the rationality requirement pos-
tulated by Tappolet.
Within Tappolet’s account, the requirement in question obviously is a practical
requirement. Since we can improve the reliability of a misfiring emotional system, we
ought to do it, so the story goes. As this already indicates, it is according to Tappolet a
supposed epistemic function of emotion which grounds the practical requirement: we
ought to improve our emotions’ reliability, if possible. Like sensory perceptions, emo-
tions are seen by Tappolet as functioning so as to provide us with knowledge about
the world, with knowledge about values in this case. We may then ask whether the
practical requirement in question is instrumental (means‒ends directed), prudential
7
Helm (2001: 42f.). 8 See Tappolet (2012).
9
Tappolet (2012: 221). 10 Brady (2009: 413f.).
(benefit directed), or categorical (apodictic). Tappolet seems to be presuming that it is

categorical: whenever we can improve the reliability of an epistemic mental system, we
ought to do it, whether or not doing so serves our ends and purposes, or is beneficial.
But this can hardly be true and, more importantly, it fails to establish the claimed dif-
ference in rationality between emotion and perception.
On the one hand, there are clear cases in which we do not dismiss persons as irra-
tional when they abstain from trying to improve their emotional dispositions.
Consider the following entry headed Help me enjoy the scenic views that I found in a
hiking forum on the Internet: ‘I am an avid hiker, and as are most people that do such
things, interested in the nice views at the top of mountains, cliffs, and so on. However,
I have a fear of heights that is limiting my ability to do an[d]‌enjoy certain hikes. My
fear is specific to being in a high open space, near the edge of a steep drop . . . Does any-
one know if it’s possible to ‘conquer’ a fear of heights, through repetition, mind tricks,
something else?’ Let us assume that there is a way for this person to get rid of her fear.
Does this imply that she is rationally required to do so? I do not think so. The person
may find out that the costs for learning to keep her fear under control are immense—
long-lasting intense therapy instead of mere ‘mind tricks’, say—and she may, therefore,
decide to restrict hiking to valleys or to move on to a completely different hobby. And
we would hardly accuse her of irrationality if she does so.
On the other hand, if the requirement to improve, if possible, the reliability of one’s
epistemic mental systems were categorical, it is quite unclear why it should not also
apply to sensory perception. The examples of the bent stick in water and the Müller-
Lyer illusion are biased towards the non-plasticity of sensory perception, although
Tappolet admits that our perceptual systems merely are less plastic than our emotional
systems. Sensory perception is shaped, and can be reshaped, by our socio-cultural
environment, even if less so than emotion. Imagine, for example, that in a flat that I
intend to rent I smell something that I believe is mould. An expert nose tells me that
this is not mould but must be something else—perhaps still the ‘old-person smell’ of
the former residents. I trust the expert nose, and yet I cannot avoid perceiving the
smell as that of mould and, in the end, decide not to move in. In this case I can hope
to rid myself of the perceptual illusion by learning to distinguish the particular smell
of mould from other earthy or musty smells. But am I, therefore, under a categorical
rationality requirement to improve the reliability of my olfactory perceptual system?
Again, the answer is ‘no’, I think. Or imagine me in a restaurant drinking a glass of
red wine which to me tastes exactly like the Merlot that I had for dinner the night
before. The waiter tries to convince me that what I am drinking is in fact a Cabernet
Sauvignon, and finally shows me the bottle. Nevertheless to me it tastes like the Merlot
from yesterday. Again, it would well be possible for me to improve my taste perception.
And again, I doubt that rationality requires any such thing. Just as in the hiking exam-
ple, the costs may in both examples be too high compared to the benefits.
All three examples illustrate that we are not required to improve the reliability of an
epistemic mental system whenever this is possible. First, the same requirement would
have to hold for sensory perception but would, secondly, clearly be asking too much.
What we ought to do instead is to improve the reliability of both emotions and percep-
tions whenever this is useful or beneficial. If I want to hike in the mountains, and, in
order to achieve that goal, have to learn to master my acrophobia, I ought to learn to
master my acrophobia. But there is no such ‘ought’ in the absence of my goal, which
could well be overridden by other goals—goals that matter more to me in my life than
hiking in the mountains. Of course, in order to be able to live a life and to survive, our
epistemic mental systems must to a certain extent function reliably, yet survival leaves
considerable freedom for further goals and, ultimately, is just another goal.
This in turn sheds light on the source of the intuition that only recalcitrant emotions
are irrational. This intuition, I claim, is due to the fact that recalcitrant emotions inter-
fere to a much greater extent with the reasoned pursuit of our goals than recalcitrant
perceptions because, by contrast with perceptions, emotions have motivational force.
Both emotions and perceptions are occurrent conscious states that can capture and
occupy the subject’s attention.11 Because of this salience in consciousness, they both
incline the subject to take their content at face value, even when the subject has already
repudiated this content in judgement. Therefore, the intuition that only recalcitrant
emotions are irrational can, pace Brady, not be explained by a difference in attentional
influence, as is also noted by Tappolet.12 It is not that only recalcitrant emotions are irra-
tional because only they incline the subject to accept a content which the subject has
already rejected. Whatever my judgement be: seeing the stick in water as bent inclines
me no less to judge that the stick is bent than fearing to fall inclines me to judge that
I am in danger.
By contrast, only recalcitrant emotions have motivational force. Emotions do not
only come with epistemic inclinations; they also motivate the subject to act in cer-
tain ways. I believe (and have argued elsewhere) that this difference to perception is
the heart of the intuition that recalcitrant emotions involve irrationality. Emotions
continue to move us to act even after we have rejected their content. I may reject my
fear of falling as unwarranted, and yet be moved to give in to it, as the example of the
acrophobic illustrates. Or I may be aware of my hypersensitivity, and yet (re)act out
of being offended, contrary to what I hold appropriate. Because of their motivational
force, recalcitrant emotions cannot easily be kept from interfering with the reasoned
pursuit of our goals. Recalcitrant perceptions present a less difficult obstacle, although
again, I think, that the examples of the bent stick in water and the Müller-Lyer-illusion
are biased. Perceptual illusion due to refraction need not be an obstacle to achieving
your goals in the bent-stick-in-water-illusion, but in trying to catch a fish with your
bare hands it will certainly be.
11
For the present, my only concern is with occurrent conscious emotions (jealousy of a rival at a party); I
am not dealing here with emotional dispositions (the jealous husband; that is, his disposition to be jealous).
12
See Tappolet (2012: 218); see also Brady (2009: 422ff.) and Brady (2007: at 279ff.).
Nevertheless, the crucial difference between emotions and perceptions is that only
the former motivate action. And it is the actions caused by recalcitrant emotions which
are irrational insofar as they interfere with the reasoned pursuit of our goals, but not
the emotions themselves—or so I will argue.
II.
Before doing so, let me complete my argument against Tappolet’s account by point-
ing out that adjusting our emotional dispositions to our better judgements, whenever
possible, would, from the standpoint of rationality, even be counter-productive.13 The
putative irrationality of recalcitrant emotions is typically discussed in cases where the
emotion goes wrong and where it is comparatively easy to agree that it goes wrong.
The notorious example is that of the ‘lovable, old dog’ Fido whom the protagonist fears
although, ‘given his age, arthritis, and lack of teeth’, Fido is obviously harmless.14 Other
than in this example, however, it need not be the emotion which gets things wrong
when it comes to conflict with so-called ‘better’ judgement; it may equally be the judge-
ment—which is ‘better’ merely in the sense that it is ‘deliberate’.15 The prime example
is that of Huckleberry Finn who, after having helped his friend Jim to run away from
slavery, decides to turn him in but, when he is given the opportunity to do so, finds
himself doing just the contrary. Instead of turning Jim over to the slave hunters, Huck
lies out of sympathy in order to protect his friend, even though he does not endorse
his emotion but castigates himself for his weakness.16 In this example, Huck is practi-
cally irrational (akratic), but I claim he would not be irrational at all were he capable of
keeping his recalcitrant sympathy under control. Recalcitrance of his sympathy is even
required, from the standpoint of rationality, in order to enable Huck to reconsider his
better judgement so as to bring it into line with his emotions in the end.
Tappolet might reply that we ought to correct our emotional dispositions only in
cases where we know for certain that an emotion misfires. The underlying view is that,
thanks to certain natural or descriptive features of the world upon which the evalu-
ative features perceived in an emotion ‘supervene’, it is definitely decidable whether
or not an emotion is appropiate. Tappolet is a strict realist about value and ascribes
to the emotions the epistemic function to directly perceive value. According to her, it
is no objection to this view that our emotional reactions, and the ascriptions of value
implied by those reactions, are open to criticism and justification via reasons. This dif-
ference to sensory perception, Tappolet says, is explainable precisely by the fact that
13
See also Döring (2010).
14
See Greenspan (1981: 162f.) and Helm (2001: 41).
15
See Arpaly (2000) and Arpaly and Schroeder (2012). See also Arpaly (2003).
16
This example is due to Bennett (1974). Similar cases are described by Arpaly (2000); Jones (2003); and
Tappolet (2003).
evaluative features supervene upon descriptive ones which thus can be cited as rea-
sons. Because of his age, arthritis, and lack of teeth, Fido is not fearsome.
This explanation faces two objections. First, descriptive features are not decisive for
the presence (or absence) of evaluative features. The Fido example suggests that they
are, for it is comparatively easy to agree that an old arthritic dog without teeth is not
fearsome. But, as we all know, in typical cases of value ascription it is impossible to
decide on the basis of descriptive features alone. Just compare the Huck Finn exam-
ple: when Huck, in feeling sympathy for Jim sees and understands, Jim as in need for
freedom, this evaluative or normative feature of Jim is not fixed and determined by
his descriptive features. It also depends on Huck’s growing friendship with Jim; that
is, on his growing concern for Jim and his well-being. This is not to say that descriptive
features have no role to play in the constitution of value. For example, sympathy, and
the ascription of value implied by it, conceptually requires of an object that it itself has
feelings; and a need for freedom necessarily presupposes the capacity for freedom. But
the role of such descriptive features is only co-constitutive: value also depends on what
we care about. And this means that the rationality of an emotion, and of the ascription
of value implied by it, cannot be assessed in terms of the emotion’s reliability in track-
ing natural (or descriptive) features alone; it also depends on our conception of the good
life.
This is my second objection against Tappolet’s strict value realism which, in combi-
nation with her view of emotions as perceptions of value, is ultimately responsible for
her claim that we ought to improve the epistemic reliability of our emotions whenever
we can. I shall not further elaborate on this objection here but use it as a starting point
for an outline of my own perceptual account of emotion, based on which I will then try
to undermine the Helm objection.
III.
Unlike Tappolet, I do not claim that emotions are a kind of sensory perception. My
claim is instead that the emotions’ rational role is analogous to that of perception in
that the contents of emotions are non-inferentially related to the contents of other men-
tal states, including other emotions. Only in this sense I say that emotions are per-
ceptions.17 There are a number of obvious disanalogies between emotions and sense
perceptions, starting with the fact that emotions do not have organs or transducers.
None of these disanalogies undermines the analogy, since all that it requires is that
both sense perceptions and emotions have an intentional content of a certain kind,
thanks to which they play a non-inferential role in justifying relations.
I shall not address the question of non-conceptuality here, which, thinking of Tim
Crane’s by now classical paper ‘The Nonconceptual Content of Experience’, of course
See Döring (2007).

17
lurks in the background of non-inferentiality.18 For the present, all I want to show is
that emotions are not propositional attitudes, nor intentional mental states directed at
particular objects. Most contemporary philosophers still seem to think that intention-
ality is to be explained in terms of propositional attitudes alone, be it truth-directed
beliefs or goal-directed desires. Against this, some have argued that the typical target
(intentional object) of an emotion is not a state of affairs but a particular object: you
love your child, are angry at your partner, or fear the gorilla. As we will see, although
emotions are not propositional attitudes, this is a too narrow and simplistic approach.
Inferential relations hold between the propositional contents of judgements, or
beliefs (I will use ‘judgement’ and ‘belief ’ interchangeably here). For an example, con-
sider the simple deductive inference rule modus ponens. This inference rule does apply
to the propositional content of belief, but it does not apply to the content of percep-
tion, nor does it apply to the content of emotion. If you see a vertical line and then
you become aware of a dot above it so that you see the letter i, there is no such thing as
deductively inferring the perception that this an i from the two perceptions (1) that if
a figure is a vertical line with a dot above it, it is an i, and (2) that this is a vertical line
with a dot above it. No modus ponens involved here. Instead, you immediately see the
letter i. The same point can be made for the content of emotion. To take up one of Peter
Goldie’s examples: if you feel that someone’s behaviour is irritating, but then start to
feel amused about how over the top his irritating behaviour is, your amusement’s con-
tent cannot be deductively inferred from the content (1) that if behaviour is way over
the top it is funny, and (2) that this behaviour is way over the top. Just as you immedi-
ately see the letter i, you immediately see the behaviour as funny.
The examples are instructive also in two further respects. First, they show that
although the evaluative content of emotion is in some way open to justification via
reasons, those reasons are never compelling. Imagine that your friend also feels that the
person’s behaviour is irritating but fails to grasp the comedy of the situation. All you
can do to convince him of the fittingness of your amusement is to point to the salient
features which, for your eyes, make the behaviour funny. None of your hints will suf-
fice to ‘force’ your friend to infer that the behaviour is funny, since this does not follow
from any of its saliences. To understand, he must experience the behaviour as funny
himself. In the same way, perceiving something as an i requires this very perception.
Arguing that the figure in question is a vertical line with a dot above it provides no
compelling reason. A vertical line with a dot above it could equally be a semicolon—
which also means that the modus ponens stated earlier would not be valid in any case.
Secondly, the examples suggest that the contents of emotions are neither proposi-
tions, nor just particular objects. It is obviously not particular objects as such which
are the targets of our emotions. Fear of gorillas, for example, is typically not fear of
dead or deadly sick gorillas, drugged gorillas, baby gorillas, or the like. Even specifying
vital properties of particular objects will not do—for we do also not fear, say, healthy,
See Crane (1992).

18
adult and aggressive gorillas as such, but only in situations where they appear to pre-
sent a danger to us (or others). Typically, we do not fear a (healthy, adult, and aggres-
sive) gorilla whom we believe to be securely caged. The situation changes significantly
when we suddenly realize that the door to the cage has been left open: only then will we
feel fear of the gorilla.
Altogether this indicates that, in experiencing an emotion, we ‘construe’ a situation
both in terms of its natural features and in terms of what we care about. This is similar
to, though not the same as, construing Wittgenstein’s duck–rabbit as a duck at one time
and as a rabbit at another.19 Robert C. Roberts has therefore characterized the content
of an emotion as an ‘evaluative construal’; I have described it as ‘Gestalt-like’.20 Roberts
also noted that emotional construals, although they are not sensory perceptions, ‘have
an immediacy reminiscent of sens[ory] perception’.21 What he calls ‘immediacy’, I tried
to flesh out by arguing that the content of emotion does not enter inferential relations.
One of my central arguments for the non-inferentiality of emotion has been that,
like perceptions, and unlike inferential judgement and belief, emotions can persist in
light of better judgement and knowledge—‘can’ here meaning logical, and not just psy-
chological, possibility. Like the perceiver of the Müller-Lyer illusion who, despite his
better knowledge, cannot help but see the two lines as being of different lengths, the
subject of a persisting—or, if you prefer, ‘recalcitrant’—emotion need not revise either
of his two mental states in order to avoid contradiction. By contrast, it would clearly be
contradictory to continue to judge that ¬p in light of the better judgement that p, the
reason being that inferential relations holding between judgements include obvious
logical relations such as, when you judge that p, you must also judge that ¬(p ˄ ¬p). This
contrast to inferential judgement and belief is also evidenced by the fact that neither
recalcitrant perceptions nor recalcitrant emotions commit the subject to Moorean
sentences like ‘p but I believe that ¬p’. It is not (Moore-)paradoxical to say ‘The two
lines are the same length but I see one line as longer than the other’. Nor is it paradoxi-
cal to say ‘The gorilla is harmless but I feel fear of him (see him as fearsome)’.
This is not to deny that there is conflict, but this conflict is, as I have dubbed it
elsewhere, ‘conflict without contradiction’.22 In order to now answer the question of
whether conflict without contradiction does involve irrationality in the case of emo-
tion, whereas it does not in the case of perception, let us, rather than relying on intui-
tion, examine the conflict in more detail.
To qualify as cognitive conflict in the first place, conflict without contradiction must
be conflict in content. Provided that emotional contents are non-propositional con-
struals, this raises, first of all, the question of whether they can at all conflict with the
propositional content of better judgements. The answer is that emotions present evalu-
ative features of the world in a different mode than value judgements do. In feeling fear
19
See Wittgenstein (2001 [1953]: part II, §xi).
20
See Roberts (2003: 69ff.); see also Roberts (1988) and Döring (forthcoming a).
21
Roberts (2003: 75). 22 Döring (2009).
of a gorilla, you do not simply see him as ‘fearsome’; that is, in terms of a concept that
could be expressed by using the predicate ‘fearsome’. Rather, you see the gorilla’s aggres-
sive eye, his terrifying set of teeth, and perhaps you see him already in a posture ready
to attack you. That is, you see the gorilla in terms of the salient features which, taken
together as a whole, constitute its fearsomeness for you. Nevertheless, both contents
are about the same evaluative feature of the situation—fearsomeness—and both are
assessable for correctness: although the content of emotion is not propositional, it is
representational.
It follows that, when fear of a gorilla persists in light of the subject’s better judgement
that, in the given situation, the gorilla is not fearsome, there is cognitive conflict in con-
tent. How could the subject still not be irrational?
To answer this question we may again refer to the different modes in which the goril-
la’s fearsomeness is presented. The judgement that the gorilla is not fearsome has a
content which is assessable as true or false because its content is regarded as true by the
subject. It is a propositional attitude which consists, precisely, in regarding a proposi-
tion as true. By contrast, the content of an emotion is not a proposition, and hence
the mode in which we relate to this content is not an attitude towards a proposition.
Instead, in experiencing fear of the gorilla, the gorilla appears to be fearsome, where
this is an appearance of truth: it seems to the subject that the gorilla is in fact fearsome,
whether or not he would affirm the truth of his emotion’s content in judgement.
Elsewhere I have tried to account for the difference between an appearance of truth
and a mere appearance by pointing to the fact that, from a meta-level of reflection
and regulative self-guidance, we, as rational agents, treat our emotions as a mental
subsystem which provides us with information about our environment and which
is generally reliable in doing so.23 To borrow Karen Jones’ illuminating terms: being
‘reason-responders’, we treat our emotions as generally reliable ‘reason-trackers’, in
order to step in only when necessary.24 In the default mode it does not occur to us to
ask whether an aggressive gorilla whose cage has been left open really is fearsome:
our occurrent conscious fear puts forward this content as true, thereby enabling finite
embodied creatures like us to respond appropriately to a complex and risky world at a
speed that we would not be able to achieve without our emotions: we perceive certain
things in our environment as salient features of danger, and fear immediately primes
us to take protective action; that is, without having to wait for inference and delibera-
tive reflection. Appearance of truth is a characteristic also of the content of perception:
in the default mode, it does not occur to us to ask whether the conditions under which
we perceive (such as the lighting conditions) are normal.
The possibility of conflict without contradiction is then explained as follows: con-
flict without contradiction between emotion and judgement arises when an emotion
persists in spite of the subject’s better judgement. In the default mode we rely on our
emotions, but once we suspect that an emotion deceives us we are ready to leave the
23
See Döring (2010: 293‒5). Jones (2003: 190).
24
default mode and to switch into a different mode. When an emotion fails to pass the
tribunal of deliberation, we withdraw our confidence in it. It comes to conflict with-
out contradiction when the emotion persists. The subject does not contradict him-
self because he only regards his judgement’s content as true, whilst the content of his
emotion merely appears to be true to him. Still, there is cognitive conflict because, in
being accepted as mental states that reliably represent actual features of the world, the
emotions are seen by the subject as being in the service of truth and knowledge, albeit
in a different way than judgements are. As such, this is not irrational, but as I argued
elsewhere, the practical enkrateia requirement may require the agent either to keep his
emotion under control or to bring his judgement into line with his emotion.25
IV.
We are now in a position to reject the Helm objection. Put in terms of conflict with-
out contradiction, Helm’s worry is that a perceptual account of emotion explains away
not only the contradiction but also the conflict. According to him, there is no cogni-
tive conflict at all in the case of recalcitrant perceptions. Because its content is genu-
inely and fully repudiated and thus regarded as a mirage, comparable to the attitude
we take towards the content of an imagination, Helm says, it is not at all irrational
to have a recalcitrant perception. By contrast, emotions are seen to involve cognitive
conflict which, in Helm’s view, seems to imply that the subject is irrational. Helm tries
to explain this contrast by ascribing the attitude of ‘assent’ to emotion which is denied
of perception. Helm’s emotions thus look more like judgements than like perceptions,
but he does not want to say that they are judgements: the assent in question is intro-
duced as a distinctively emotional kind of assent.
It is clear from what I have said earlier that I agree with Helm in that emotions must
not be mere appearances. But it does not follow from this that each particular emotion
must involve an attitude of assent in addition to the attitude, or rather the mode of
regarding as fearsome, offensive, admirable, annoying, enviable, or whatever evalua-
tive property defines the emotion and thus, in Scholastic parlance, is its ‘formal object’.26
The formal object is the property which the subject must necessarily ascribe to the
object of a mental state in order to render the state intelligible. In the case of belief, the
subject must necessarily regard its content as true. In the case of fear, the subject must
regard his emotion’s target as fearsome, in the case of admiration as admirable, in the
case of anger as annoying, and so forth. My point now is that, beyond this, there is no
distinctively emotional kind of assent—and is not even needed. As I have argued, emo-
tions are not mere appearances but appearances of truth: we, as ‘reason-responders’,
25
See Döring (forthcoming b). In fact, I think that bringing recalcitrant emotions into harmony with
better judgements would better suit Tappolet’s view than the claim that recalcitrant emotions ought to be
brought into line with better judgements.
26
See Kenny (1963: 189); Lyons (1980: 99); de Sousa (1987: 122); and Teroni (2007).
generally rely on our emotions being ‘reason-trackers’—which also explains why emo-
tions come with the epistemic inclination to make a corresponding judgement. This
is all the assent needed. It is assent not to each particular emotion but to our emotional
system as a whole, and it is not distinctively emotional.
The same kind of assent, I claimed, we give to our perceptual system. This means,
first, that in both cases—that of recalcitrant emotions and that of recalcitrant percep-
tions—there is cognitive conflict between the two contents involved. And yet I insist,
secondly, that in neither of these cases is the subject irrational because, thanks to the
different modes involved, he does not contradict himself. Pace Helm, cognitive con-
flict between contents does not necessarily imply irrationality of the subject. Both the
content of a recalcitrant emotion and that of a recalcitrant perception have the appear-
ance of truth, but as long as the subject does not make a corresponding judgement—
that is, regard a corresponding proposition as true—he is not irrational.
I think that recalcitrant emotions are intuitively sensed as irrational because, in
addition to cognitive conflict, they typically involve practical conflict. Typically,
we are poised to act out of our emotions—‘act’ being here understood in a broad
sense so as to include irrational and even so-called ‘arational’ expressive actions.
This brings the avid hiker or the oversensitive person into trouble. These persons
feel torn between judgement and emotional experience and should perhaps see a
therapist, because their recalcitrant emotions interfere with the reasoned pursuit
of their goals by moving them so as to act against those goals. But this means that,
strictly speaking, the action is irrational—not the emotion itself. Behavioural ther-
apy should suffice—especially so as it may always turn out that the emotion, and not
the judgement, gets things right.
References
Arpaly, N. (2000). ‘On Acting Rationally Against One’s Best Judgment’. Ethics 110: 488–513.
Arpaly, N. (2003). Unprincipled Virtue: An Inquiry into Moral Agency (Oxford: Oxford University
Press).
Arpaly, N. and Schroeder, T. (2012). ‘Deliberation and Acting for Reasons’. Philosophical Review
121: 209–39.
Bennett, J. (1974). ‘The Conscience of Huckleberry Finn’. Philosophy 49: 123–34.
Brady, M. S. (2007). ‘Recalcitrant Emotions and Visual Illusions’. American Philosophical
Brady, M. S. (2009). ‘The Irrationality of Recalcitrant Emotions’. Philosophical Studies
145: 413–30.
Crane, T. (1992). ‘The Nonconceptual Content of Experience’. In The Contents of
Experience: Essays on Perception, ed. T. Crane (Cambridge: Cambridge University Press),
136–57.
Döring, S. A. (2007). ‘Seeing What to Do: Affective Perception and Rational Motivation’.
Dialectica 61: 363–94.
Döring, S. A. (2009). ‘The Logic of Emotional Experience: Noninferentiality and the Problem of
Conflict Without Contradiction’. Emotion Review 1: 240–7.
Döring, S. A. (2010). ‘Why Be Emotional?’ In The Oxford Handbook of Philosophy of Emotion, ed.
P. Goldie (Oxford: Oxford University Press), 283–302.
Döring, S. A. (forthcoming a). Gründe und Gefühle. Zur Lösung ‘des’ Problems der Moral (Berlin
and New York: de Gruyter).
Döring, S. A. (forthcoming b). ‘Wide Process Enkrasia: How to Revise One’s Best Judgement
in Light of a Reason-Tracking Emotion’ (paper presented at the conference ‘Dimensions of
Normativity’, Frankfurt).
Greenspan, P. S. (1981). ‘Emotions as Evaluations’. Pacific Philosophical Quarterly 62: 158–69.
Greenspan, P. S. (1988). Emotions and Reasons: An Inquiry into Emotional Justification
Helm, B. W. (2001). Emotional Reason: Deliberation, Motivation, and the Nature of Value
Hume, D. (1978 [1739/40]). A Treatise of Human Nature, ed. L. A. Selby-Bigge and P. H. Nidditch
(Oxford: Clarendon Press).
Jones, K. (2003). ‘Emotion, Weakness of Will, and the Normative Conception of Agency’. In
Philosophy and the Emotions, ed. A. Hatzimoysis (Cambridge: Cambridge University Press),
181–200.
Kenny, A. (1963). Action, Emotion and Will (London: Routledge).
Lyons, W. (1980). Emotion (Cambridge: Cambridge University Press).
Montaigne, M. de (2007 [1588]). Les Essais, ed. J. Balsamo, C. Magnien-Simonin, and M.
Magnien (Paris: ‘Pléiade’, Gallimard, based on the 1595 published version).
Roberts, R. C. (1988). ‘What an Emotion Is: A Sketch’. Philosophical Review 97: 183–209.
Roberts, R. C. (2003). Emotions: An Essay in Aid of Moral Psychology (Cambridge and
New York: Cambridge University Press).
Tappolet, C. (2003). ‘Emotions and the Intelligibility of Akratic Action’. In Weakness of Will and
Varities of Practical Irrationality, ed. S. Stroud and C. Tappolet (Oxford: Clarendon Press),
97–120.
Tappolet, C. (2012). ‘Emotions, Perceptions, and Emotional Illusions. In Perceptual
Illusions: Philosophical and Psychological Essays, ed. C. Calabi (Basingstoke: Palgrave
Macmillan), 207–24.
Teroni, F. (2007). ‘Emotions and Formal Objects’. Dialectica 61: 395–416.
Wittgenstein, L. (2001 [1953]). Philosophical Investigations (Oxford: Blackwell).
9
Surprise
Adam Morton
1. Introduction
It is not surprising that unlikely things often occur. Fair coins come down heads four
times in a row, healthy people drop dead, unrelated colleagues develop the same rare
hereditary disease. At any rate it should not be: the emotion of surprise here would be
irrational. But the individual unlikely events do provoke real surprise, which can be a
pretty intellectual or a pretty visceral emotion, and the reasonableness of the reaction
needs a more subtle treatment. Surely your eyebrows rise when the coin lands heads
for the fourth time, even though you know such things will happen one time in sixteen.
Surely you shudder in shock when the same nasty random fate befalls several people
you care for. The main purpose of this piece is to explore the emotion of surprise as
an emotion, and to address the question of its rationality. A secondary purpose is to
discuss the value and disvalue of surprise: how we both desire and fear the unexpected.
2. The Emotion

You open your door to go out, and there is someone just standing there, her hand about
to knock. You leap back, your breath comes fast and your heart pounds; it takes several
seconds to regain your composure. ‘You startled me’ you say.
You find that the president of your university, a well-respected scholar and
much-admired leader, has been blackmailing colleagues to fund his drug habit. You
are shocked.
You go into your lab one evening and find that six of the ten mice that were showing
symptoms of an acute viral infection are healthy and alert. You find this curious, and
you pay a lot of attention to the details of the mice’s surprising condition.
You log onto your bank account online, and find that the balance is appreciably
higher than you expect. You immediately look for unexpected transfers in, and see if
regular debits have all been paid.
These are very basic emotions—all reactions to changes in the environment. Not all
are unpleasant. They are all reactions to unexpected information—reactions that defend
138 Adam Morton
against possible new threats and that gather more data. That the link between new-
ness and information-gathering is a very basic human attribute is suggested by the
work on the reactions of infants and newborns to novelty, pioneered by Robert Fanz in
the late 1950s and famously applied to ever-younger babies by Elizabeth Spelke from
the 1980s (Gopnik and Melzoff 1997; Talbot 2006). Infants look more, and show other
signs of increased attention, to situations that depart from established patterns or from
expectations that we can hypothesize to be innate. Perceptual phenomena such as the
Ganzfeld effect suggest that variety-hunger is built into the normal function of very
basic cognitive processes (Metzger 1930). As my examples suggest, the range of situa-
tions in which a violated expectation can lead to an emotion of surprise is very wide,
producing reactions from self-preservation, as when one is physically startled, to intel-
lectual inquiry, as when an interesting result stimulates one’s curiosity.
It is important to take situations in this whole range not only as unexpected hap-
penings, but as producing the emotion of surprise. For in situations like those I have
described, which span the range, the person enters a state in which resources are
summoned to meet the challenge, some of which go beyond the normal reaction to
falsification of a belief, and some of which may not be perfectly adapted to the par-
ticular situation. That is of the essence of an emotion: one finds a way to extraordinary
resources, but at the price of possible mismatch or over-reaction. Standing at the door,
looking at the old friend who happens to have arrived just as you opened it, you react
as if you had faced an assassin with an axe. (I once addressed the friend by name with a
friendly greeting, and then jumped back and became short of breath, a second later. The
emotion was out of sync with the thought.) Looking at the unexpectedly healthy mice,
you find yourself scouring your memory for ways that these might not be the same
mice, or in which some powerful antiviral could have been administered by mistake,
in much the way you would have if you were trapped in a burning building searching
for a way out and trying to remember anything you had heard about its construction.
At some level, research and self-preservation are not so different.
These are characteristic functions and effects of emotions, but surprise has familiar
common-sense emotional features also. I will mention three. It can be attributed on
the basis of facial configuration and bodily posture, though as with other emotions this
is an unreliable and context-sensitive thing. It can be expressed in language with a that
clause, though what is expressed is not a belief or a desire. You are surprised that your
president is a criminal; you are surprised that the mice are well. (In this respect it is
what Robert Gordon (1988) called a factive emotion.) And it comes with a feel, build-
ing in part on physical sensations, which besides experiencing ourselves we attribute to
others when we imagine their surprise empathetically. Imagining a person confront-
ing someone unexpected at the door we sense the breath getting tight and the pulse
racing. Imagining a person learning that an admired leader is a criminal we sense the
muscular tightening of outrage and the panicky vertigo of disbelief.
What visceral and cerebral cases of surprise have in common is a combination of
defensiveness and inquisitiveness: defence against new or novel threats, and inquiry
Surprise 139
into ways the situation might not be as it seems. Michael Brady (2011) has drawn our
attention to epistemic aspects of apparently practical emotions; notably, ways in which
they inquire into their own groundedness. Using his way of thinking, surprise asks ‘is
there anything here I have to know more about?’, and plays it safe until the question is
answered. The exact combination of defence and inquiry, and the detailed profile of
actions that are motivated, is not something that we can settle without a lot of data. It
is an important fact about our and related species, rather than a truth about the nature
of thinking agents.
3. Motivational Aspects

The Large Hadron Collider may simply confirm the standard model of fundamental
particles and forces. That would in a way be satisfying, but many physicists hope for
something less comfortable. They hope that it produces data that force them to change
the model, so that they come up with a more powerful and somehow deeper one. Just
perhaps, new data may prompt some young wizard to produce a new model that can be
grasped as intuitively compelling. (Though there is no reason why nature at that level
should respect the expectations of even highly evolved jungle and savannah creatures.
In some domains, lack of surprise would be surprising.) So we hope to be surprised: it
might be good for us. (See, for example, Butterworth 2013, where it is remarked that we
may have to ‘live with’ the absence of surprising results, and the editorial accompany-
ing the article where the absence of data challenging the standard model is referred to
as a ‘nightmare’.)
We often hope to be surprised. We do not go on entirely predictable boring vaca-
tions, and we do not want our friends to be script-driven cardboard cutouts. (One
aspect of this, not on the axis of this piece, is the Sartrian desire that other people be
other people, real independent contrary people. Up to a point!) Of course we hope
to be surprised when our expectations are low, with feeble students and hopeless
offenders. But there is also a motive towards surprise for its own sake. We want our
lives to contain an element of the unexpected. There are many sources of this. There
is the practical unpredictability of human action, there is the depth and variety of the
physical world, and there is the disparity between the combinatorial complexity of
possible situations and the limits of our thinking power. It is an exciting mathemati-
cal surprise when axioms suddenly show that they entail the opposite of what was
expected.
Our desire for surprise is typically human; we want varied and unpredictable intel-
lectual and social lives. But it is not only a human phenomenon. Boredom affects many
other creatures, from cats kept alone in small apartments to tigers in zoos. Most wild
animals probably have all the surprise they need, with the changing requirements for
staying fed without becoming food. The greater the environmental variation to which
a species is adapted, one expects, the greater the need for surprise as part of that spe-
cies’ telos.
140 Adam Morton
Humans are complex creatures adapted to variable environments, and thus we can
expect that they will need a fair degree of surprise in order to flourish. We get bored if
things are too predictable, and we also want signs that they are real, that we are strug-
gling to make objective accomplishments, coming to terms with the actual causes
of our experience, and interacting with people whose behaviour does exhibit their
unfaked attitudes. Delusions, simulations, and acting, however pleasing, go against a
deep human impetus. As a result, we are reassured when the details go against our
expectations, in part as a sign that they are not generated by these very expectations.
Thus a wise person—in fact a typical person—will want that there be a good supply of
surprising events in her life, even at the price that some proportion of the surprises be
unwelcome.
4. Surprisingly Unsurprising

Sometimes an event is unexpected but unsurprising. Examples can be found in
1960s epistemology (Goodman 1973). Suppose that coins enter your pocket in a
generally random way that is neutral with respect to their composition: silvery
coins and coppers are likely to get there in their proportion among coins. In fact,
for the sake of the example suppose that proportion to be 1:1. You know there are
four coins in your pocket. You reach in and pull out a copper. You do again, and
again. Now you know that at least four of the coins that were in your pocket were
copper. The probability that they were all copper has increased from 1/16 to 1/2,
yet, against the orthodoxy that developments that increase the probability of a
claim are evidence for it, you have not acquired evidence that all the coins were
(or are now) copper. So if a fourth reach brings out a silver coin, you should not
be surprised. (You may in fact be surprised, superstitiously, as you might be if for
example three friends who walked under a ladder failed a subsequent exam, and
then a fourth did not.)
An interesting case is that of events for which one has no good evidence, but which
one has come to expect. If you find that servers at your coffee shop are always wear-
ing blue shoes, though there is nothing else uniform about them, you may well find
yourself expecting it, and subject to surprise if one has black or red shoes. Suppose now
that one of them confides in you. ‘We do it to annoy the boss. Hard to explain, but it
drives him wild and there’s nothing he can do about it. He’s been nicer lately, though,
so we’re just vaguely considering calling it off.’ Now you do have evidence that it is not
just a series of coincidences, though interestingly the evidence—though evidence for
the regularity—may lower your subjective confidence in its probability. So now when
you look down when reaching for an espresso and see that the barista’s shoes are green,
should you be more or less surprised?
Improbable events are often not surprising. If one in a thousand five-year-old light
bulbs burns out in any given month, then it is less than amazing that the light in your
bathroom has gone, though it is a nuisance. There is nothing that you have to check or
Surprise 141
take account of, since you always knew it would happen sooner or later. If you had been
perfectly prepared you would have had the spare bulb ready and have thought out how
to place the chair safely under the socket for the replacement. Contrast this with your
reaction when you replace the bulb and it begins to blink in a Morse code fashion, as
if trying to tell you something. Is it trying to tell you something; is it likely to explode;
should you unscrew and replace it or wait and see what happens? Or suppose that the
new bulb is a dud, flashes, and then goes dark. This is surprising even if it is more prob-
able than that your long-serving bulb should fail on this particular day. In these last
two cases you have to revise your expectations of what may happen and why.
It is here that questions of rationality become delicate. Is it irrational to be surprised
when your lottery ticket wins, your light bulb blows, you succumb to a disease that mil-
lions have? Surprise is certainly not unusual. The aspect that is irrational, or at any rate
unhelpful, is that no reappraisal of possible causes and their possible effects is called
for. (As becomes clearer later, I think that in hard cases we should withhold ‘rational’,
‘irrational’ in favour of more nuanced labels.) You may now be rich, or in the dark,
or facing death, and there are emotions suited to these situations. But the probability
of winning such a lottery, the causes and frequency of bulb failure, and the chance of
someone with your profile succumbing to that disease remain as they were. Nothing
has become more mysterious by happening now. The likelihood of these things does
not need to be better understood, though you may need to gather more information
and think out new precautions because of what has happened. You could sensibly be
alarmed but not surprised.
Seen this way, the lack of surprise in the copper coin and barista shoe examples
seems right. You knew all along that there was a 1/16 probability of four coppers a priori
and 1/2 when three have been drawn. You knew originally that you had only a guess-
work expectation that the shoes would be blue, so that green shoes did not upset any
fixed belief. And then once you had been given the explanation, green shoes have the
power to prompt questions: have they made friends with the boss or has he become
too powerful to question; is there a chance that he will come storming in and hot coffee
will fly? So it makes sense to say that in the enlightened situation green shoes make a
smaller dent in your expectations intellectually but are a greater prompt to the emo-
tion of surprise.
Similar considerations allow us to see how surprise at those same things might in
similar circumstances be sensible, and guide the person to inquiries she should be
undertaking. You may not have calculated the probabilities in advance, and have just
taken four heads in a row or four coppers as ‘very unlikely’. Then the event may make
you do the calculation you could have done earlier. Or you may have never thought
about the half-lives of light bulbs, so when you turn the switch and nothing happens
your reaction is to something completely unanticipated. Somewhat differently, you
may have ruled out the possibility of winning the lottery, or of all four coins turning
out to be copper, simplifying your epistemic situation by throwing out the evidence on
which your all-things-considered belief is based. Then when that belief proves false,
142 Adam Morton
you no longer have that evidence to make it unsurprising and so you react as if to a
mystery. You have to restore the information you had suppressed.
The reaction has a kind of rationality in these cases: given your neglect of relevant
considerations earlier, or given your labour-saving collapse of probability to certainty,
it follows a routine that generally promotes the interests of the organism. That ear-
lier corner-cutting may itself be an efficient reaction to limited time and processing
power, or it may not be. So there are many possibilities. There are simply helpful ways
of reacting that will promote a person’s well-being in most likely circumstances. There
are second-order helpful reactions that compensate for earlier deviations from proce-
dures that would have worked out well. Among these there are those that compensate
for deviations that were simply faulty: slips, omissions, glitches. And there are also
reactions that compensate for deviations from ideal procedure that given the nature or
situation of the person represent acceptable trade-offs between possible outcomes and
costs of thinking and investigating.
Epistemologists know well that to label thinking as rational or not brings compli-
cations such as these into view. In the philosophy of emotion the complications are
forced on us when we pay attention to surprise. Surprise fits here because it is the emo-
tion that prompts further inquiry, the reaction to things that need explanation.
5. Contrastivity
Events are often surprising and unsurprising, as we have seen, and both welcome and
unwelcome. This is no paradox, but merely the effect of context and, especially, con-
trast. It is surprising that there is a run of seven heads rather than a more evenly dis-
tributed series; it is not surprising that the coin comes down heads rather than tails on
this occasion. It is welcome that your cancer is a treatable type rather than an untreat-
able one; it is unwelcome that you have cancer rather than a stomach ache. Sometimes
the two contrasts coincide. Suppose there is a consolation prize given at the same time
as the jackpot to a very small proportion of the people who have bought a thousand
lottery tickets but have never won. You are one of these poor people. You learn that
you have won the consolation prize. It is surprising that you have won the consolation
prize rather than losing the value of your ticket yet again, and it is not surprising that
you have won the consolation prize rather than the jackpot. Similarly, it is good news
that you have won the consolation prize rather than losing the value of your ticket yet
again, and bad news that you have won the consolation prize rather than the jackpot
(see Driver 2012).
In traditional cultures people usually described themselves as not wanting surprises.
(See the introduction to Giardina 1993, and a host of bad connotations for secondary
meanings of novus and cognates in Latin dictionaries. For another ancient culture see
chapter 4 of Almerding 2011.) But people then were as susceptible to boredom as we
are, and had as deep a need for stimulation. The difference is in the default contrasts,
Surprise 143
I think. If little happens on a given day, and you think the alternative is plague or mas-
sacre, then of course you will value the lack of surprise. If the alternative is exciting
developments and thought-provoking puzzles, then you will disvalue the lack. A person
who meets every new day with wide-eyed wonder is implicitly contrasting it to conceiv-
able more boring days (and this may sometimes be a feat of imagination). A person
who remains blasé in the face of the most dramatic occurrences is implicitly contrasting
them with the really interesting things that might have happened. (Imagination over-
powers perception: the person does not see present marvels as marvellous, blinded by
the force of what might be there instead; Morton 2013.) It is all in the contrastive rather
than: what in the usual context is the alternative one will naturally think of?
I previously described surprise as the inquiry-prompting emotion that asks for
explanations. Explanations are typically contrastive too (see Hitchcock 2012), which
fits with the idea that in being surprised at something one is asking why it, rather than
some contextually determined alternative, should have occurred. But there is another
contrast-themed connection between surprise and explanation. Surprise sets the task
for satisfactory explanation. If one is surprised that the crops have been half-successful
rather than giving a full harvest, then one will want an explanation of why this year’s
harvest is half rather than full. If one is surprised that they have been half-successful
rather than failing, then one will want an explanation of why the harvest is half rather
than none. The cultural contrasts built into intuitive surprise can be ignored in an
explanatory project, of course, but this takes deliberate effort. The influence can go in
the other direction too, as when one learns that some events taken with a particular
contrastivity have no explanation—why the fair coin lands heads rather than tails on
this particular toss—and as a result comes not to be surprised by them.
6. Life and Death

In earlier work I have argued that when rationality is slippery we can sometimes frame
our questions in other terms, particularly by asking what virtues are needed by some-
one who thinks in one manner or another (Morton 2012). I am not going to repeat the
arguments for this point of view, but end this essay by asking of some ways of being
surprised what virtues can well accompany them.
All humans are mortal, and you are human. But it comes as a surprise to many peo-
ple when their death comes in sight. Why now, why this disease? But the situations of
an eighty-five-year-old with a failing heart and of a fifteen-year-old with leukemia are
different. The first is no surprise.
For some people, life will go better under the illusion of immortality. Their conversa-
tions with the spectre would oppress them; they would not make good medium-term
plans. So a minimum virtue for lack of ‘unreasonable’ surprise is the ability to keep
perspective and proportion in the face of scale-changing considerations, and perhaps
a degree of immunity to framing effects. (I have known highly intelligent people who
144 Adam Morton
lacked these virtues, and knew it.) Given these, a person’s setting themselves in the
way of surprise, if forced to face the fact that they will die sooner or later (or that their
children will have less than total admiration of them, or that their loved ones will see
some of their flaws) will work to their advantage. What about the less mysterious, but
still puzzling, surprise at the fact that some particular not very unusual death is immi-
nent? It ought to be like surprise that an old light bulb has blown, or that this time all
the coins are copper. But it can seem like surprise that the insurance number randomly
assigned to you begins with your birth date, or that all of your close friends have the
same middle initial.
I think the death case is in fact rather like these latter cases. They are all of kinds that
are rarely as unusual as they seem at first sight, and so one’s reaction to them is likely
to be based on shortcut estimates and heuristics. The reaction thus makes sense in a
second-order way, as a good reaction to the fact that one’s basic reaction is not ideal.
Its advantages stem from the fact that one has not previously digested some relevant
considerations. Would it be better if one had digested them? Perhaps, for some peo-
ple—those who could make good choices and enjoy their lives in the face of these gen-
erally a priori considerations. For others not: the lives of some others will go better, for
a short while, if they react as if some monstrous predator had chosen them, against all
the odds, as victim.
Corresponding things can be said of good news. As you awake from the anaesthetic
you may be surprised that the 70% chance of success has been achieved. Oh, I am here.
Had you really thought that what was only 30% probable was more likely to happen?
Surprise and relief are hard to disentangle here: relief has a related set of functions and
is most easily activated when the feared outcome was taken to be more probable, so to
get to feel it—make the right offerings and change the right priorities—it may help to
keep one’s grasp on what was likely a bit fuzzy. The moral is the same: depending on
who you are, different ways of thinking what to expect may better fit your intellectual
and emotional constitution, and with them different occasions when surprise is a help-
ful emotion for you.
One important and modern virtue that emerges from the examples I have been
using is the virtue of reacting appropriately to the randomness of the world. If you
see that there is no deep reason for much of what happens—how long a run of heads
continues, whether you have a road accident or contract a fatal disease, whether this or
that opportunity opens up for you in life—then you may succumb to fatalistic lethargy
or you may take life as a mixture of opportunities to be taken and dangers to be fore-
stalled or endured. And you will see that whatever happens could have been better and
could have been worse, so that it is both a good surprise and a bad one, and both sur-
prising under one description and not under another. You can take it as both; you can
take control of the contrastivity. Thus the virtue in question is one of framing events
with suitably contrasting ones so that you can react to them as surprising or banal, and
as good news, bad news, or no news at all, in ways that steer between keeping life inter-
esting and keeping it safe.
Surprise 145
If events are framed in this way, a sophisticated higher-order emotion becomes

possible. You can see the value of unwelcome surprises, both in the simple pragmatic
terms that have shaped most of my discussion and as licences to direct at them other
evaluative emotions. You can react to the occasional item of bad news with disappoint-
ment or annoyance while simultaneously being glad, even relieved, that your life has
the variety of the occasional less than optimal moment. (More than occasional, and
something super-human might be needed.) Your reactions to the particular event and
the general pattern it exemplifies are different. Seeing the presence of randomness in
most important developments is important in this, I think, because it prevents one
from thinking that particular events have to be tied to general patterns. Your annoy-
ance at this particular frustration does not have to undermine your satisfaction that the
project it is part of is proving unpredictable and challenging. So, if this is right, there
is a systematic connection between a randomness-appreciating evaluative attitude—
which is most easily expressed in terms of what you find surprising or unsurprising—
and a capacity to direct opposed emotions towards intrinsically linked objects. They
both depend on not reading too much into things one by one.
References
Almerding, A. (2011). Moral Exemplars in the Analects: The Good Person is Just That
(London: Routledge).
Blaauw, M. (ed.) (2012). Contrastivism in Philosophy (London: Routledge).
Brady, M. (2011). ‘Emotions, Perceptions, and Reasons’. In Morality and the Emotions, ed. C.
Bagnoli (Oxford: Oxford University Press), 135–49.
Butterworth, J. (2013). ‘Instant Expert: Higgs Boson’. New Scientist 2933 (7–13 Sept.): i–vii.
Driver, J. (2012). ‘Luck and Fortune in Moral Evaluation’. In Contrastivism in Philosophy, ed. M.
Blaauw (London: Routledge), 154–72.
Giardina, A. (ed.) (1993). The Romans (Chicago: University of Chicago Press).
Goodman, N. (1973) Fact, Fiction and Forecast, 3rd edn. (Indianapolis, IN: Bobbs-Merrill).
Gopnik, A. and Meltzoff, A. (1997). Words, Thoughts and Theories (Cambridge, MA: MIT Press).
Gordon, R. (1988). The Structure of the Emotions (Cambridge: Cambridge University Press).
Hitchcock, C. (2012). ‘Contrastive Explanation’. In Contrastivism in Philosophy, ed. M. Blaauw
(London: Routledge), 11–34.
Metzger, W. (1930). ‘Optische Untersuchungen am Ganzfeld’. Psychologische Forschung 13: 6–29.
Morton, A. (2010). ‘Emotion, Virtue, and Knowledge’. In The Oxford Handbook of the Philosophy
of the Emotions, ed. P. Goldie (Oxford: Oxford University Press), 385–400.
Morton, A. (2012). Bounded Thinking: Intellectual Virtues for Limited Agents (Oxford: Oxford
University Press).
Morton, A. (2013). Emotion and Imagination (Cambridge: Polity Press).
Talbot, M. (2006). ‘The Baby Lab’. The New Yorker (4 September): 90–101.
10
Emotions Fit for Fiction
Greg Currie
It is uncontroversial that fictions of any interest are multiply interpretable. But those
multiple and sometimes conflicting interpretations rarely arise simply through the
judicious weighing of evidence as to what the text means. At least in part, interpre-
tive disagreement is founded on disagreement of response: imaginative and emotional
response, that is. But just as multiple interpretability does not mean that all interpreta-
tions are good, the fact that there is a plurality of merited responses to the work does
not make every response an appropriate one. Here I examine some of the simpler and
more obvious ways to identify appropriateness in emotional responding. Much will
depend on the idea that fictions do not offer us alternative worlds, but rather alterna-
tive representations of the world.
1. Constraints on Imagining

If choosing what to believe is possible at all, it is much harder, and much less usual,
than choosing what to imagine. But imagining is sometimes guided by interests and
purposes which impose their own constraints. While reading Tolstoy’s description of
the death of Anna Karenina, imagining her boarding a space ship for Mars will be a
wrong, though possible, imagining. Wrong by what standard? By the standards for
appropriate engagement with a narrative fiction. Tolstoy represents Anna as doing,
and being able to do, certain things: boarding a space ship is not one, and imagining
her doing that disengages us from the system of representations that is Tolstoy’s story.
And within the class of appropriate imaginings there are constraints on how one ima-
gines. It would be legitimate, as part of an imaginative exploration of the possibilities
open to Anna, to imagine pathways different from the ones she takes. I may imagine
her deciding, at the station, to accept a less than ideal relationship with Vronsky, or
a reconciliation with Karenin; that is one way to make vivid the consequences of the
choice she actually makes.1 But the story represents her as not doing those things, and
1
These options might not be dramatically very satisfactory, but that is another issue. On the complexity
of the relations between what is true in a fiction and what it is appropriate to imagine, see Walton, ‘Mind the
Gap’ in his (2014).
Emotions Fit for Fiction 147
one’s imaginative project must be of a kind which reflects the difference between what
might happen and what does happen.2 How imagining registers this we can only con-
jecture; perhaps there is a stage of cognitive processing where some imaginings are
marked with a label: this is story content; unmarked imaginings are assumed not to so
correspond. An imagining can be wrong by being marked when it ought not to be, or
vice versa; one may imagine something as story content when it is not, or fail to imag-
ine it that way when it is. We are not required to register in this way everything that is
story content: there is too much, and much of it is trivial. But we need to register the
narratively salient elements; failure to mark them would constitute failure to follow the
story.3
The marked/unmarked distinction concerns the manner of imagining, not its
content. Imagining that Anna commits suicide does not become marked by being
translated into imagining that Anna’s suicide is part of the story. That would be a self-
defeating move, since it is not part of story content that Anna’s suicide is part of this
or any story. That is why ‘attaching a label’ is at least a suggestive description of what
is going on; the Minister does not change the content of your memo when she adds a
label ‘Action this day’, though she affects the role that content plays in the Ministry’s
information system. Marked imaginings play a role that unmarked ones do not: ena-
bling us to track what actually happened, according to the story.
Sometimes it is hard to know what imaginings should be marked; I may imagine,
while reading, that Professor Peacock is the murderer, but that is only my conjecture
until Poirot gives his verdict. Sometimes we are never sure how to mark: the story
permanently confuses us on some important question. We may even decide that no
imagined resolution of some important question should be marked; the story is inde-
terminate.4 We may resolve the story to our own satisfaction, but merely imagining
things in the story to go that way will not be enough to make them marked; we have
to add the marker ‘this is story-content’, at which point we have wrested control of the
story from the author and have become the authors of a continuation.
Ambiguity, indeterminacy, and the possibility of pirated continuations are interest-
ing effects which occupy theorists of literature a good deal. But they are deviations
from the standard imaginative relation we have with fiction, which consist of our
imagining what we recognize to be part of the story content, because we recognize that
it is intended to be part of the story content. Without that background of concordance
between author and reader, our relation to the fiction falls apart, and it is only against
this background that other, non-concordant imaginings are appropriate.
We have seen one example of the idea that there are different ways to imagine:
marked and unmarked. Here are some other ways—by no means an exhaustive list.
2
It would be a different matter to imagine the proposition ‘Anna might have returned to Vronsky’; that
proposition is (as I read it) part of story content.
3
See Walton (1993: 39–40).
4
All stories are indeterminate on some questions; we bother to call stories indeterminate when they fail
to settle questions we expect them to settle.
148 Greg Currie
One’s imaginings may have narrative density: one does not merely imagine that Anna
throws herself under the train; one imagines it as a series of physically specific and
continuous movements, recruiting forms of visual, auditory, or somatic imagery to the
project. And one’s imaginings can be emotionally inflected in various ways. Imagining
Anna Karenina’s suffering I may feel empathic disturbance, experiencing something of
what I take to be her anguish. These ways of imagining are interconnected: empathiz-
ing with Anna is likely to be encouraged by a richly narrativized sequence of imag-
inings, by a visualization of the oncoming train and by a somatic representation of
the fall. Here I am particularly interested in emotional inflections of imagining, but
in many cases to be discussed the emotional inflection will depend—though I will not
always make this explicit—on it having a good deal of narrative density as well as per-
ceptual enrichment.
These emotional and other modulations of imagining are not generally chosen by
readers and viewers; we find ourselves responding in certain ways. But there are some-
times disparities between our tendencies to respond and the response we are intended
to have, as when we find a supposedly comic scene distasteful. This is one aspect of
what is called ‘imaginative resistance’, now recognized as a heterogeneous class of phe-
nomena.5 If, as people say, we are resistant to a fiction according to which female infan-
ticide is a good thing, this is not, I think, a matter of finding it difficult to produce some
minimal imagining of the idea. But we may have difficulty in taking our imaginative
engagement further. We might not want to engage in a project of narrative elabora-
tion—taking the idea in the direction the author wishes to take it. We prefer to close
the book and turn to something else. Or we may be unable to enrich our imagining,
because that requires imagining a way the proposition could be true. Moral truths,
like those of some other kinds, are true, if they are, because other things are true. What
could make it the case that female infanticide is right? Perhaps the story will go on to
tell us that females grow up to be mass murderers, in which case narrative elabora-
tion becomes possible. But if the fiction gives us no indication of any such relevant
difference from the real world, we are stuck with a bare imagining of the proposition,
and cannot integrate it with the intended background. And, most relevantly to our
topic, we may be unable to respond with the right kind of feeling if the story is one
which encourages us to feel pleased when females are successfully destroyed. Suppose
Tolstoy had said that Anna’s suicide was due to her having worn a less than fashionable
outfit; your ability to respond with strongly empathic emotion would be challenged.
I noted that the content of imagining is sometimes constrained. To some extent at
least, the manner of imagining is constrained also: imaginings which correspond to
what is true in the story should be marked as such. Are there constraints on the emo-
tional inflection of imagining? Here things are more complex. We shall see that there is
a variety of ways in which an emotion can be appropriate when it is part of the project
of engaging with a fiction. First of all I will concentrate on forms of appropriateness
5
See e.g. Walton (1994); Currie (2002); Gendler (2000); Weatherson (2004).
which are determined by the fit, or absence of fit, between the emotion and the story’s
events and characters. Then I will say something tentative about the difficult issue of
whether our emotions should be responsive to what authorial intentions are expressed
by the work.
2. Norms of Emotional Responding

Let us not forget, in all this, our emotional responses to the real world: we are frightened
or delighted by real things, love some real people and despise others, are made anxious
by certain prospects, and regret past failures. And these emotions are often appropri-
ate: it is right to fear the fearful, to delight in the delightful, to feel secure with some
people and wary of others, depending on what they are actually like. Commentators
have noted significant similarities between our emotional reactions to fiction and to
real life: we are frightened by creatures lurking in dark alleys, delighted by romantic
unions, and saddened by deaths, whether they are fictional or real. The emotions seem
to operate indiscriminately across that divide, and in that case we would expect the
constraints on emotions to be the same in both domains. It will turn out that they are
not.
A way for emotions directed at the real world to be appropriate is representational;
emotions represent their objects as being certain ways, and an emotion is appropriate,
in one sense, if it represents its object correctly. Fear for someone is associated with a
representation of them as in danger, fear of something goes with a representation of
that thing as dangerous; jealousy is associated with a representation of another as bear-
ing relations to a third party which they ought to bear exclusively to you. An emotion
fails to be appropriate when its associated representation misrepresents its object.6 It is
inappropriate to be fearful for someone who actually faces no threat, or to be jealous of
someone who in fact has no inclination to take a romantic interest in anyone else. Let us
call this the reality test. It says that an emotion should represent its object as it really is.
There are a number of ways an emotional response can be, or fail to be, appropriate,
and failure to be appropriate in one of them does not rule out success in another. Nor is
failure to be appropriate in one way always a bad thing; insisting on obeying the park-
ing regulations would be reprehensible if it impeded an important rescue by the fire
service. Passing the reality test is neither necessary nor sufficient for appropriateness in
an emotional response as we judge it all-things-considered. Rational beliefs are some-
times false beliefs, and one may falsely but rationally believe something which would
justify fear or jealousy. An emotion may misrepresent its object and be a required
response, all-things-considered; I should have been angered by Smith’s behaviour as
you reported it to me, given how badly he seemed to have acted and how reliable you
usually are, though it turned out that your account was misleading (we will have more
6
See Gilmore (2011).
150 Greg Currie
to say on this topic). And jealousy might be the wrong emotion to have, even when it
passes the reality test; given how badly you yourself had behaved, jealousy is a luxury
you were not entitled to even though your jealousy goes with a correct representation
of your partner as unfaithful.7
Still, the reality test is not a bad starting point when inquiring into the appropri-
ateness of an emotion; if my information about Smith is wrong, any anger I might
have had would have been misdirected, however rationally, and if your partner had
really been unfaithful your jealousy would have been free of one defect, even if it had
others.
Does the reality test apply to fictional cases? It does not, for two reasons. Firstly,
many objects of emotion in fiction simply do not exist—there is no truth about them
to constrain emotion.8 Second, where objects of emotion in fiction do exist—and some
do—the truth about them may constrain inappropriately. Consider an example given
by Stacie Friend; watching the movie JFK I am moved by Prosecutor Garrison’s self-
less devotion to truth, even though I believe (truly, let us suppose) that Garrison was,
in reality, a conspiracy fanatic who ruined the life of at least one person in his obses-
sive bid to prove government complicity in Kennedy’s assassination. As far as the real-
ity test goes, any emotion of mine experienced as I sit watching the movie and which
represents Garrison as admirable must be inappropriate. Doesn’t that miss the point?
Shouldn’t we be talking about how Garrison is represented in the film, not about what
he was really like? It is the film’s representation of Garrison as devoted to truth which
makes my admiration appropriate.9 For fiction we need, not a reality test, but a repre-
sentational correspondence test: my emotional response to the fictional character or
event is appropriate to the extent that the emotion is associated with a representation
which corresponds to how that character or event is represented as being in the story.10
This applies to wholly fictional characters like Sherlock Holmes as well as to real people
represented in fictions; my admiration for Holmes is (in one sense) appropriate just in
7
Perhaps jealousy is something we simply cannot help but feel, irrespective of circumstances. One’s
obligation in this situation would then be to minimize its duration and intensity and its effects on one’s
behaviour.
8
The ontology of fictional characters is disputed territory; what I think is generally agreed is that there is,
for example, no such person as Sherlock Holmes, no such bird as Tweetypie, and no such place as Tara. Those
who think that fictional characters exist at all do not think that they are people, birds, and places, but that
they are abstract objects of some kind. Being abstract they do not have the kinds of properties that would
make, say, contempt appropriate. The fictional character Karenin does not have the property of being hypo-
critically vengeful, for the character is an abstract object and not a person. On the metaphysics of fictional
characters see Thomasson (1999).
9
For qualification of this line of thought, see below.
10
Fictive emotions I take to be emotions directed at fictional characters and happenings. I am not assum-
ing at this stage that such states are not genuine emotions. That is an issue we shall come to. I do not count
as fictive emotions those emotions directed at fictional works, as when one is moved by the superb econ-
omy of the story-telling. Fictions are real things and emotions with fictions as objects are appropriate in the
same way that emotions which have our friends as their objects are appropriate. Being moved by the superb
economy of a story is not appropriate (in our current sense) if the story is not superbly economical in its
story-telling.
case he is represented in the stories as being admirable. The test cannot be whether he
is admirable, for he does not exist. A rule of reality constrains aptness for the emotions
of life; a rule of representational correspondence—correspondence between how the
emotion represents its object and how the fiction does the same—constrains fictive
emotions.11
All this talk of representations can become confusing. Emotions, I have said, rep-
resent.12 They represent things and events as being this way or that. That is true of
‘real’ cases, like my fear of snakes, and of emotions directed at fictional things—my
fear of Dracula. The difference is that my fear of snakes is (in the sense relevant here)
appropriate if snakes are dangerous, while my fear of Dracula is appropriate if he is
represented (in the novel, movie, or whatever) as dangerous. Note that this is not a dis-
tinction between what real and fictive emotions are about, but a distinction between
what it takes for them to be appropriate. I may genuinely fear a representation—a
painting I believe to be cursed, for example. What would make that appropriate would
be the painting having the property of being dangerous. But if I fear the painting in
The Portrait of Dorian Gray, my fear is made appropriate by the representation of that
representation as dangerous. What distinguishes fictive from real cases is that there is
always an extra level of representation in the specification of appropriateness.
There are ways to disguise this difference. One is to insist that the test is the same for
fictional and for real cases, with the difference merely that the test is relative to different
places. We find out whether it is raining in London by looking at London and whether
it is raining in Paris by looking at Paris, but that does not make for different tests. What
we need to do, the argument goes, is distinguish worlds; the actual world and the world
of the fiction. The emotions of life are appropriate if they correspond to the object
in the actual world, while the fictive emotions are appropriate if they correspond to
the object in the world of the fiction. It will be a theme of this essay that this is badly
mistaken.13 First, to ensure a thoroughgoing symmetry between fiction and reality we
would have to think of all these worlds as equally real—a metaphysically unattractive
commitment for most of us. Secondly, we have now to include among the worlds a lot
of places where things are massively indeterminate and where contradictions hold, for
11
Gilmore (2011: 475) puts the point in terms of a contrast between what is true and what is imagined: ‘If
the intentional state is a belief, then it is a constraint on the aptness of the emotion that the object has the
qualities in fact that the emotion presents it as having. If the intentional state is an imagining, the qualities
the emotion presents the object as possessing must be possessed by it as imagined for the emotion to be apt.’
Since I take it ‘imagined’ here means ‘appropriately imagined’, and what is appropriately imagined depends
on what is represented in the fiction, I think my way of stating the point is close to his. See also Schier (1983).
But what Gilmore says at p. 480 suggests that he thinks that the essential representedness of fictional states
of affairs does not make the assessment of a fictive emotion’s appropriateness so very different a matter from
the assessment of appropriateness for an emotion of real life; for this shows ‘only that the amount of poten-
tially determinate facts about the objects of our emotions differs across reality and fiction’. I am not sure that
this is right; is not this ‘different amount of facts’ a reflection of the deeper difference that, with fiction, we
never get below the level of what is represented to the level of what is true?
12
Or, more guardedly, are associated with representations.
13
See Walton (1993: 41–2).
152 Greg Currie
fictions are always incomplete in their specifications and sometimes contradictory; at

least for us realists about actuality, that puts fictional worlds in a quite different cate-
gory from the real world.14 Finally, it ignores what should be a rather obvious but signif-
icant point: that fictions are not alternative realities; they are representations of reality.
Fictions regularly represent real things, and in doing so misrepresent them: Henry V
was not in reality as Shakespeare represents him, but it is Henry who is represented in
the play. And fictions misrepresent one very big thing: the world. They say the world
is this or that way when in fact it (usually) is not. To this extent, fiction makers are like
liars and the mistaken. People who lie or who are mistaken are not people who are talk-
ing truly and reliably about some possible world other than the actual world; they are
lying or mistaken about the real world. Similarly, tellers of fictional tales are not people
talking about other possible worlds; they are misrepresenting reality. The difference
of course is that they are knowingly, non-deceptively misrepresenting it. And because
we collude in their (mis)representation, it is the representation that matters, not the
reality. So, when it comes to thinking about the appropriateness of emotion, the proper
contrast between reality and fiction is the contrast between how things are in reality
and how they are represented in the fiction. This is not like the contrast between two
places; it is like the contrast between a place and a picture of a place.
Another way to bring the two cases—emotions directed at real situations and emo-
tions directed at fictions—closer together is to argue as follows: the rule of represen-
tational correspondence (‘rule of representation’ let us say) which guides emotion in
fictional cases holds also in the real world; it is just that, in the real world, there are two
rules: this one and the rule of reality, and they sometimes conflict. As we have already
noticed, there is a sense in which it was right for me to be angry with Smith given how
you represented him in your testimony, though in fact you were, surprisingly, unreli-
able, and Smith had done nothing wrong. So the rule of representation applies in the
real world as well as in fictional cases, the difference being that in the real world one’s
response might satisfy one of these two rules and not the other and so be appropriate
by just one of two criteria, whereas for the fictional case the rule of representation is the
only rule. So it may be said.
In fact there is nothing that governs real-world cases of emotions that parallels
the rule of representation for fiction. The closest things governing real-world cases
is a rule of evidence—an emotion is appropriate when there is good evidence that its
associated representation is correct. And representations are sometimes evidence; if
the total evidence available to me, including the evidence of your testimony, suggests
that John behaved badly then my anger with him is in one sense apt, though it may
be that the evidence available to me is misleading and so, in another sense, not apt. It
might even be that the only evidence I have is your testimony, and, given your general
14
Nelson Goodman (1982: 162–4) also disbelieves in fictional worlds, but for different reasons. Lewis
(1983) defines truth in fiction in terms of truth in possible worlds, but denies that there is any world which is
the world of the fiction.
reliability, that representation makes my anger apt. But that is because your testimony
is good evidence, not merely because it is a representation of John’s behaviour. Merely
being a representation of something gives that representation no evidential weight
in the real world. The rule of representation that does apply to fiction is not a rule of
evidence. In the fictional case, how the story represents the situation is not merely
evidence for how things are in the story; it is constitutive of how they are. It is what
makes things that way in the story. Things are in a story as they are represented to be.15
In sum, for real-world emotions there is a rule of reality, and for emotions of fiction
there is a different rule: the rule of representation. The rule of representation does not
apply to real-world cases. What does apply there (in addition to the rule of reality) is
a rule of evidence, and representations (such as testimonial ones) sometimes provide
evidence.
Talk of evidence does allow us to find a parallel between real-world and fictional
cases. Just as it is often far from obvious what is true, so it is far from obvious what is
represented in a fiction. A reader can hope to conform to the rule of representational
correspondence only by weighing the evidence of what is represented. If I decide, at the
beginning of the story, that Professor Peacock is the murderer, and it turns out that he
is, I have impeccably conformed to the rule of representation; my emotions directed
at him—horror, contempt—represent him as the story itself represents him (let us
assume). But suppose my initial reasoning was based on a hopelessly wrong picture of
the evidence and my confused judgement of guilt was simply lucky; I conformed to the
rule of representation, but I should have done so by conforming to the rule of evidence,
which I did not do.
In all this we need to distinguish carefully between how the story represents its
characters and their doings and how the narrator represents them, for narrators can
be unreliable, and the representation of a narrator is only ever evidence for how things
are represented in the story itself. We could not coherently say that the narrator was
unreliable unless we thought that the narrator’s account of what happens is incon-
sistent with what happens according to the story, and what happens according to the
story is determined by how the work—the novel or whatever—represents things as
happening.16
Unreliable narrators mislead us as to what is represented in the story, but we are
very commonly misled, at least temporarily, by narration which is selective rather than
unreliable. With many kinds of fiction we are under-informed through the earlier
stages of the work about a character, being given the false impression that she has this
or that character trait, or performs this or that character-revealing action. In such cases
a certain emotional trajectory is, if only vaguely, laid out for us by the work, and there
15
This is a general truth about stories, fictional and non-fictional.
16
Which is not to deny the possibility of indeterminacies in the story; where the story is indeterminate as
between outcomes A and B, the narrator can be said to be unreliable if her account has something other than
either A or B occur.
154 Greg Currie
is a sense in which it is appropriate for us to experience a range of conflicting emotions.

A story may ask of us that we experience a kind of warm confidence in one of its char-
acters in order that we be the more disconcerted at the end when they are revealed as
the murderer. Or it may be that this change of reaction is not brought about by a change
in the presentation of the character’s traits, but by a piling up of evidence that their less
than admirable characteristics are a product of, and/or an understandable response to
hostile circumstances. Perhaps that is how we are to react to the rather calculating Lily
Bart in Edith Wharton’s The House of Mirth, who may seem to have more sympathetic
characteristics when we have worked our ways through the catalogue of hypocritical,
rapacious, and mean-spirited inhabitants of New York Society ranged against her.
For a work of complexity there need be no uniquely appropriate emotional sequence
of responding as the plot develops, especially when we compare first and later readings.
The reader of McEwan’s Sweet Tooth is likely to experience some degree of imaginative
resistance on first reading, because the narration, ostensibly from a female character,
sounds at various points rather unconvincing in this regard; there are moments when
one is likely to feel disappointment that McEwan, a male author, has not performed
the task he set himself well—or perhaps that he took on the task at all. But as usual
McEwan is ahead of us; at the end we learn that the narration is in fact that of a male
character imitating the voice of a female protagonist, and so the failure to achieve full
authenticity can be seen as appropriate and intended. What seemed to be an awk-
ward clash between what is fictional—a female narrator—and what is real—a male
author—turns out to be no such thing. A rereading is very unlikely to recover the ini-
tial and highly appropriate sense of awkwardness in the narration, and will probably be
replaced by an ironical recollection of one’s own earlier naïve response. And even first
readings generally allow of a range of appropriate emotional responses; responsible
and informed readers will differ in their reactions to characters and situations, and
authors who attempt to constrain our responses very closely are sometimes open to
criticism on that account.
In many stories the dynamic aspects of emotional response are sensitive to the rule
of representation as well as to the rule of evidence, because, in stories as in life, peo-
ple change, and our emotions are tracking changes in emotion-relevant states of the
object itself. A person you admire may start to behave in ways that destroy your admi-
ration, not merely because of a change in the direction of evidence; the person has
changed. That can happen in fiction, with stories which represent growth in moral
character (Prince Hal, Henry IV‒V) or moral decline (Angelo, Measure for Measure,
Antony, Antony and Cleopatra). Both the rule of representation and the rule of evi-
dence help us account for ways in which our emotions towards characters and situa-
tions may change through the process of reading. What is appropriate at one stage of
reading according to either rule may not be at another. For a diligent reader it should
be part of the experience of any such emotion that it be experienced as in some degree
provisional; we can only, if ever, be confident about what is represented by the story at
the end of it.
3. The Rule of Reality Again

It seems, then, that the rule of reality has nothing to do with assessing the appropriate-
ness of emotions in response to fictions; we only ever thought that it did because we
confused the rule of reality with the rule of correspondence. But do not considerations
of truth sometimes constrain the responses of the spectator at fiction? Does not the fact
that Prosecutor Garrison was, let us suppose, a wicked and/or deluded man make it,
in some sense, inappropriate for me to luxuriate in admiration for him while watching
JFK, which represents him as a selfless and committed seeker of truth? We may have
two sorts of worry about resting content with the idea that a response of admiration is
appropriate to Garrison as represented in the film, denying that notions of truth have
any role to play in assessing the appropriateness of the emotional response. The first
is a well-attested tendency of people to believe that things are as the fiction represents
them to be, when the fiction has the kind of real-world basis we find in JFK. Indeed,
some of the evidence for this general claim is based on responses to this very film: even
well-educated viewers were strongly influenced by it in their beliefs about the likeli-
hood of a wide-ranging conspiracy behind the Kennedy assassination.17 And it is very
plausible that the mechanism of this belief change has something to do with the emo-
tions the fiction elicits. So finding yourself with feelings of admiration for Garrison as
represented in the film may make you concerned about your vulnerability to insidious
change of belief. For that reason you may worry that you and others risk acquiring false
beliefs from fictions which falsely represent their characters.
The second line of reasoning attributes to the film’s makers the motive of gaining aes-
thetic traction, so to speak, by relying on a tendency in the audience not merely to imag-
ine the events of the film, but actually to believe them and, through believing them, to
experience more intense emotions in response to the film.18 Seen that way, the movie
looks like the result of an immoral enterprise: one that aims to achieve its effects by
misleading people, and not merely something which has, incidentally, this effect. If the
makers can call on the extra power of belief, they do not have to work so hard to produce
emotional effects that will make the work admired. That gives me a reason to stay out of
range of the film’s emotional influence even if I am rationally convinced that I will not
be affected in this way myself, and whether or not in doing so I am making a difference
to how others react to it. We do not wish to be associated with immoral enterprises even
when we think we will not personally be harmed by contact with them.
This suggests a connection with the phenomenon of imaginative resistance dis-
cussed previously. Fictions themselves are resistant to certain authorial stipulations: an
author cannot merely stipulate that female infanticide is a duty while at the same time
suggesting that things in the fiction are, in relevant non-moral ways, just the same as
17
See Butler et al. (1995). For an attempt to provide a modest theory of the ways fictions can affect belief
see Greg Currie and Anna Ichino, “Truth and trust in fiction (forthcoming in a volume of essays on art and
belief edited by Helen Bradley, Paul Noordhof and Ema Sullivan-Bissett).”
18
Later I will say more about the affective differences between belief and imagination.
156 Greg Currie
they are in the real world; it seems as if the author has to specify, or hint at, or at least
allow for some corresponding change in the supervenience base.19 And we as readers
are also resistant in various ways. Sometimes it seems to be a matter of not being able
to imagine, or adequately to imagine the fiction’s content, as with Sylvan’s box, which
contains, and does not contain, a statue.20 Sometimes, to bring things closer to our
topic, successful imagining seems to depend on affective responses that we may not be
able to bring to bear, as with the childish knock-knock joke which, the fiction tells us,
was the funniest thing the world had ever known. I can imagine that it is funny, but I
cannot imagine it as funny; I cannot inflect my imagining with an amused response.
Or it may be that we wish not to respond affectively in the way the work requires or
suggests. I do not want to imagine Garrison as a moral hero, because I do not want to
feel those waves of sympathy for him and resentment against his detractors the film
counts on eliciting.
Truth has not lost all ability to constrain our emotional engagement with fiction;
failure to meet the truth constraint, when this is part of a design to boost a work’s emo-
tional power and thereby to disguise its artistic weaknesses gives me a reason (not nec-
essarily an overriding reason) to resist responding to the work as I am prone to do.
How far this constraint extends is difficult to determine. Worries parallel to those that
arise with JFK might surface even in a fiction which does not feature—and so does
not misrepresent—real people as their characters. We might be uncomfortable with a
movie featuring a purely fictional DA who investigates the Kennedy case, if he or she is
presented in something like the way Garrison is presented in JFK, on the grounds that
the well-attested tendency for people to believe in conspiracies is still being exploited.
We worry also about the misrepresentation of ethnic groups where no real people or
incidents are featured. Where fictions are apt, for whatever reason, to generate emo-
tions which misrepresent, worries of this kind arise. I will not try here to investigate
the difficult question of when and by what mechanisms fictions are likely to do that.21
4. Emotions Directed at Representations

I return now to the contrast between the rule of reality and the rule of representa-
tion. What exactly are the implications of this contrast for the account we should
give of fictive emotions? What does it tell us about the nature of the representa-
tions associated with fictive emotions? Are fictive emotions, with their distinctive,
representation-dependent appropriateness conditions, genuine emotions at all? This
section answers these questions, though the answer to the final question is equivocal.
Admiring my friend Albert, my emotion represents him as admirable, and my
admiration is appropriate (in one of the several ways emotions can be appropri-
ate) if he is admirable. On the account I have given, my admiration for Holmes is
See Weatherson (2004). 20 See Priest (1997).

19
See Currie and Ichino (unpublished).

21
appropriate (again, in one of several possible ways) just in case he is represented as

admirable. Should we then say that my admiration of Sherlock Holmes represents
him as represented as admirable (in the stories by Doyle)? The answer is ‘no’. It is
characteristic of the states we enjoy (or suffer) in response to fictions that there is a
systematic mismatch between what the state represents and what makes the state an
appropriate one to have. Consider the case of imagining. Reading the Holmes sto-
ries, I imagine that Holmes and Moriarty struggle together at the Reichenbach Falls.22
What makes that an appropriate bit of imagining is not that Holmes and Moriarty
struggle together at the Reichenbach Falls—they do not. What makes it appropriate
is that, in the story, they are represented as struggling. But (to repeat a point made
earlier) my imagining is not The story represents that Holmes and Moriarty struggle
together at the Reichenbach Falls. It is that Holmes and Moriarty struggle together at
the Reichenbach Falls.23 When it comes to imagining in response to fiction, appropri-
ateness is a matter of how things are represented, not of their being represented that
way. My imaginings should represent as the story represents; it should not represent
the story’s representing.
This is of relevance to a debate over the nature of fictive emotions which began with
a paper by Kendall Walton.24 Walton argued that the fictive emotions are not genuine
emotions. He proposed that we should call them quasi-emotions, in recognition of
their phenomenological similarity to genuine emotions. Walton does not deny that
we feel fearful while watching horror movies; what he denies is that emotions are
constituted wholly by feelings, arguing, as cognitive theorists of the emotions do, that
emotions depend essentially on belief in the reality of their objects. One response
to this claim has been to point to certain empirical results concerning emotions in
brain-damage patients which have been interpreted as supporting the following view:
while emotions function to help us respond to real, current threats and opportuni-
ties, as when fear motivates us to flee, they also help us to choose between competing,
not-yet-realized options. Imagining certain courses of action and their outcomes can
cause us to put those options aside as unacceptably risky. And where people fail to
have the relevant responses to imagined options, they persist with inadvisable strate-
gies, often with disastrous results. So emotions in response to purely imagined situ-
ations can be as useful in helping to orient us to the world as are emotions generated
by real, current situations. And emotions in response to imagined situations could
not play this role unless there was a systematic match between them and our emo-
tions in response to actual states of affairs—otherwise the former would be no guide
to how the imagined scenario would affect us if it were actualized. For these reasons
we should count what Walton calls quasi-emotions—states which involve imagining
22
The kinds of imaginings at issue here are those I earlier described as marked.
23
The argument here parallels the argument concerning the contents of marked imaginings; see above.
24
Walton (1978).
158 Greg Currie
things which are dangerous, delightful, and so on, rather than believing in them—as
genuine emotions.25 Call that the functional argument.
For the other side of this debate, Stacie Friend has argued that we strike absurd-
ity if we lump these two kinds of states together. In an essay, I have already noted she
argues that ignoring Walton’s distinction obliges us to say that someone who despises
Garrison for his deluded or dishonest crusade, but who responds with admiration
to his representation in JFK, suffers from emotional inconsistency: having emotions
associated with inconsistent representations of their object. The despising represents
Garrison as failing to respect norms of evidence and of probative behaviour in inves-
tigating the assassination, while the admiration represents him as epistemically and
practically virtuous with respect to the same project. To restore consistency one or
other of these states ought to be abandoned. But that is surely a wrong diagnosis in the
case of the viewer caught up in the mood of JFK. The worst one can say of such a person
is that they are moved too easily by a questionable portrayal; they are not emotionally
inconsistent. To avoid attributions of spurious inconsistency, we need to characterize
the states in question in ways which avoid the idea that they are associated with incon-
sistent representations of their objects.
Given what was said earlier, we have a way to do this. For only one of these states—
the despising—needs to conform to the reality rule. The other—the admiring—is
governed by the rule of representation. The despising is appropriate if Garrison is (put-
ting it briefly) vicious; the admiring is appropriate if Garrison is represented in JFK
as virtuous. There is no incompatibility between these conditions. If that is the right
response, it underlines the difference between ordinary, real-life emotions and those of
fiction—the ones Walton wishes to call quasi-emotions. These two kinds of state have
quite different kinds of appropriateness conditions. And states with different kinds of
appropriateness conditions ought not to be lumped together.
At this point we might usefully step back a bit. Two things may be remarkably alike
in some ways and remarkably different in others. A profitable thing to argue about is
exactly how alike and how different they are. Accepting both arguments just now pre-
sented, we can say that emotions and quasi-emotions (the term simply picks out the
imagination-based cases and carries no implication at this stage that quasi-emotions
are not real emotions) are the same in that they are capable of aiding us in our inter-
actions with the world, and different in the conditions they have to satisfy in order
to aid us in that project. Emotions have to represent their objects as those objects
are, while quasi-emotions have to represent them as they are represented (in a novel,
in my imaginative project, or whatever). Suppose now we are doing a ‘boxology’ of
the mind: constructing a diagram that distinguishes importantly distinct mediators
25
The argument is due to Gendler and Kovakovich (2005). They draw primarily on empirical work done
by Damasio and colleagues: see Bechara et al. (1997). Gendler has said that the view expressed in this essay is
‘superseded’ by her recent work which has argued that much of what we have been inclined to attribute to the
operation of the imagination is due in fact to the activity of a kind of mental state she has named ‘alief ’ (see
Gendler 2010: 227). On the relation between imagination and alief see Currie and Ichino (2012).
between perception and action.26 Beliefs go in one box, desires in another, because
beliefs and desires play very different roles when it comes to generating behaviour,
and we want to say very different things about how they might be responsible for bad
things happening to the agent whose actions go awry. We could ignore those differ-
ences and have just one box marked ‘states mediating between perception and action’.
There would be nothing wrong with that as long as we were prepared to look closer and
see boxes within boxes, at which point we will want boxes for belief, desire, and emo-
tion. But wait. At that point would we not also want an imagination box? It does not
seem right—and no one I know has suggested—putting imagination inside the belief
box, making it a special kind of belief. We want a belief box and an imagination box
because believing P is appropriate when P is true, and imagining P is appropriate when
P is represented (in the relevant imaginative project) as true. Of course, there are other
differences between imagining and believing—notably, behavioural differences: when
I believe that my house is on fire I run outside; when I daydream that it is or (strange
but possible) read a story in which it is, I do not. But that sort of difference is explica-
ble in terms of the difference in appropriateness conditions: running out of my house
in response to a represented fire is not behaviour that will contribute to my fitness,
and creatures like us have different ways of responding to states with different kinds of
appropriateness conditions. If differences in appropriateness conditions for belief and
imagining justify putting them in different boxes, we have the same reason for having
separate boxes for emotions and quasi-emotions.27
5. Difference in Emotion Caused By Difference in

Desire?
The constraints on fictionally directed emotions so far discussed can be understood
only as operating at the point where some emotional response is in question; they say
nothing about when we should respond emotionally rather than not. We do not respond
emotionally to all the events of a fiction, nor should we. But there surely are events that
we should respond to emotionally, and the question then arises as to what would be an
appropriate emotional response. Someone who sits watching (a representation of) the
murder of Desdemona without a flicker of emotion is not responding appropriately,
26
Here I am indebted to the discussion in Weinberg (2014), though he will not agree with my conclusion.
27
Note that the appropriateness conditions for beliefs and for imaginings differ in other ways. It is at
least plausible to suppose that we have a categorical obligation to believe whatever we have evidence for
the truth of, assuming that the evidence meets the (possibly contextually dependent) threshold for belief.
But we have no more than a hypothetical obligation in the case of imagining: if our aim is to engage with
a certain fiction, then we should imagine those things for which we have an appropriate level of evi-
dence that they are so according to the fiction. (These formulations certainly need refinement, and there
is a complex debate under way as to how, if at all, we should formulate the norms of belief. On this see
Sullivan-Bissett 2014.)
160 Greg Currie
but they have not violated either the reality rule or the rule of representation. You violate
those rules only by having an (inappropriate) emotion.
At this point it is tempting to offer a principle like this: the way we ought to respond
to the events of the fiction is the way we would respond if they (or relevantly similar)
events occurred in real life. (Call this the Transfer Rule.) One difficulty is that there
are many ways to respond to real-life events, including not responding in any distinc-
tive way at all. That, notes Alan Gibbard, is probably what we should do concerning
the theft of a stranger’s camel far away; but that ought not to make indifference auto-
matically the right response to a fictional representation of such a theft.28 The fiction
may make the theft emotionally stirring, and we cannot dogmatically insist that being
stirred by it is irrational.29 There is nothing special to fiction here; a history or journal-
istic piece, well written, might present the real theft in such detail and with such vivid-
ness that, once again, our feelings are stirred. Perhaps the right response to the fictional
representation is the right response to the very same representation (same words) now
taken to be non-fictional. But that cannot be right. The appropriate response to a non-
fictional Anna Karenina would be either wonder at Tolstoy’s magical powers of mind
reading, or rejection of the whole thing as fantastic lies. And while the appropriate
response to murder in fiction ranges from horror to amusement, the same range does
not seem to be available for non-fiction.
Shaun Nichols has looked closely at the problem of how to explain the difference
between emotional responses in superficially similar real and fictional cases. While
not claiming a comprehensive solution, he emphasizes particularly the role of desire.
The thought is that our emotions depend as much on what we desire as on what we
believe (or imagine) and that differences in emotion which it is tempting to attribute
to the difference between believing P and imagining P are due in fact to differences
in the desires that go with possession of these two states. This is said in the service of
preserving, as far as possible, the idea that belief and imagination operate through a
‘single code’ which, other things being equal, will produce ‘similar’ effects, including
emotional effects.30 The thought is that differences in desire mean that other things are
not equal.
While desires are probably of relevance here (along with the controversial category
of desires-in-imagination sometimes postulated in these discussions) I do not think
they are key to understanding the differential emotional effects of belief and imagi-
nation.31 There are cases where differences of desire are hard to identify. Compare
imagining that someone you are close to is dead with believing that they are. However
28
Gibbard (1992: 126). Gibbard goes on to say that ‘It even makes sense to engage one’s feelings in fiction
from time to time’, as if this kind of engagement needs strict rationing.
29
I am assuming, like most parties to the current debate, that emotional responses to fiction are not all
automatically irrational, as Colin Radford once argued they are.
30
Nichols (2006). Nichols does not explain why the single code theory is not committed to the view that,
other things being equal, the effects will be qualitatively identical.
31
Nichols rejects the category of desires-in-imagination.
affecting the former state is likely to be, it surely is affectively very weak compared
with the latter. But what differences of desire are there here? In both the believing and
imagining situation I have the same strong desire for the safety of that person. Nichols
at one point puts the difference in terms of a difference between the real world and
imaginary situations: ‘we have different desires concerning the imaginary situation
and the real situation.’ How does this help? It is true that my caring about my loved
ones is based on concern for their safety in reality—I do not care about what hap-
pens to them in imaginary situations. But then how would we explain my caring about
the fate of Desdemona? When I watch the play I am not encouraged to imagine that
she in murdered in an imaginary situation, and I would not care about her fate if that
was what was represented any more than I care about the fates of my loved ones in
imaginary situations. And if it were, there would be an evident difference between my
belief and my imagining which would not need an explanation in terms of differential
desire. For believing that Desdemona is murdered, and imagining that Desdemona is
murdered in an imaginary situation are states with different contents, and the common
code hypothesis does not imply or even suggest that these two states will have similar
emotional effects. There is something here of the view rejected earlier on: that the dif-
ference between fiction and reality is a different between locations—worlds, in fact.
But once we see that fictions do not create worlds but representations—non-decep-
tively false ones, generally—of the real world, the view that our desires concerning
fictions are desires about what happens in imaginary worlds collapses.
6. Appropriate Emotions, Intended Emotions

In wondering about the different emotional effects of belief and imagination Nichols
is focused on the factual question: what are those differences and how are they to be
explained? This requires an investigation into the causes of emotions. My interest is
rather in the appropriateness of emotional responses to fictions, particularly where
those responses would seem distinctly inappropriate when directed at the real world.
But any such normatively focused account of emotional appropriateness must recog-
nize that emotions are hard for us to initiate and control; they are most often simply
responses to how things are, or how they seem. So any normative account of emotions
in response to fiction needs to say something about how we can expect appropriate-
ness in emotional responding to be achieved. My own suggestion is that, very often
in our encounters with fiction, we respond in the appropriate way simply because the
work elicits from us the intended response, and not because the work provides evi-
dence that we ought to respond in that way. In particular, the elicitation of emotion is
not achieved by having us recognize that a certain response is intended; elicitation of
emotion is not Gricean. On the contrary, the best evidence that an emotional response
is appropriate is often the fact that we have that response, for our having it suggests that
the work was designed to elicit it from us, assuming that the intended response is the
appropriate one.
162 Greg Currie
Why should the intended response be the appropriate one? It is not universally so,
and I will mention cases which contradict this assumption. But there is a reasonable
presumption that the intended response is appropriate, for it is generally the case with
artefacts that the way to use them to best advantage is to use them in the ways they are
intended to be used. We do not get the best from our cars by using them as dustbins,
or the best from paintings by using them as writing surfaces; they were not made with
the intention of fulfilling these purposes and are therefore not likely to fulfil them very
well. It would be similarly unpromising to invent a system of semantic rules which
gave completely unintended meanings to the sentences of Anna Karenina, or to insist
on imagining the negation of everything which the standard reading of that novel sug-
gests we are intended to imagine. Narratives of worth are artefacts carefully crafted to
convey stories in which certain things and not others happened; those are the things
we should imagine (in the marked way), and the careful crafting makes it likely that we
will imagine them. The presumption that the intended response is the appropriate one
is especially strong in cases where the maker has a reputation for producing work of
great quality and effectiveness—we can be more than usually confident in such cases
that the responses we find ourselves having are the appropriate ones. If, to bring this
back to the case of fiction and the emotions, we do not respond emotionally to the
work as we are intended to respond to it, we are likely to put ourselves at odds with the
tone and purposes of the work, to drift away from its themes and to miss the cadences
of its story-line.
Of course, in real life we are constantly presented with testimony (a kind of repre-
sentation, we have agreed) concerning situations to which we may respond emotion-
ally, and we may even detect an intention on the part of the testifier that we respond
to them in a certain way. Perhaps these intentions sometimes determine, at least in
part, how we do respond. Perhaps there are even norms, of politeness, say, according
to which we ought to respond in those intended ways. This does not, I think, make for
much similarity between the real and the fictive cases. With fiction there is nothing
beyond the representation; in the real world one always can, and often should, think
‘how should I respond to these events, given how they were, rather than merely how
they are represented as being?’ It can be right—as we shall see in a moment—to fail or
refuse to respond as intended to a fictional scenario, but then it is right with respect to the
events-as-represented.
Recall that emotions may be elicited by fictional works without our needing to recog-
nise that these are the emotions intended to be elicited. Still, it can make a difference to
the overall effect of a work that we recognise, post hoc perhaps, that this emotion was
intended; that can increase our sense of community with the author. What is it for a fic-
tional work to embody an intention concerning how we are to respond emotionally to the
work? Certainly, the expressed intentions of the author in this regard may be vague and in
some cases non-existent. Where they are vague, a range of responses can be thought of as
appropriate with some indeterminacy about what is appropriate at the margins. Where
they are non-existent, any emotion or none is appropriate. But such cases are rare and
we should not think that no appropriate emotion is indicated if we cannot find anything
in the work which positively expresses the intention that we respond in such-and-such a
way. For some intentions count as expressed merely by default; the author expresses an
intention that one respond in this way because it is generally understood that we would
be unreflectively inclined to respond in that way and the author does nothing to prevent
us from responding in that way. In treating the matter of intention in this way we are fol-
lowing agreed path for judging intention in regard to the contents of imagining. We imag-
ine—and presume ourselves, on reflection, intended to imagine—that the characters in a
fiction have two eyes, breathe air, and generally have features overwhelmingly standard
in the human population, unless the author has gone to the trouble of saying or implying
something to the contrary. For the author can be counted on to know that we will assume
this unless something is done to block that imagining. Using our imaginative inclinations
in that way, the author can economize on expressive effort. I see no reason why it should be
different with the emotional inflections we give to our imaginings.
This underlines the importance of a distinction between (i) the emotions the work
is expressive of, and (ii) what, if anything, it expresses about the intended emotional
response of a reader to the emotions expressed. It is possible for a work to be expressive of
thoroughly objectionable sentiments while it is manifestly intended that one respond to
those sentiments with contempt. And this will be true in certain cases even where there
is nothing one could point to in the text indicative of that intended response, because the
intended response coincides with the response readers in the intended audience would be
expected to have unless efforts were made to shift their response, and no such efforts seem
to have been made. We are clearly intended to treat the character Jim in Ring Lardner’s
story Haircut as contemptible, even though the narrator expresses a much more positive
attitude towards him, and no direct authorial intervention contradicts this opinion.
This last point is of some relevance to an example considered by Paisley Livingston
and Al Mele. They cite a poem by Baudelaire, ‘A Celle qui est trop gaie’, expressive of
the ‘gruesome hopelessness of a frustrated misogynist’.32 Are we, they ask, enjoined
to respond to this poem with the ugly feelings it expresses? Such complicity, they say,
would be ‘revolting’.33 Is this poem, understood as Livingston and Mele wish to under-
stand it, a problem for the present proposal?
The first thing to be said is that there should be room for an appropriate response
which, at least in part, does mirror the misogynist’s gruesome hopelessness. There can,
in particular, be orders of emotional response; one may respond in a certain way, and
also have an emotional response to that way of responding, thinking ‘that [referring
32
Livingston and Mele (1997): see especially p.164; ‘gruesome hopelessness’ is quoted from an essay by
Eric Auerbach.
33
Livingston and Mele (1997) offered this example as evidence against an earlier proposal of my own con-
cerning appropriate emotions in fiction. I emphasize that I am not seeking to defend here the view I earlier
expressed, and I am grateful to Livingston and Mele for their treatment of that earlier and certainly defective
account.
164 Greg Currie
ostensively to one’s own state] is deplorable’.34 It does not seem out of the question to
have, and be intended to have, a layered response to Baudelaire’s poem which involves
an emotion which is, in some measure, congruent with the troubling response of which
the poem is itself expressive. In the context of a reflective response of a questioning or
negative kind, that first response can be seen to have the merit of helping us to under-
stand, from the inside, emotions we find both alien and immoral, without threatening
our own moral integrity. Sharing an ugly emotion with another agent is not always a
case of endorsing the agent’s perspective, and will not be if one fails to identify with
that emotion in the way the other agent does.35
Livingston and Mele might agree to the proposal that we distinguish endorsed and
unendorsed emotions, but argue that there is no reason to suppose, in the case of
this particular poem, that the intended response should involve the higher-order
negative response to that emotion. For, they say, ‘One looks in vain in this poem for
evidence of any authorial attitude of distance or disapproval.’36 But I have already
noted that a response can sometimes be seen as intended merely on the grounds that
it is a response one is likely to have unless something is done in the text to block or
redirect it.
But this is only a partial response, because there are works which are clearly
intended to produce emotional effects which we would feel uncomfortable with and
where those works are not redeemed by the thought that the uncomfortable response
is one we are not intended to endorse. Livingston and Mele mention the story ‘The Jew
among Thorns’ by the Brothers Grimm, clearly intended to provoke an unpleasant and
endorsed delight in the fate of the Jew—all too available a response to the intended
audience anyway. But cases like this do not strike me as requiring us to treat questions
about the appropriateness of a particular mode of imagining differently from the way
we treat the appropriateness of the contents of imagining. Appropriate engagement
with a work requires us to imagine those things which are saliently true in the story,
yet for various reasons we may decide not to imagine them and may even think that
it would be wrong to do so; some things are just the sorts of things you ought not to
imagine. These are cases where the best thing to do—stop reading—is inconsistent
with appropriately engaging with the fiction, and are no more problematic than cases
where it would be better to save a child from a burning building than to go on read-
ing Anna Karenina. It is just the same with the emotional inflections we bring to our
reading.
34
This is close to a suggestion of Jerry Levinson, described in Livingston and Mele (1997), though
Levinson, as understood by Livingston and Mele, says that this should be a process in which one first has
the unpleasant feelings and then, through moral reflection, moves on to an attitude of disgust. In my view it
can be possible and instructive to take on some simulacrum of the unpleasant feeling while being at the same
time fully aware of their deplorable natures
35
See Frankfurt (1988: esp. chap. 12).
36
Livingston and Mele (1997: 164).
7. Conclusion
Fictions are representations of reality, not alternatives to it. And the emotions we have
in response to fictions are appropriate (in one sense) because of how things are rep-
resented, not because of how they are; in that they differ from the emotions we direct
at events and things in the real world. But there are many ways for emotions to be
appropriate, and in one of them truth sometimes matters even for fictive emotions; it
matters when we have reason to think that a representation is confusing us about what
is true, and getting an emotional free-ride in consequence. Being governed by what is
represented rather than by what is so, fictive emotions are also sensitive to how we are
intended to respond to the represented situation. And the response we are intended to
have to a fictional representation can be very far indeed from the response we would
have, or should have, to similar events in the real world, even if they come to us via tes-
timony or some other form of representation.37
References
Bechara, A., Damasio, H., Tranel, D., and Damasio, A. R. (1997). ‘Deciding Advantageously
Before Knowing the Advantageous Strategy’. Science 275: 1293–5.
Butler, L., Koopman, C., and Zimbardo, P. (1995). ‘The Psychological Impact of Viewing the Film
“JFK”: Emotions, Beliefs, and Political Behavioral Intentions’. Political Psychology 16: 237–57.
Currie, G. (2002). ‘Desire in Imagination’. In Conceivability and Possibility, ed. T. Gendler and J.
Hawthorne (Oxford: Oxford University Press), 201‒22.
Currie, G. (2010). ‘Tragedy’. Analysis 70: 632–8.
Currie, G. and Ichino, A. (2012). ‘Aliefs Don’t Exist, Though Some of their Relatives Do’. Analysis
72: 788–98.
Frankfurt, H. (1988). The Importance of What We Care About (Cambridge: Cambridge University
Press).
Gendler, T. (2000). ‘The Puzzle of Imaginative Resistance’. The Journal of Philosophy 97: 55–81.
Gendler, T. (2010). Intuition, Imagination and Philosophical Methodology (Oxford: Oxford
University Press).
Gendler, T. and Kovakovich, K. (2005). ‘Genuine Rational Fictional Emotions’. In Contemporary
Debates in Aesthetics, ed. M. Kieran (Oxford: Blackwell), 241‒53.
Gibbard, A. (1992). Wise Choices, Apt Feelings (Cambridge, MA: Harvard University Press).
Gilmore, J. (2011). ‘Aptness of Emotions for Fictions and Imaginings’. Pacific Philosophical
Goodman, N. (1982). ‘Fiction for Five Fingers’. Philosophy and Literature 6: 162–4.
37
This paper was read at a conference on the emotions held in Geneva in 2011. My thanks go to all who
offered criticisms and suggestions on that occasion, including Simon Blackburn, Pascal Engel, Peter Goldie,
Adam Morton, Julien Deonna, and Nancy Sherman. I have been greatly helped by discussions with Jonathan
Gilmore, as well as by reading his ‘Aptness of Emotions for Fictions and Imaginings’ (Gilmore 2011). I am
grateful also for written comments from Sabine Roeser and Cain Todd.
166 Greg Currie
Lewis, D. (1983). ‘Truth in Fiction’. In his Philosophical Papers, volume 1 (Oxford: Oxford
Livingston, P. and Mele, A. R. (1997). ‘Evaluating Emotional Responses to Fiction’. In Emotion
and the Arts, ed. M. Hjort and S. Lavers (Oxford: Oxford University Press), 157‒76.
Nichols, S. (2006). ‘Just the Imagination’. Mind & Language 21: 459–74.
Nichols, S. and Stich, S. (2003). Mindreading (Oxford: Oxford University Press).
Priest, G. (1997). ‘Sylvan’s Box: A Short Story and Ten Morals’. Notre Dame Journal of Formal
Logic 38: 573–82.
Schier, F. (1983). ‘Tragedy and the Community of Sentiment’. In Philosophy and Fiction: Essays in
Literary Aesthetics, ed. P. Lamarque (Oxford: Pergamon Press), 73‒92.
Sullivan-Bissett, E. (2014). ‘Belief, Truth and Biological Function’. Doctoral dissertation,
University of York.
Thomasson, A. (1999). Fiction and Metaphysics (Cambridge: Cambridge University Press).
Walton, K. (1978). ‘Fearing Fictions’. The Journal of Philosophy 75: 5–27.
Walton, K. (1993). Mimesis as Make-Believe (Cambridge, MA: Harvard University Press).
Walton, K. (1994). ‘Morals in Fiction and Fictional Morality (I)’. Supplement to the Proceedings of
the Aristotelian Society 68: 27–50.
Walton, K. (2014). In Other Shoes (Oxford: Oxford University Press).
Weatherson, B. (2004). ‘Morality, Fiction, and Possibility’. Philosophers’ Imprint 4: 1–27.
Weinberg, J. (2014). ‘All Your Desires In One Box’. In Aesthetics and the Sciences of Mind, ed.
G. Currie, M. Keiran, A. Meskin, and J. Robson (Oxford: Oxford University Press).
PA RT I I I
Emotion, Value, and the Self
11
Emotional Self-Trust
Linda Zagzebski
1. Introduction
In this chapter I want to defend the rationality of basic emotional self-trust that paral-
lels an argument I have given for the rationality of basic epistemic self-trust. My argu-
ment proceeds from the assumption that states of emotion can be fitting or non-fitting,
just as beliefs can be true or false, and what we call being justified or rational in an emo-
tion is doing the best we can to make our emotions fit their objects, just as what we call
being justified or rational in a belief is doing the best we can to make our beliefs true.
I will argue that the fittingness of an emotion does not reduce to the truth or fitting-
ness of a set of beliefs, so we cannot justify an emotion by reference to the justification
of epistemic states alone. Ultimately, the justification of our emotions is circular for
the same reason that the justification of our beliefs is circular. My position is that the
rational response to the awareness of epistemic circularity is basic epistemic self-trust.
Similarly, I will argue that the rational response to the awareness of emotion circularity
is basic trust in our emotion dispositions. It is likely that emotions are in general less
trustworthy than our epistemic faculties, but I will argue that we find many emotions
trustworthy, and it is rational to do so. I will conclude with some brief thoughts on the
ethical implications of emotional self-trust.
I assume that a self is a being that is conscious of itself, and that includes conscious-
ness of a variety of mental states that traditionally have been divided into rather
vague and overlapping categories—beliefs, desires, emotions, sensations, attitudes,
judgments, and choices, as well as imaginary versions of each. Once we reflect upon
these states, they change. Some of them change with little or no reflection. We catch
a glimpse of someone whom we think is a certain friend, but when we take a closer
look, we see that it is not, and our belief changes effortlessly. We feel angry at someone,
but after talking with them for a while, the anger turns to pity. Many states change in
this automatic and unreflective way. That experience gives us our initial model of how
to change states of the self and what we accomplish by doing so. As we acquire new
conscious states through experience or reflection, dissonance in the self is produced,
something changes without effort, and harmony is restored.
170 Linda Zagzebski
There are times, however, when states of the self change only under the direc-
tion of the self. Self-reflection brings our beliefs, emotions, and experiences
to conscious awareness, and reflection may show us disharmony we had not
noticed. We may even need to struggle to get our beliefs to be consistent, or to
get our emotions to harmonize with our beliefs, or to make our intentions line
up with our beliefs about norms of action. Clearly, we have only limited con-
trol over states of the self, and sometimes we attempt change but cannot do so,
but we do sometimes attempt change and succeed. Our experience of automatic
change gives us a model of the purpose of changing. We do not like disharmony
within the self, and sometimes we change in order to restore harmony. But that
is not the only reason we change. We think that some of our psychic states have
objects and the state can fit or not fit the object. We change because we want to
make the state more fitting. We change a belief because we want to make it true.
Changing a desire or an emotion is similar: we may find it unfitting in some
way. If we desire something undesirable or admire someone who is not admi-
rable or pity someone who is not pitiable, the state is not fitting. The state does
not succeed in being what it aims to be. So we have two reasons for changing or
attempting to change a psychic state: we want our states to fit their objects, and
we want our states to fit each other. My position is that ultimately our only way
to tell that our psychic states do not fit their objects is that they do not fit each
other, but the aim to make them fit each other is not the same as the aim to make
them fit their objects.
Before we reflect on our psychic states, we trust our powers, the products of our
powers, and the way in which we adjust those products without reflection. We natu-
rally believe that they are fitting, and feel an attitude of trust in them until reflec-
tion shows us that they might not be trustworthy. We learn to make adjustments with
experience. We naturally trust that what we think we see is what we see, but quickly
learn to trust a clear look more than a brief glimpse. Similarly, we naturally trust our
reasoning, but we quickly learn to trust careful reasoning more than quick and sloppy
reasoning. We learn other mistakes to avoid through training and reflection, such as
biases in inference. But in every case we trust something first. It is our default state.
When we make adjustments, we make adjustments in something we already trust.
So we learn to monitor the ways we exercise our powers, and learn to change some of
what we naturally do, but we cannot change all of what we naturally do. That would be
impossible.
My position, then, is that self-trust is the starting point of self-reflection. The issue
I want to raise next is what happens when we reflect upon our powers and our trust in
them. Can we get proof that our powers are untrustworthy? Can we get proof that they
are trustworthy, thereby making self-trust unnecessary? This question arises because
trust appears to be a second-best state—what you fall back on when you do not have
the proof you would prefer to have. So maybe we would prefer to escape trust. Can we
do so? If not, what attitude should we adopt about self-trust?
Emotional Self-trust 171
2. The Rational Inescapability of Epistemic Self-Trust

The pre-reflective self includes some desires and beliefs that are natural. I assume there
is a natural desire for truth and a natural belief that the natural desire for truth is satisfi-
able.1 The pre-reflective self includes an attitude of trust in our epistemic faculties for
the aim of getting truth in advance of proof of their trustworthiness.
What happens when we reflect? If we reflect long enough, we notice that there is no
non-circular way to determine that the natural desire for truth is satisfiable, although
we sometimes have non-circular ways to tell that the natural desire for truth fails. If a
faculty gave us inconsistent outputs, we would be able to tell that it is unreliable, but
there is no non-circular way to tell that it is reliable. Any argument for its reliabil-
ity would use the same faculty. This phenomenon of epistemic circularity has been
noticed by a number of philosophers. Richard Foley (2001) links it with the position
that there is no answer to the radical skeptic, but one need not be especially wor-
ried about skepticism to notice the circularity of ultimate epistemic justification, as
William Alston (2005) points out. As I see it, the desire to be fully reflectively justi-
fied is just the natural desire for truth made rigorously self-reflective. When we are
self-reflective, the desire for truth moves us to seek reasons for our beliefs, and the
discovery that the search for reasons can never be completed without circularity is
the discovery that we cannot get something we desire—a kind of guarantee of success
in reaching truth that would make self-trust unnecessary. So if we retain the natural
belief that the natural desire for truth is satisfiable, we cannot escape basic trust in our
epistemic faculties.
Epistemic circularity is only one of the reasons we cannot escape self-trust, and it
is not even the most basic reason. Focusing on epistemic circularity obscures a more
important reason. Imagine that there was a way to complete the process of justify-
ing our beliefs. For instance, suppose that strong foundationalism had succeeded. It
is tempting to think that that would make self-trust unnecessary, but unfortunately it
would not, because we would still need to assume that there is a connection between
successfully reaching truth and what we do when we attempt to answer our questions.
Whether or not we can complete the process of answering our questions, we need trust
in the connection between the process of attempting to answer our questions and suc-
cess. We can make the same point in terms of reasons for belief. Even if reasons ulti-
mately are not circular, but derive from a foundation in something certain, we would
still need to trust that there is a connection between our entire set of reasons and suc-
cess in getting the truth—trust in our sense of certainty, as well as trust in the connec-
tion between the foundation and the reasons based upon it.
I am not going to argue that there is no rational alternative to epistemic self-trust.
Perhaps we can change either the natural desire for truth or the natural belief that it is
1
What I mean by the desire for truth is the desire to have our questions answered. I am not taking sides
on debates about truth.
172 Linda Zagzebski
satisfiable. A Pyrrhonian skeptic will opt out of making judgments. A certain kind of
philosophical neurotic will continue to make judgments but will be plagued by doubts.
There are other possibilities as well. I am not claiming that trust in one’s epistemic fac-
ulties is the only rational response to our lack of proof of the link between our faculties
and success at getting the truth, but I think that it is more rational than alternatives
because it produces harmony in the self with the least amount of change in the self. We
begin with the natural desire for truth, and first with self-trust. Upon reflection we dis-
cover that there is no rational way to escape the need for self-trust while retaining our
natural desire for truth and the natural belief that the desire is satisfiable. Reflection
tells us to make our trust reflective; it does not eliminate the need for it.
Reflective epistemic self-trust is trust in our faculties when we are using them the
best way we can to get the truth. That is the property I call epistemic conscientious-
ness. A reflective person distinguishes when she is trustworthy from when she is not
by her degree of conscientiousness. She cannot do better than the best she can, and
she trusts herself the most when she is doing the best she can. But the conscientious
use of her faculties would not get her to the truth unless the faculties were already
designed to reach their natural ends—to get her to the truth. Trust in herself when she
is conscientious rests upon a more basic trust in the faculties themselves. The consci-
entious exercise of those faculties over time, with reflection on changes in our experi-
ences and other beliefs, is the only test we ultimately have that we have succeeded in
getting a true belief. Survival of conscientious self-reflection is the ultimate norm of
rationality.
3. The Rational Inescapability of Emotional Self-Trust

I want to argue now that the issue of reflective justification also arises for the more com-
plex and interesting category of emotions. I will not try to adjudicate between com-
peting accounts of emotion here, but I want to show why it is appropriate to raise the
question of trusting or doubting an emotion, given certain assumptions about the
nature of emotion. First, I assume that emotions have intentional objects and that hav-
ing an intentional object is roughly what distinguishes an emotion from a sensation or a
mood. We fear something, hope for something, pity someone, love someone, feel indig-
nation at some state of affairs, feel sympathy with someone’s plight, feel angry at some-
one, and so on. When I reflect, I can reasonably ask whether it is fitting or appropriate
that anger is directed at the object of my anger, whether what I fear is really fearsome,
whether the object of my indignation deserves that emotion, and so on. Some emo-
tions may be pure reactions, and we do not think that the issue of fit arises. For instance,
feeling irritated at something is more like finding something nauseating than feeling
angry at it. I need not ask myself whether it is appropriate for me to feel irritated at a
word-processing program, even while I know that most people do not find it irritating.
There are also emotions that can be appropriate or inappropriate, but where we think
the determination of fit is up to me. For instance, I determine whom to love and what to
hope for. Even so, I can make a mistake. I imagine that the range of fitting objects of love
or hope varies more from person to person than the range of fitting objects of anger or
fear, but there are still objects beyond the range of the fitting, and I can often recognize
that myself. I may later judge that I should not have loved my roses so much, or that it
was unwise to hope for a windfall profit on the stock market, and it seems to me that my
later judgment can be correct. It survives critical self-reflection better than my earlier
emotion. So with some qualifications, emotions can fit or not fit their objects, and we
think that an emotion ought to fit its object.
Given that emotions may or may not fit the object, we can raise the same sort of
questions about emotions as about beliefs and desires. That is, we can reasonably ask
ourselves whether we have a non-circular way to tell that our emotion dispositions
in conjunction with our other faculties reliably produce a state that fits its object.
Similarly, we can ask whether we have non-circular reasons for a particular emotion,
just as we can ask whether we have non-circular reasons for a particular belief.
One other feature of emotion that I assume for this argument is that an emotion has a
cognitive component. I think that it also has an affective component, but I will focus on
the former. In a state of emotion, something appears to the agent to be a certain way—a
way that is distinctive of the emotion type and which is not purely descriptive. In a state
of fear, the object of fear appears fearsome; in a state of pity, the object of pity appears
pitiable; in a state of love, the object of love appears lovable; and so on.2 I am not sug-
gesting that an emotion includes a judgment or belief. Something can appear fearsome
to the agent when she does not judge that it is fearsome. In fact, she may judge that it
is not fearsome. But the category of appearances raises the question of fit. If the light
appears to me to be green when it is red, there is a misalignment between the world and
my faculties even if I do not judge that the light is green. There is nonetheless a misstep
of some kind—a lack of fit between my faculties and my environment. The faculty or
disposition through which something appears to me is misrepresenting the object.3
Similarly, if someone appears pitiful when she is not, there is a misalignment between
the world and the emotion disposition operating in that situation even if I do not judge
that she is pitiful. For philosophers who take the position that emotions are judgments
or have judgments as components, the conclusion that we can be mistaken in our emo-
tions is straightforward.4 My point here is that there can be an error without judgment,
and we need not think that emotions include judgments to think that an emotion can
be in error. So one way that the possibility of error arises for emotions is that emotions
involve appearances and an appearance can misrepresent the object.
2
In Zagzebski (2003 and 2004) I call concepts that we apply to the objects of emotion—pitiable, fearsome,
contemptible, rude, and so on—“thick affective concepts.” I think these concepts are not purely descriptive,
but are affectively laden. Their cognitive and affective aspects cannot be pulled apart.
3
For a general account of appearing in perception, see Alston (1999).
4
See, for example, Nussbaum (2001) and Solomon (1980) and (1984). For a detailed account of emotion
that interprets them as having cognitive content without judgment, see Roberts (2003).
174 Linda Zagzebski
We sometimes have a non-circular way to tell that an emotion is unfitting. That is

due to another feature of emotion: the appearances that are components of emotion
depend upon particular descriptive features. For instance, fearsomeness depends
upon the fact that the object can harm me. Pitifulness depends upon the fact that the
object is suffering. If the descriptive feature does not apply, the emotion is unfitting,
and the subject can find out that it is unfitting by finding out that the descriptive feature
does not apply.
For instance, I may feel pity for someone who, from my point of view, has suffered a
loss or is in pain. Maybe she was denied a job for which she applied, and I am sorry that
she did not get it. I assume she is suffering over it and feel pity for her. But if I find out
that she secretly did not want the job and is relieved that it was not offered to her, I will
conclude that she is not pitiful, and normally my feeling of pity will disappear effort-
lessly. That suggests that an emotion fits the circumstances only when certain descrip-
tive facts obtain. If the agent’s beliefs about those facts are false, the emotion does not
fit, and this is something it is possible for an agent to see herself. So we can sometimes
see that an emotion is unfitting or inappropriate in a non-circular way; that is, a way
that does not refer to emotion dispositions.
Having an accurate grasp of the descriptive features of a situation is necessary for the
right emotion, but I want to argue now that it is not sufficient. The move from the out-
put of our perceptual and epistemic faculties to an emotion state requires the operation
of an emotion faculty or disposition. We have evidence that there is a distinct faculty
or disposition in the production of emotion because of the cases of individuals who
have all the right beliefs about the descriptive facts but lack the ability to feel emotions
because of damage to sections of the prefrontal cortex region of the brain.5 Perceptual
and cognitive processes are not sufficient to produce an emotion, much less an emo-
tion of any particular type. A person without an emotion disposition will not have an
emotion no matter what her descriptive beliefs, and absent a particular emotion dis-
position, there are indefinitely many affective states one could have that are compatible
with any set of descriptive beliefs.
Imagine beings just like us in perceptual and cognitive abilities, but with different
emotion dispositions. For instance, suppose that there are beings otherwise like us,
but who reverse the human emotions of contempt and offense. In a situation in which
a human being typically feels offended, these beings feel contempt, and in a situation
in which a human being typically feels contempt, they feel offended. We could not con-
vince the emotion-reversers that it is not appropriate to feel offended when someone is
doing what we see as contemptible, nor could they convince us of the contrary, because
it takes an emotion disposition to move from the belief that a situation has certain
descriptive features to an emotion such as offense or contempt, and we differ in those
dispositions.
5
For example, see Damasio (1994). For other examples of how damage to the prefrontal cortex affects the
emotional dispositions of persons, see Berlin et al. (2004) and Driscoll (2009).
Suppose, then, that I have settled for myself the relevant descriptive facts in some
situation. How can I tell that my emotion is fitting? To put the same question another
way, suppose that two people agree on all the descriptive facts but differ in emotion.
You and I agree that someone has expressed disrespect for us, but I see it as rude and
you do not. How could we settle the disagreement? My point is not that emotion disa-
greements between persons need to be settled. Usually they do not. I am raising the
issue of settling disagreement as a way to show how an emotion is justified. I do not
see how it can be done without reference to other instances of the same emotion type
in relevantly similar circumstances. So if I feel offended and you do not, I would point
out that the situation is like other situations in which you felt offended or thought the
feeling of offense was appropriate. If you are not disposed to feel offended in circum-
stances of the relevant kind, it does no good to point out to you that it is of the relevant
kind.
The justification of fear is also circular. Suppose you and I both observe the pres-
ence of something that can harm us. We agree on all the relevant descriptive features
of the situation—that the object or animal is harmful, the degree of harm it can inflict
upon us, and the probability that it will in fact harm us. But I feel fear and you do
not. Once we settle on that and become aware of the differences in our emotional
response, at least one of us will be puzzled. Certainly I will be puzzled. I might actu-
ally admire you because I might believe that fear, while appropriate in the circum-
stances, sometimes prevents a person from making the most rational response to a
dangerous situation. You might be better able to figure out a way for us to escape than
I am in the grip of my fear. But that does not make your emotion more fitting than
mine in the sense I mean. Another possibility is that I would admire you if I believed
that your ability to avoid fear is due to discipline, that in the past you felt fear in such
situations, but you learned through careful training and practice not to be afraid.
That also does not show that the emotion of fear is inappropriate. It shows that there
can be reasons not to have an appropriate emotion. But if your lack of fear is due to a
lack of the disposition to feel fear, I would think there is something wrong with you,
and I would think that because I assume that my own disposition is correct. If you
lack the disposition or if your disposition is radically different from mine, I would
be at a loss to justify the appropriateness of any particular output of my own dis-
position. My judgment that fear is appropriate in this case is correct only under the
assumption that my disposition to fear usually produces an emotion appropriate to
the circumstances.
The appropriateness of the emotion of disgust has received some attention in the
recent literature. The research of Jonathan Haidt et al. (1993) shows that people who
refer to disgust as a reason not to do something cannot justify their judgment any fur-
ther than the experience of the emotion of disgust itself. Haidt’s examples include eat-
ing one’s pet dog that was killed by a car, cutting up an American flag and using it for
rags to clean the bathroom, and having sex with a dead chicken. As far as I know, Haidt
does not take a stand on the appropriateness of disgust in these cases. The focus of
176 Linda Zagzebski
most of the discussion of his research has been on whether finding an act disgusting is
a good reason to say you should not do it.6
But I am not discussing here the issue of whether an emotion is a reason to act in cer-
tain ways. My topic is the appropriateness of the emotion to the circumstances. Martha
Nussbaum (2006) has given an extended argument that disgust should be irrelevant to
the law, and she clearly thinks that disgust is often an inappropriate response to a situ-
ation, quite apart from its legal ramifications. She argues that disgust can be attributed
to a fear of our humanity; in particular, a rejection of our bodily nature. She argues
further that disgust expresses “magical ideas of contamination” (2006: 14), and aspira-
tions to purity that are not realistic for human life as we know it. It appears to me that
Nussbaum’s position is not only that one should not use disgust as a ground for a legal
judgment; she thinks that disgust is often a mistake. What appears disgusting is not
disgusting. The emotion misrepresents the object. But how could we settle a disagree-
ment between someone who finds sex with a dead chicken disgusting and someone
who does not? They might agree on all the descriptive features of the act and its conse-
quences, including the fact that the act is harmless. Each disputant might compare her
emotion of disgust with previous situations in which she felt disgust, as well as similar
situations in which she did not. Once she has done that, there is not much that can be
done to resolve the disagreement except by reference to cases in which both parties
take the feeling of disgust to be appropriate. That is, the two parties together can only
do what each of them would do to settle for herself whether the emotion is appropriate.
They compare one instance of the emotion with another. Sometimes one or the other
will find anomalies in their emotion, and that will help them settle the disagreement.
But if they simply have different emotion dispositions, I do not see how it is possible
to adjudicate differences in the dispositions.7 What happens, I think, is that there are
6
Kwame Anthony Appiah (2008: 141) says in reference to the Haidt research, “If ‘You can’t have sex with
that, it’s a dead chicken’ is (as I rather suspect) a bad reason, we’ll want to be able to say why.” Appiah’s discus-
sion is about disgust as a moral sentiment. Given the context, I cannot tell if Appiah means that he suspects
that the fact that something is a chicken does not give one a good moral reason not to have sex with it, or if he
thinks it is not a good reason simpliciter.
7
Haidt et al. (2008: 1107) present one way a person could be led to see disgust as inappropriate, and then
give their verdict on the inescapability of relying upon emotion in moral judgment:
The present data show that some people are more easily influenced by extraneous disgust
than are others. But for those who do pay attention, should they? Should people use their
intuitions when confronted with moral issues? In the present case the answer appears to be
no because people who followed their feelings were “tricked” by extraneous disgust. Prior
experiments in which explicit attributions of affect to extraneous sources were encouraged
(Schwarz & Clore, 1983) suggest that it might help to be reminded that in some situations, the
feeling of disgust simply has nothing to do with the judgment at hand. For example, jurors
judging a defendant with a facial deformity, or who engages in harmless sexual practices they
do not approve of, might need help in overcoming their spontaneous flashes of irrelevant
disgust. Thus, once we realize that we indeed rely on repugnance in cases where it is clearly
not a kind of wisdom (Kass, 1997), we can perhaps do something about it.
However, if people ignore all feelings when making judgments, they may have little else
to go on . . . If each person tried to figure out the optimal moral judgment without taking any
counsel from affectively laden intuitions, it is not clear that the products of such deliberation
would be wise. Indeed, Damasio’s (1994) patients, who lack the ability to integrate somatic
people who find that their emotion of disgust changes over time and with reflection,
and possibly disappears entirely, whereas the emotions of other persons do not change,
even with similar experiences and similar reflections. But we cannot expect someone
else to change an emotion because we have changed it in similar circumstances. We
need to admit that once we agree with each other on the descriptive features of a situa-
tion, then if there is a difference of emotional response, neither one of us can reasona-
bly judge that we are right and the other is wrong without the assumption that our own
emotion dispositions are more trustworthy than the dispositions of the other person.
I do not see how we can produce such an argument if the emotions of both persons
survive their own critical reflection on their total set of beliefs and emotions without
dissonance.
Emotion circularity is not the deepest reason we need trust in our emotion disposi-
tions. In the previous section I argued that even if we could complete the search for
reasons for our beliefs in a non-circular way, we would still need to trust that there
is any connection at all between what we do when we try to get the truth and success
in reaching the truth. Similarly, the need for trust in our emotion dispositions is not
primarily due to the fact that we cannot complete the task of justifying an emotion in
a non-circular way. The problem is that we have no way of telling that there is any con-
nection at all between what we do to justify the fittingness of an emotion and its actual
fittingness. To the extent that we rely upon our emotions in the conduct of our lives
and believe that the emotions upon which we rely are fitting, we need basic trust in the
tendency of our emotion dispositions to produce fitting emotions for the same reason
we need basic trust in the tendency of our epistemic faculties to produce true beliefs.
Someone might accept my assumption that emotions have cognitive content, but
reject my argument that they can fit or not fit their objects on the grounds that what
makes an emotion appropriate is not that it fits its object, but just that it leads to action
that is right for circumstances with certain descriptive features. I have heard the claim
that as long as a person acts appropriately, any emotion from which the act arises is
appropriate, as well as no emotion at all. On this view, we gain nothing in explanatory
power by saying that the cognitive content of emotions can be fitting or unfitting, nor
is it accurate to call the feeling aspect of an emotion appropriate except in the sense
just noted. If a person responds to a suffering person by helping her, whatever emotion
leads her to do so is appropriate. If she runs from danger, whatever emotion leads her
to do so is fitting just because it leads her to run. Otherwise, it makes no difference that
she feels one thing rather than another. Inner states mean nothing.
An adequate answer to this objection would require a full defense of my view on
the nature of emotions, but I want to offer one observation about it here. As long as an
markers into their decision making, seem quite foolish, and they are unable to make up their
minds about simple matters. Our findings lead us to conclude that affectively laden moral
intuitions are often useful, but we need to be aware of our heavy reliance on such intuitions
and of the factors that sometimes distort them.
178 Linda Zagzebski
emotion is a response to the world and it has cognitive content, it is hard to see how
someone can think an emotion state is not the sort of state that can fit or fail to fit its
object without having the same view about beliefs. If the appropriateness of an emo-
tion is determined by its power to cause a person to act in a way that is right for the situ-
ation, the appropriateness of a belief ought to be determined the same way. According
to this position, when I see a suffering person it ought not to matter whether I believe
that I see a suffering person or whether I believe that 2 + 2 = 4 or that the sky is blue,
as long as I help the person. What matters is that the belief state leads to the right act.
When I am in danger, it ought not to matter whether I believe I am in danger or instead
believe that Benjamin Franklin invented the lightning conductor, as long as I run away.
But, of course, we do think that a belief can fit or fail to fit its object independently of
the acts to which it leads, and we think that because we think a belief state is a state of
taking the world to be a certain way. My position is that an emotion is also, in part, a
state of taking the world to be a certain way. It differs from beliefs in several respects,
but as long as it includes an aspect of appearance, its appropriateness cannot be simply
a matter of leading to appropriate external behavior. We must ask ourselves, then, how
we can tell that an emotion is appropriate. When we do so, we encounter the same
problem that we encounter with our other faculties. Upon reflection we find that we
need basic trust in our emotion dispositions, just as we need basic trust in our sensory
and epistemic faculties.
4. Trustworthy and Untrustworthy Emotions

There are important differences between our emotions and some of our other facul-
ties that lead some people to become emotion skeptics. Our reasons for trusting our
emotions are circular, but some of our emotions lack circular justification. For the
most part, our sensory faculties produce consistent outputs, but many of our emo-
tion dispositions do not, and we later judge that the emotion was a mistake. An emo-
tion may change when our beliefs about the descriptive facts do not change, or we
judge the emotion to be inappropriate after reflection. We have experiences of falling
in and out of love with the same person, becoming angry and then losing anger at
the same incident, feeling contempt that turns to pity, feeling disgust that changes
into compassion, and so forth in a multitude of cases that illustrate the instability
and lack of consistency of many of our emotions. These experiences happen often
enough that most of us are much less trusting of our emotion dispositions than of
our perceptual and epistemic faculties. An emotion skeptic will say that these con-
siderations show that we lack even circular justification for our emotions. She will
say that our emotions are unstable, do not survive critical self-reflection, and there-
fore should not be trusted. But general emotion skepticism is disingenuous, since
there are many emotions that we all trust, and many emotions that we need to trust
in order to live a normal life.
One is the emotion of sympathy. Sympathy is critical to moral beings. It underpins

many moral virtues such as generosity, gratitude, compassion, and benevolence, and
its absence makes the development of the stronger emotion of love impossible. But
how could we justify the appropriateness of sympathy? Perhaps there is an argument
that consistency requires us to feel about others in a way that is (very roughly) similar
to the way we feel about ourselves, but it is doubtful whether such an argument would
get very far. At least, it is doubtful that it will make headway against anybody who does
not already have a natural inclination to sympathy. We could not prove that the par-
ticular emotion we call sympathy is appropriately directed at anything, much less that
it is appropriately directed towards particular other persons. It seems to me that the
most honest thing to say is simply that we trust it.
There are many other emotions that we trust. In my experience, even philosophers
with the greatest proclivity to moral skepticism trust their emotion of indignation.
Unlike sympathy, indignation is an emotion that quite obviously is not always appro-
priate because it is dependent upon or sensitive to moral judgments, usually judg-
ments of injustice. But curiously, people will often trust their emotion of indignation
more than the associated moral judgment, and sometimes use the emotion as grounds
for the judgment. Whether this is justified is an intricate question, but I use it only as
evidence that there are emotions we trust without non-circular defense.
Most of us have considerable trust in some epistemic emotions as well—emotions
directed at beliefs. One is the emotion that Thomas Reid calls “epistemic ridicule.”8 We
think “That is ridiculous!” or “How absurd!” when we hear something we take to be
epistemically outrageous. Reid says we have this emotion when faced with someone
who denies a First Principle of common sense, and he clearly thinks it is appropriate in
such cases.9 What is particularly interesting about Reid’s point for my purpose here is
that he does not use the fact that someone violates common sense as a way to justify the
emotion of ridicule. He argues the other way around. Trust in the emotion of ridicule
grounds his claim that we have reason to think that the object of ridicule is epistemi-
cally unjustified. Again, my point is not that there is no way we can be mistaken in this
emotion, but rather that we place a great deal of trust in it, and there is no justification
for it that does not assume the general trustworthiness of the disposition that produces
the emotion.
It appears, then, that we trust some emotions even though we do not trust others.
If trust in an emotion survives conscientious reflection, we have the same type of
grounds for trusting it as we have for trusting beliefs. Perhaps we think that our past
emotions have a poorer record of survival of conscientious reflection than our past
beliefs. If so, we have reason to think that our emotion dispositions in general are less
8
In my vocabulary, “ridicule” is the name of a form of behavior, not an emotion. But it is hard to know
what to call the emotion that accompanies that behavior, so I am following Reid in calling it ridicule.
9
Thomas Reid, “Essay Six: Of Judgment” (1983: 256).
180 Linda Zagzebski
trustworthy than our epistemic dispositions, and so we ought to be cautious, but I do
not see that we have reason for general emotion skepticism.
I suggest that a conscientious person should treat emotion dispositions the same way
she treats her epistemic faculties. She does not have a non-circular justification for the
reliability of either kind of faculty, but the outputs of both kinds of faculty can survive
conscientious self-reflection. In both cases she needs to rely upon the faculty to lead a nor-
mal life, and in both cases she cannot do that without placing basic trust in the faculty or
disposition itself. She monitors the faculty, adjusting or attempting to adjust their outputs
in response to reflective judgment. What I have called epistemic conscientiousness is the
self-conscious attempt to make our beliefs fit the truth. I want to suggest now that there
is a form of conscientiousness in which we make a self-conscious attempt to make our
emotions fit their objects. Like epistemic conscientiousness, emotional conscientiousness
would not be trustworthy unless our dispositions were already basically trustworthy. That
is because there is no point in trusting ourselves when we use a faculty to the best of our
ability unless the faculty is generally suited for success. The trustworthiness of emotional
conscientiousness depends upon the general trustworthiness of our emotion disposi-
tions, and trust in ourselves when we are emotionally conscientious depends upon a more
basic trust in our emotion dispositions in conjunction with our other faculties.
Could an epistemically conscientious person opt out of trusting any of her emotions?
I have not said how many emotions survive conscientious self-reflection for any particu-
lar person. I think it is highly unlikely, perhaps impossible, that none of them does. The
consequences of total lack of trust in emotions would be drastic. I have argued in another
place that moral judgments depend upon emotions, so if I am right about that, skepticism
about emotions would lead to skepticism about moral judgments.10 I have not defended
that position here, so my claim that radical emotion skepticism is not viable does not
depend upon it, but I am relying upon the intuitive force of examples of emotions that are
generally trusted by conscientious persons, and which are furthermore used as partial
justifications for action. If a conscientious person trusts her decision to act in a certain
way, she needs to trust the emotions that form at least a partial basis for the decision. In
any case, I think it is fair to say that we often act by depending upon the appropriateness
of certain emotions. If the emotion is not appropriate, that needs to be established by the
failure of the emotion to satisfy the demands of conscientious self-reflection. I do not see
that any general skeptical considerations are apt to succeed.
Trusting an emotion that survives conscientious reflection is justified in the same
way trusting a belief that survives conscientious reflection is justified. It is justified
in the only way it can be justified. Once we realize we need not succumb to skepti-
cism about emotions, emotions can serve their proper place in our lives without philo-
sophical embarrassment. An emotion that I think we trust, and rightfully so when it
10
Zagzebski (2003 and 2004).
survives reflection, is admiration. I have argued that a comprehensive moral theory can
be constructed around the emotion of admiration for exemplars.11 Other emotions are
crucial for moral theory and practice. I have already mentioned sympathy, indignation,
and disgust, but there are many others, such as gratitude, pity, respect, and contempt.
I do not claim that every moral theory must have a place for each of these emotions,
but moral theory must at a minimum have a place for emotions that survive reflection
because there are times when we find these emotions more trustworthy than the purely
conceptual parts of a theory. I have proposed here a way we can justify our trust in these
emotions. We may acquire them unreflectively, but ultimately, emotions are justified the
same way beliefs are justified: we continue to have them when we conscientiously reflect
upon our total set of psychic states. If I am right that survival of conscientious reflection
is ultimately our only way to justify a psychic state, then many emotions are justified. We
need trust in them because we do not have the kind of non-circular justification philoso-
phers yearn for. But then we do not have that kind of justification for beliefs either.12
References
Alston, W. P. (1999). “Back to the Theory of Appearing.” Noûs 33: 181–203.
Alston, W. P. (2005). Beyond Justification: Dimensions of Epistemic Evaluation (Ithaca,
NY: Cornell University Press).
Appiah, K. A. (2008). Experiments in Ethics (Cambridge, MA: Harvard University Press).
Berlin, H. A., Kischka, U., and Rolls, E. T. (2004). “Impulsivity, Time Perception, Emotion and
Reinforcement Sensitivity in Patients with Orbitofrontal Cortex Lesions.” Brain 127: 1108–26.
Damasio, A. R. (1994). Descartes’ Error: Emotion, Reason, and the Human Brain (New York: Quill
Publishers).
Driscoll, D. M. (2009). The Effects of Prefrontal Cortex Damage in the Regulation of Emotion
(Iowa City: University of Iowa Press).
Foley, R. (2001). Intellectual Trust in Oneself and Others (New York: Cambridge University
Press).
Haidt, J., Koller, S., and Dias, M. G. (1993). “Affect, Culture, and Morality, or Is It Wrong to Eat
Your Dog?” Journal of Personality and Social Psychology 65: 613–28.
Haidt, J., Schnall, S., Clore, G. L., and Jordan, A. H. (2008). “Disgust as Embodied Moral
Judgment.” Personality & Social Psychology Bulletin 34: 1096–109.
Kass, L. (1997). “The Wisdom of Repugnance.” New Republic 216: 17‒26.
Nussbaum, M. C. (2001). Upheavals of Thought: The Intelligence of Emotions (Cambridge:
Cambridge University Press).
Nussbaum, M. C. (2006). Hiding from Humanity: Disgust, Shame, and the Law (Princeton:
Princeton University Press).
11
See Zagzebski (2010). I am currently working on a book manuscript on that theory, tentatively entitled
Exemplarist Virtue Theory.
12
Material for this chapter is taken from Zagzebski (2012: ch. 4).
182 Linda Zagzebski
Reid, T. (1983 [1785]). “Essays On the Intellectual Powers of Man.” In Thomas Reid’s Inquiry and
Essays, ed. R. E. Beanblossom and K. Lehrer (Indianapolis: Hackett Publishing).
Roberts, R. C. (2003). Emotions: An Essay in Aid of Moral Philosophy (New York: Cambridge
University Press).
Schwarz, N. and Clore, G. R. (1983). “Mood, Misattribution, and Judgments of Well-Being:
Informative and Directive.” Journal of Personality and Social Psychology 45: 513–23.
Solomon, R. (1980). “Emotions and Choice.” In Explaining Emotions, ed. A. Rorty (Los Angeles,
CA: University of California Press), 251–82.
Solomon, R. (1984). The Passions: The Myth and Nature of Human Emotions (New York:
Doubleday).
Zagzebski, L. (2003). “Emotion and Moral Judgment.” Philosophy and Phenomenological Research
66: 104–24.
Zagzebski, L. (2004). Divine Motivation Theory (Cambridge: Cambridge University Press).
Zagzebski, L. (2010). “Exemplarist Virtue Theory.” Metaphilosophy 41: 41–57.
Zagzebski, L. (2012). Epistemic Authority: A Theory of Trust, Authority, and Autonomy in Belief
(New York: Oxford University Press).
12
Self-Empathy and Moral Repair
Nancy Sherman
‘An early morning run can make you feel like you are one step ahead of the world.’
Peter Goldie in conversation with me one morning at The Hague, May 20111
1. Introduction
My subject concerns self-empathy as a component in the healing of moral injury. The
idea of self-empathy may strike some as odd. As an epistemic notion, empathy is typically
directed at another and is a vehicle for understanding how to see the world from her par-
ticular corner. As an affective mode, it is a way of being able to share someone else’s emotion
and so have congruent feelings. But what work does empathy do when directed at oneself?
Even if we are never fully in sync with our own minds and emotions, for most of us, there
is not the same kind of gap within us as there is between us. The idea of empathizing with
oneself, some might say, is redundant. I want to suggest that this is not so. Even if we are
already ‘in sync’ with many aspects of ourselves, there are still corners we do not peek into
because their contents are too alien, and so possibilities for change are thereby closed off.
Self-empathy can play a role in peering into those corners and opening doors. It can be a
way of calling out to oneself that one is hurt and in need of attention and response.2
The work of empathy is especially important in responding to moral injury. Among
the moral injuries I have in mind are those expressed through guilt feelings, specifically
subjective guilt that does not always accurately track culpability and, too, the shame
that that guilt often covers up. I have written on this subject in The Untold War (2010),
and in particular, on how soldiers experience guilt as a response to good and bad moral
luck, such as surviving buddies or causing their deaths by accident.3 This kind of guilt
is obviously not restricted to war, but war provides an important and timely context for
studying it. Imposing guilt on oneself, I argue, is a way of taking responsibility, even if
1
This chapter was inspired by conversations with Peter Goldie during conferences we attended in May
2011 at Geneva and at The Hague. We were planning on talking again, but we never did get the chance.
2
Here, I am influenced by the work of Kukla and Lance (2009) and Macnamara (2012) on Strawsonian
models of reactive attitudes. See Strawson (1962).
3
For other related pieces, see Sherman (2013, 2014a, 2014b, 2014c, 2015).
184 Nancy Sherman
one overimposes it. It can be morally fitting and admirable, even when epistemically
ill-fitting.4 In some cases, it is a way of defending against a harder-to-uncover feeling
of falling short of one’s measure. Though I cannot take up the issue here, my interest
in guilt, shame, and self-empathy is part of a broader interest in reactive attitudes con-
stitutive of holding self and others accountable for morally significant conduct that
impinges on moral agency. By and large, philosophers have focused on negative reac-
tive attitudes (of disapprobation for transgression or falling short), and with only some
notable exceptions, have considered positive emotions (of approbation) that can help
in moral recovery.5 Empathy might be thought of as among those positive emotions,
alongside trust and certain forms of hope in persons. In the context of moral injury,
these latter emotions, each in their own way, can act as positive overtures to self or oth-
ers that at once expose vulnerability and seek responsiveness to one’s needs. Moreover,
we can sometimes ‘grow’ responsiveness in those we address. This is certainly the case
in trust, where if we are a bit wise with regard to whom we trust for what and when,
our very act of trusting can often elicit and reinforce another’s trustworthiness (Jones
2012; also Pettit 1995). Something similar may happen in the case of self-empathy. We
uncover our hurt to ourselves, and in that acknowledgement can sometimes elicit
resources for responding to and ameliorating the suffering. In the case of punishing
guilt, in empathetically reviewing the very evaluations that are at the core of our self-
reproach, we may find room to hold ourselves to account in a more compassionate and
equitable way.
To limn the plan of what follows, I begin with two soldiers’ stories of shame, one
contemporary and one ancient, take up the issue of guilt and its masking of shame, and
then develop a conception of self-empathy, and its component parts, as they figure in
moral repair.
2. The Wounds of Shame

Army Major Jeffrey Hall deployed to Iraq twice, commanding infantry and artillery
units (at the time, at the rank of captain) near Baghdad and Fallujah.6 He signed up
for the army at 17, and at 40, despite having implemented versions of COIN (counter-
insurgency operations) in those last deployments—serving as mayor of a local advisory
council of elders, painting schools and laying sewers, outfitting scores of children with
4
See D’Arms and Jacobson (2000) for clarification of this distinction.
5
For some important philosophical work on guilt and shame, see Morris (1976); Taylor (1985); Williams
(1993); Deigh (1999); Murphy (1999); Nussbaum (2001); Velleman (2001, 2003), and Deonna and Teroni
(2008, 2011). For insightful work on uncovering masked shame in psychoanalytic treatment, see especially
Lansky (1992, 1995, 2003, 2004, 2007) and Lewis (1971). On self-forgiveness, which can be a part of getting
beyond emotions of self-reproach, see Goldie (2011) commenting on Griswold (2007). For work on hope,
see Martin (2014), and on trust, Walker (2006) and Jones (2012). For work on gratitude, see Macnamara
(2012).
6
I interviewed Jeff Hall in September 2010 and several times later that year and more recently. He has
given me permission to disclose his name.
Self-empathy and Moral Repair 185
shoes (who, never having worn them before, had no clue that shoes, or their feet, had a
right and a left), risking life to bring food and medical care to families in need—he still
thinks what he should do in armed conflict, and what he is trained to do as a soldier, is
engage and destroy an enemy.
And yet that was not what his war in Iraq was about. Once Baghdad fell in 2003, he
found himself deep in softer and more cultural methods of warfare, often inadequately
supported, and unclear of the cause or mission. He often felt betrayed by his com-
mand, and that, as a result, he, in turn, was forced to betray those who counted on him.
Stateside, he was diagnosed with severe, near-suicidal post-traumatic stress (PTS),
and with the support of his wife and his commander at home, sought treatment. As
he puts it: ‘You have to understand. My PTS had everything to do with moral injury. It
was not from killing, or seeing bodies severed, or blown up. It was from betrayal, from
moral betrayal.’7
One incident stands out from 2003. A civilian family driving home from church in
Baghdad’s Mansour district crossed a cordon and got caught in the crossfire of a US
attack on a high-value target. Hall’s unit did not carry out the attack, but he was near
the scene at the time. Shortly after the incident, Hall received orders from battalion
headquarters to find the surviving family members and begin to make amends. Over
Chai, an uncle, now guardian of the surviving young daughter in the family, insisted
that what was most important was the return of the bodies for a prompt burial. Hall set
to work, but his efforts were stymied at every turn by an incompetent American occu-
pation bureaucracy (the Coalition Provisional Agency) and an equally incompetent
Iraqi Ministry of Health.
As he waited over a month for the bodies, Hall’s commander issued him a ‘condo-
lence payment’ to be delivered to the family. Hall was speechless when he opened the
envelope and found a piddling $750. He let his commander know how he felt: ‘Sir, they
lost a father, a mother, and a son. And a car that is probably as important to them as the
other losses.’ In disgust, he pushed the money back into the commander’s hand, and
said, ‘You go pay them with this!’ But the commander, cocooned for much of the war
inside Saddam’s former palace in the Green Zone, was unmoved. Hall was under an
unequivocal order to deliver the money.
And so he did. In silence, he passed the uncle the envelope and watched as he counted
the bills, flung them to the ground, and then stormed out of the room. Left alone with
the daughter staring at his incompetence and betrayal, Hall put on his helmet, fastened
his chinstrap, and left the house, dripping in shame.
But the ordeal, and the shame, would not end. The bodies were finally returned to
the family, unembalmed, and rotted beyond recognition by the scorching desert heat.
The family had one last request of Hall. They needed death certificates to finalize the
7
In referring to moral injury, Hall is using a term of art that the US Veteran Affairs and military behav-
ioural health units now use. See, for example, Litz et al. (2009), Maguen and Brett (2012), and Nash et al.
(2011). The work of Jonathan Shay (1994, 2002) has been pioneering in this area.
186 Nancy Sherman
burial. And so Hall returned to the Ministry of Health and secured the certificates. But
on each was stamped in bold red letters: ‘ENEMY’. ‘Can’t you give me something that
doesn’t have “enemy” stamped on it?’ Hall beseeched. ‘No,’ the official curtly replied.
‘They are enemies. They are considered enemies.’
Hall’s story verges on the comedic.8 But the comedy barely lightens the profound
moral injury he suffered. Disarmed of much of his usual arsenal as a warrior, more than
ever, he needed to be able to trust his own basic goodness and have some confidence
that he could compassionately help these non-combatants caught in war. However
much a part of the just conduct of a soldier it is to minimize collateral damage in war
and ameliorate its effects, for Hall, the duty was more basic; it was an intimate duty to
a family he had come to know and care for. And yet he felt thoroughly impotent in that
role. When he says the injury was worse and more lasting than what he suffered from
seeing the detritus of war for three years, what he means, in part, is that the betrayal by
command put him in a position of feeling trapped and helpless, much more powerless
and captive than he had ever felt in facing enemy fire. He was stripped, defenceless,
with nowhere to go. That shame paralyzed him until one day, back home, he simply
could not put on his combat boots. It was at that point that a new, far more benign com-
mander than his previous one got him help.
3. Ajax’s Shame and Prior’s Guilt

The theme of a soldier betrayed by his command is the subject of Sophocles’ Ajax.
Ajax is stripped of his time, his honour and status, when the Greek chiefs vote to award
Achilles’ armour, a prize given to the best fighter, to Odysseus rather than him.9 Ajax
goes ballistic. He has been nakedly shamed before his peers. The fall is steep and pub-
lic. In a pique of blazing rage, he sets out to take revenge on Odysseus and his troops,
and to prove, once and for all, his unmatched skill as a swordsman. But the goddess
Athena blinds him and he flails his sword in the dark, mistaking barnyard animals for
his rivals. He ‘hacked at this chief and that chief ’, recounts Athena. And after tiring of
the slaughter, he took the rest of the beasts captive and tortured them.
Ajax ‘comes to’ in a bloodbath of butchered carcasses and mutilated livestock. He
mocks the sight of himself:
Look at the valiant man! The brave heart!
The one who unflinchingly faced the enemy!
You see the great deeds I have done to harmless beasts?
Oh, the ridicule runs riot against me! (Sophocles 2007: 364–7)
8
‘You couldn’t invent more comedic war narratives’, Rajiv Chandrasekaran, war correspondent and
author of Imperial Life in the Emerald City: Inside Iraq’s Green Zone, said in a seminar at the Wilson Center,
September 2011, reflecting back on that period in the war in Iraq.
9
For a a retelling of the Ajax story, see Woodruff (2011).
There is ironic distance,10 but it fails to insulate. Ajax’s self-evaluation could not be
more unforgiving. True, here he seems to look on at himself as someone in the past. But
his past is not past. It consumes him in the present. He is damned in his eyes. He has
lost his prize—his warrior eminence—and now his wits. In an unparalleled moment in
Greek tragedy, this great Greek general falls on his sword on stage.
The experience of shame, as these two illustrations show, is about being seen and
about having nowhere to hide. Greek etymology is a reminder. Aidôs is related to aid-
oia, genitals. To be ashamed is to be caught without your figleaf. The audience can be
real or imagined. When Aristotle says ‘eyes are upon you’, he should not be read liter-
ally (Aristotle 1984: 1384a35‒1384b1). That is how shame feels. In some cases, shame
can be too toxic to be consciously experienced, screened as a more socially respectable
and manageable feeling of guilt with its presumption of a discrete act of wrongdoing
and promise of redemption through moral repair (Lansky 1995, 2003, 2007). Indeed,
perhaps one way to think of certain instances of epistemically ill-fitting (or irrational)
guilt is as substitutes for shame, sublimations of a sort. So an army commander who
loses a private due to an accidental blast of a turret gun on an army vehicle may be at
most only minimally culpable, though he feels horrific and unabated guilt.
This is a case of what I call ‘accident guilt’ in The Untold War (2010: 89–110). In the
specific case I have in mind, the commander, Captain John Prior, approved, with the
advice of his team of engineers, the use of a marine replacement battery for the army’s
Bradley fighting vehicle. What no one foresaw in using the marine battery was that
turning on the ignition would now cause the current to jump to the turret and auto-
matically fire the gun. The blast scooped out the face of young private, Joseph Mayek,
who did not survive the ordeal. What Prior feels is that he should have been able to take
care of his soldiers better, or as we might put it, that he less than perfectly fulfilled his
imperfect duty of care. Put this way, the feeling may have more the colour of shame
than guilt—the shame of falling short of an ideal he set for himself and that captures for
him his responsibilities of office and role.11
Moreover, while guilt may seem overwrought and epistemically irrational in this
kind of case, shame seems more epistemically fitting—Prior did fall short of an implicit
image of himself as a commander who takes care of his troops. Moreover, the idea of
seeing oneself as a leader who should be able to avoid this kind of malfunction on
his watch is not that far-fetched or grandiose; at least, it does not seem over-idealized
to me, in the way, perhaps, thinking one can avoid enemy-inflicted combat death is.
Epistemically fitting shame, in this regard, seems more permissive than epistemically
fitting guilt and we should be less quick to call it ‘irrational’. Still, shame of this sort
can linger far too long, and be as self-destructive as guilt. That is precisely why it is
10
On narrative and ironic distance, see Goldie (2007, 2011: 87).
11
On a related note, Peter Goldie (2011) has argued that shame may not always have ‘insidious’, ‘globaliz-
ing’ tendencies that deny ‘all moral worth’. Also see Abramson (2010).
188 Nancy Sherman
important to try to unmask it and find ways to own it and tolerate it. Self-empathy
plays a role.
4. Self-Empathy
Much has been written on empathy in the past three decades, and so I will be brief in
my summary as a prelude to my current interests.12
Empathy is a term of fairly recent coinage. It comes into usage at the turn of the
twentieth century with the translation by Titchner of the German word Einfühlung—
to enter into a feeling, a term itself first used by Robert Vischer in 1873 in the context of
the psychology of aesthetics and developed by Theodor Lipps in the context of how we
know other minds.13 Two prominent models of empathy (alluded to at the beginning
of this chapter) have emerged in recent years as something of competitors in the psy-
chological and philosophical literature. The first is empathy as vicarious arousal or con-
tagion. The key historical figure is David Hume and his notion of sympathy (though
what he means is what we would now call ‘empathy’) as a mechanism that enables us
to ‘catch’ another person’s affect. We know others’ emotions by coming to feel qualita-
tively similar or congruent emotions. Hume’s metaphor is intuitive: ‘We are attached to
another, as if by a cord, with movement at one end reverberating at the other, causing
a fainter impression of the original feeling’ (1968: 316–24). The second camp, led by
Adam Smith (1976), conceives of empathy in more robust, cognitive terms. Empathy
(again, ‘sympathy’ is his term) is a process that engages imagination, requiring simula-
tion and the taking up of roles or perspectives. We come to know another’s emotions by
trading places ‘in fancy’ as Smith (1976: 51) puts it, we ‘beat time’ with their hearts. But
Smith insists that the swap is not only situational, but also dispositional:14 We not only
stand in another’s shoes. We try to become them in their shoes: to ‘enter, as it were, into
his body and become in some measure the same person with him’ (Smith 1976: 48).15
How do these models fare with respect to self-empathy, and in particular, its features
relevant to surmounting overly harsh self-reproach? One obvious worry for the con-
tagion model is that it suggests a picture of empathy as a repetition of the same stuck,
often intrusive feeling.16 The idea of emotional fixity or stubbornness is part of a more
general worry Peter Goldie raises about the inbuilt biases of emotional construals, or
ways of ‘seeing as’, that predispose one to judgements (as perceptions do), but in some
12
For my own overview of the subject with discussion of the literature and bibliography, see Sherman
(1998a, 1998b, 1998c).
13
For discussion, see Eisenberg and Strayer (1987), Lipps (1903), and Titchener (1909). For Freud and his
interest in empathy, see Pigman (1995) and Freud (1986: 325).
14
For reflections on becoming another, see Bernard Williams’ collection (1973), in which appears the
important article, ‘Imagination and the Self ’.
15
On the notion of ‘becoming’ the other person and the therapeutic work of empathic resonance and
mirroring, see the important work of psychoanalytic theorist, Heinz Kohut (1971, 1977, 1984).
16
At its most worrisome, a retraumatization. See Freud on repetition compulsion in his ‘Beyond the
Pleasure Principle’ (Freud, 1955).
cases, tell us something we do not believe.17 As Goldie (2004: 99) puts it, emotional
subjects tend to confirm rather than disconfirm their evaluative construals: ‘The feel-
ing directed toward the object of the emotion, and the related perception of the object
as having the [evaluative] property, tend to be idées fixes to which reason has to cohere.
The phenomenon is a familiar one: when we are afraid, we tend unknowingly to seek
out features of the object of our fear that will justify the fear.’18 So we have a tendency
to build an ‘epistemic landscape’ that coheres with an evaluation and feeling. We lock
ourselves into a specific emotional take. Self-empathy, as a contagious re-experience
of emotion, may exacerbate a tendency that we already have and that itself requires
intervention.
Similar worries emerge for the simulation view of empathy, for it would require that
we take up, again, the very perspective from which we are trying to free ourselves. In the
cases I detailed previously, the emotional subject’s focus is framed by guilt and shame
that ‘capture and consume attention’, to use Michael Brady’s (2009) felicitous terms.
Self-empathy requires dwelling again in that perspective, and so re-experiencing the
same emotions. In the case of traumatic emotions, it may involve retraumatization.
These objections may be limited, but they make clear that if a notion of self-empathy
is to be part of a model of emotional and moral growth, something more than simulat-
ing and re-experiencing traumatic events and emotions (whether through narration
or other representational forms such as artwork or dance) is required. Here, not sur-
prisingly, the notion of empathy in psychotherapy is helpful. Arguably, psychotherapy
of various stripes, and especially psychodynamic models, depends on a patient revisit-
ing and reliving painful emotions, but characteristically in the context of an empathic
listener who bears witness—that is, who can recognize and acknowledge the pain as
well as the fragility in exposing it, communicate that acknowledgement through a
signalling of compassion, and help break the repetition and defences through vari-
ous interventions and gentle corrections of bias and framing. The therapist’s empathy
involves the ‘tracking’ of a patient’s emotion—sometimes through her own congruent
re-enactments or counter-transferences (Chused 1991; McLaughlin 1991), other times
more cognitively. But it also involves a conveyed sympathy of sorts, compassion, trust,
rapport, and a non-judgemental stance that help build a ‘working alliance’ (Greenson
1967). Empathy, in this rich context, involves access but also benevolence and trust.19
That stance is both protective and transformative, helping the patient safely to remem-
ber, revisit, and feel painful reactions to traumatic events, but also to reconstrue what
happened in ways that may involve fairer self-judgement and less rigid notions of suc-
cess and failure that ultimately help loosen self-destructive feelings.
17
‘Construal’ is Robert Roberts’ term for the cognitive content of an emotion. For how Roberts distin-
guishes that notion from a stricter judgement, see, for example, Roberts (2013).
18
See also Brady (2007).
19
And so the therapist is not just a blank screen or withholding (and ‘abstinent’), on the traditional
Freudian view. See Sherman (1995b, 1998b, 1998c).
190 Nancy Sherman
All this is relatively familiar stuff. Less familiar is the notion of self-empathy and what
role it can play in moral healing, not as a competitor or replacement for second-person
empathy and its role in formal or informal therapy, but as something in addition
that has an important place in its own right. One way to think about self-empathy
is as a conceptually or causally derivative notion. It is a first-personal application of
something we know best in the second-personal case and may come to experience
through second-personal interactions. So an individual may come to self-empathy
by internalizing a second-personal instance of it, as when a patient learns a measure
of self-empathy through the empathy of a therapist toward her. In this case, she may
internalize another’s stance. But she may also internalize her own stance that she takes
toward others. So a rape victim in a support group may come to feel self-empathy only
after first feeling empathy toward others who were similarly victimized. ‘Oh my God,
that’s what happened to me’, she might come to say.20 That externalization of experi-
ences similar to her own, and ensuing empathy toward others, may enable her now to
look at herself through new eyes. Second-person empathy, both the receiving and giv-
ing of it, may thus prepare one for first-person empathy. One gains an outside perspec-
tive on oneself, and one that is qualitatively different from the punishing and shaming
stance that has held one hostage until now. Veteran support groups may similarly ena-
ble self-empathy through the validating experience of empathizing and being empa-
thized with. Given the machismo culture of the military, the work of empathy and
self-empathy within these groups can be quite powerful.
In thinking about self-empathy, it is useful to turn to Aristotle’s remarks about
self-love (or self-friendship) in Nicomachean Ethics IX.8 (Aristotle 1984). He is aware
that the idea of self-love may be a bit strained, both because it requires that we stand
as a subject and object towards ourselves, but more importantly because it connotes a
problematic sort of selfishness. However, there is room for a good kind of self-love, he
insists, that is the capacity of a self to listen to reason with equanimity. He associates this
kind of self-love with nobility and the readiness to sacrifice for others characteristic of
excellence of character (virtue) and practical reason, and contrasts it with that baser,
more problematic kind of self-love that involves taking material advantage for oneself.
However, in the soldiers’ stories that are my focus, there is no shortage of nobility and
sacrifice. If anything, that aspiration for virtue is too hard-driving, giving way to too
much self-punishment when luck runs out. Still, Aristotle’s idea of finding the right way
to befriend oneself is useful here. The best kind of friendship—that of character (or vir-
tue) friendship, he tells us—serves as an arena for character critique and moral growth
(Aristotle 1984: 1172a10‒1172a12),21 and like all friendship, requires positive feelings or
affection (philêsis) toward one’s object and feelings of goodwill (eunoia).
Self-empathy, as I am imagining it, involves a similar kind of self-friendship, and
minimally requires a certain level of goodwill or compassion. I am also imagining
20
I thank Susan Brison for this point.
See Sherman (1995a: 187‒234) and Sherman (1989: 118‒56).
21
it in the service of moral growth and in the cases I have limned of moral repair, of
being called forth when one has held oneself accountable in a way that begins to seem
unfair, or at least requires further reconsideration and reassessment of the nature of
that accountability. And so the self-empathy I have in mind emerges as part of a moral
process and is earned as a counterweight to overbearing self-judgement. This helps
deflect popular images of self-empathy as merely self-kindness or self-compassion, a
going-gentle-on-oneself, or, relatedly, a form of self-esteem, in the sense of a contrived
boost to a deprecated sense of self, or again, a narcissistic self-absorption in the sense
of a gaze that turns too much to self and not enough to others (Neff 2003). But equally,
I am not thinking of self-empathy as a minimization of self, a putting of self in its place,
as Cicero redacts the Epicurean teaching: these are ‘the restrictions under which all
humans live’, ‘you are not the only one to have this happen’, ‘to endure these things is
human’ (Cicero 2002: sections 3.77, 3.78, 3.34). Rather, I am envisioning self-empathy
as an emotional attitude that predisposes one to a fairer self-assessment, especially,
in the cases I have focused on, where luck and accident and power ceded to others,
squeeze out one’s moral efficacy or cast doubt on one’s goodness.
As a kind of reactive emotion, self-empathy operates by calling out to oneself, in
the way that emotions and not less charged mental states do (Hurley and Macnamara
2010), reining in our attention on what is salient and morally significant to our moral
agency and well-being. We entreat ourselves to look back at the specific evaluations
involved in our self-condemnations and to respond by first reopening the cases. We
call out to ourselves to become affectively re-engaged in the anguish of our guilt or
shame feelings, and cognitively re-engaged in appraising circumstances from a new
perspective that time and distance allows. Our guilt calls out for a reply in terms of self-
empathy in the way that resentment asks those who have transgressed us to now give
us reasons for reassurance or trust (Walker 2006).22
The notion of self-esteem does not get at this idea, nor does self-respect. The under-
lying notion behind self-respect is that one is not servile or subordinate to others, but
rather an equal among equals. Yet one may have no doubt about that, stand in no need
of its reaffirmation, and yet still need a fairer hearing about whether ‘could-have-done’s
entail ‘should-have-done’s in the case of guilt feelings, or about how fixed or severe the
damage done to self is in the case of shame feelings. Put this way, one might still require
of oneself a new way of holding oneself accountable that involves less reproach and
more trust or confidence in oneself. In a way, one requires a stance that allows one to
take up a more positive reactive attitude toward oneself.
As I suggested earlier, self-empathy seems to presuppose a capacity to take up a per-
spective toward self that allows for narrative distance. Peter Goldie has written insight-
fully about a ‘narratable’ conception of self:
On the call and response nature of reactive attitudes, see Macnamara (2012).
22
192 Nancy Sherman
We are able to deploy in thought and feeling a narratable conception of oneself: with a narratable
past, which one now remembers, interprets, and evaluates in various ways; with a present; and
with a narratable future, about which one can make plans, have hopes and aspirations, and so on.
This conception of oneself is the narrative sense of self. (Goldie 2011: 86)23
One is ‘in effect seeing oneself as another’ (Goldie 2011: 86). And this creates an eval-
uative and epistemic gap essential to reappraisal and re-evaluation: ‘One now knows
what one did not know then . . . one can now take an evaluative stance which differs
from the stance that one then took’ (Goldie 2011: 87).
My notion of self-empathy adds to this narratable conception of self an ability to see
from beyond or outside without radical dissociation or alienation from the old self and
its ways of seeing and feeling. That is part of the force of the notions of affective and
cognitive re-engagement. In this sense, self-empathy allows for self-reintegration (a
kind of connectedness), rather than serial reinvention or radical conversion. Though
one may have psychologically and emotionally moved on, one can still remember how
one saw or thought and felt things. One can still be affected, even if slightly, in some
such way. As I am imagining it, in a case like Prior’s, he can still feel a bit of the bite of
the old guilt. It does not rattle him any longer, but in narrating the story he is nonethe-
less affected by the remembering, in some way as he once was. That is not all he feels
with respect to the events. He now sees circumstances far more completely and his
emotions reflect those changed appraisals. But it is not just that he is now tolerating
what he used to feel or think or accepting and owning it for what it was, as therapists
might put it. Rather, he also knows how it feels, as if in muscle memory. That is a part of
his self-empathy. Similarly, in Hall’s case, we can imagine him experiencing a flush of
shame as he retells me the story and brings to mind the faces of the father and daughter
or hears the commander’s intonation as he gives him the order to deliver the envelope.
The shame is no longer intrusive and paralyzing, as it is in post-traumatic stress. But it
is accessible. Self-empathy, as I am using the term, in addition to a compassionate, less
judging regard, involves this kind of affective, empathic access.24
Obviously the degree of access will depend on how changed a person’s psychologi-
cal make-up has become. Access is along a continuum. When the narrative distance
is great, an individual may be able to remember only coldly and cognitively, with little
emotional valence. She is not much alive to how she felt then. At the extreme, a limit to
self-empathy has been reached, at least for the while.
23
For a sharp and lively criticism of the idea of a narrative self, see Strawson (2004).
24
See Schechtman (2001), whose work I came upon in writing this chapter. She invokes Richard
Wollheim’s (1984) notion of ‘event memory’ (discussed in The Thread of Life) that as she explains, ‘is not a
cold cognitive relation to the past, but one which is thoroughly infused with affect’ (Shechtman 2001: 248).
Wollheim describes his World War II soldier years, driving by mistake into the German lines in August
1944: having described the event and the memory of it, he says, ‘and as I remember feeling those feelings, the
sense of loss, the sense of terror, the sense of being on my own, the upsurge of rebellion against my fate, come
over me, so that I am affected by them in some such way as I was when I felt them on that remote summer
night’.
Before concluding and summing up this exploratory account of self-empathy,

I want to say a few very brief words about self-forgiveness. It might be argued that
what I am really after in appealing to self-empathy is a reactive attitude of forgive-
ness toward self.25 But even if a notion of self-forgiveness makes sense in cases where
one may have transgressed against another, it is an ill-fitting notion when there is no
real intentional wrongdoing for which to demand forgiveness. As a more general idea
of forswearing anger and blame (Calhoun 1992; Roberts 2003; Griswold 2007; Goldie
2011), it may have its place in the surmounting of self-reproach, irrespective of whether
that reproach is deserved or not. But even if it does, self-forgiveness does not expose
the more complex evaluative and affective mechanism I have been exploring, of sur-
mounting certain emotions with compassion while preserving empathic access to
them.
And why is that access important or worth preserving? I suspect it is because I
do not believe that difficult conflicts and the emotions that express them are ever so
completely resolved that all residue of such conflicts disappear. Self-empathy is a way
of remaining attuned to those tugs and pulls as they morph into new shapes on new
landscapes. It is a compassionate form of keeping self-vigil. That said, we may also
need self-empathy in the cases where we have, in fact, transgressed or acted morally
wrongly, and forgiveness, toward self or from others, does not seem quite right—per-
haps because the wrongdoing was so heinous.
It is worth summing up the various elements (and historical roots) of the notion of
self-empathy I have been sketching. That there are these distinct elements suggests that
self-empathy may be a composite notion that resists easy unification. These are the ele-
ments I have discussed (though not necessarily in this order).
Affective access to past emotionally imbued experiences, such that one is able to
‘feel’ and recapture something of the tone and valence of those experiences. This is the
force of ‘being alive’ to those experiences, not numb or dissociated. This picks up on
Hume’s notion of empathy as a way of ‘catching’ affect.
Cognitive and imaginative engagement such that one can reinterpret, reframe, and
so reconstrue emotionally powerful and, in some cases, traumatic experiences. This
will often involve a reassessment of the evaluative dimensions of that experience—
such as one’s sense of betraying or being betrayed, or letting oneself or others down,
and so on. This idea resonates with Smith’s cognitive gloss on empathy as involving
imagination or ‘fancy’.
Compassionate and benevolent regard toward self, especially in cases where it is
needed to counter overly harsh judgement. In the cases I am interested in, this atti-
tude can often amount to a fairer and more equitable self-assessment of accountabil-
ity important for moral repair. Relevant here is Aristotle’s notion that all friendships
involve feelings of affection and goodwill, and that the best friendships provide arenas
for moral growth.
For the coherence of that notion, see Goldie (2011).

25
194 Nancy Sherman
Reactive attitude structure, in the sense that self-empathy is an emotionally

charged way of calling out to oneself. In the cases I have developed, it is a way of mor-
ally entreating oneself to reconsider how one holds oneself accountable. One is disclos-
ing to oneself a potential misjudgement or unmerited self-reproach, and is demanding
a change from blame to credit for doing what was at the time reasonably regarded as
appropriate and best.26 This idea borrows from the rich literature on reactive attitudes
that Strawson initiated.
The notion of a narratable conception of self highlights the idea that one knows
now what one did not know then. A narratable conception of self involves a historical
perspective, such that one now has an epistemic and evaluative advantage that only
time affords.27 This notion builds on Goldie’s work on narrative.
Self-forgiveness may figure as a companion notion in my account of self-empathy.
However, forgiveness typically connotes an objective wrongdoing that one for-
swears and seeks atonement for as a condition of re-entry into the moral community.
However, insofar as the kinds of moral injuries I have been focusing on do not typi-
cally involve objective wrongdoing, self-forgiveness does not seem apt. Granted, I have
spoken of self-exoneration in places, but I am bending that term to capture a sense of
release from reproach and a move toward credit giving and self-trust, that does not
itself involve a commitment to the fact of a wrongdoing.
Relatedly, and putting together some of these elements, perhaps one way to cap-
ture the move from negative to positive self-reactive attitudes that I am seeking in this
account of self-empathy goes roughly like this: at work is a developmental process in
which we typically experience shame or guilt in the non-perfect fulfilment of imper-
fect duties, and then through a subsequent shift in perspective toward self that involves
compassion and presupposes a narratable (historical) conception of self, we come to
reconstrue (and accept) the limits of our agency and accept what we have done as hon-
ourable or even creditworthy. So, for example, I could not save my buddy, but I was
still a good soldier and did nothing that intentionally or through negligence or incom-
petence or self-serving ends, exposed him to undue risk. This reconstrual and new-
found compassion is constitutive of the notion of self-empathy I have been developing.
6. Conclusion
Through narratives, and their analysis, I have asked us to look seriously at moral inju-
ries, including those that may seem only apparent, where we do no intentional wrong,
but yet feel as if we have betrayed others and ourselves. Soldiers know these moral
26
See Macnamara (2012) on the general view of reactive attitudes as having a call and response structure.
See Trip Glazer’s (unpublished) ‘Grunts, Groans, and Other Forms of Moral Assessment’ for further discus-
sion of this notion.
27
Peter Goldie refers to this as ironic. Though this kind of stance may be necessary for irony, it does not
seem sufficient. I thank Sabine Roeser for comments on this.
wounds well. They routinely impose moral responsibility on themselves in the face
of factors that make light of their own agency, whether flukish accident, the tyranny
of bureaucracy and public indifference, gappy intelligence, or all too lethal high-tech
and low-tech weaponry. Moral luck morally injures. And it begs for healing, in part,
through the consolations of self-empathy that call for responding to a past self with
hope and renewed trust in oneself.28
References
Abramson, K. (2010). ‘A Sentimentalist’s Defense of Contempt, Shame, and Disdain’. In The
Oxford Handbook of Philosophy of Emotion, ed. P. Goldie (Oxford: Oxford University Press,
189‒214.
Aristotle (1984). The Complete Works of Aristotle: The Revised Oxford Translation, ed. J. Barnes
(Princeton: Princeton University Press).
Brady, M. (2007). ‘Recalcitrant Emotions and Visual Illusions’. American Philosophical Quarterly
55: 273–84.
Brady, M. (2009). ‘The Irrationality of Recalcitrant Emotions’. Philosophical Studies: An
International Journal for Philosophy in the Analytic Tradition 145: 413–30.
Calhoun, C. (1992). ‘Changing One’s Heart’. Ethics: An International Journal of Social, Political,
and Legal Philosophy 103: 76–96.
Chused, J. (1991). ‘The Evocative Power of Enactments’. Journal of the American Psychoanalytic
Association 39: 615–39.
Cicero (2002). Cicero on the Emotions: Tusculan Disputations 3 and 4, ed. M. Graver
(Chicago: University of Chicago Press).
D’Arms, J. and Jacobson, D. (2000). ‘The Moralistic Fallacy: On the “Appropriateness” of
Emotions’. Philosophy and Phenomenological Research 61: 65–90.
Deigh, J. (1999). ‘All Kinds of Guilt’. Law and Philosophy: An International Journal for
Jurisprudence and Legal Philosophy 18: 313–25.
Deonna, J. and Teroni, F. (2008). ‘Shame’s Guilt Disproved’. Critical Quarterly 50: 65–72.
Deonna, J. and Teroni, F. (2011). ‘Is Shame a Social Emotion?’ In Self-Evaluation: Affective and
Social Grounds of Intentionality, ed. A. Ziv, K. Lehrer, and H. Schmid (New York: Springer
Science + Business Media), 193‒212.
Eisenberg, N. and Strayer, J. (1987). Empathy and its Development (Cambridge: Cambridge
University Press).
Freud, S. (1955 [1920]). ‘Beyond the Pleasure Principle. In The Standard Edition of the Complete
Psychological Works of Sigmund Freud, trans. J. Strachey (London: Hogarth Press).
Freud, S. (1986). The Complete Letters of Sigmund Freud to Wilhelm Fliess, 1887–1904, ed.
J. Masson (Cambridge, MA: Harvard University Press).
Glazer, T. (unpublished). ‘Grunts, Groans, and Other Forms of Moral Assessment’.
28
I wish to thank Francisco Gallegos for help in researching and editing this chapter and Kris Bradley for
earlier help with reference gathering.
196 Nancy Sherman
Goldie, P. (2004). ‘Emotion, Feeling, and Knowledge of the World’. In Thinking about
Feeling: Contemporary Philosophers on Emotions, ed. R. Solomon (New York: Oxford
Goldie, P. (2007). ‘Dramatic Irony, Narrative, and the External Perspective’, Philosophy: The
Journal of the Royal Institute of Philosophy 60 (supp): 69–84.
Goldie, P. (2011). ‘Self-Forgiveness and the Narrative Sense of Self ’. In The Ethics of Forgiveness: A
Collection of Essays, ed. C. Fricke (New York: Routledge), 81‒94.
Greenson, R. (1967). The Technique and Practice of Psychoanalysis (New York: International
Universities Press).
Griswold, C. (2007). Forgiveness: A Philosophical Exploration (Cambridge: Cambridge
University Press).
Hume, D. (1968 [1739]). A Treatise of Human Nature (London: Oxford University Press).
Hurley, E. and Macnamara, C. (2010). ‘Beyond Belief: Toward a Theory of Reactive Attitudes’.
Philosophical Papers 39: 373–99.
Jones, K. (2012). ‘Trustworwthiness’. Ethics 123: 61–85.
Kohut, H. (1971). The Analysis of Self (Madison, CT: International Universities Press).
Kohut, H. (1977). The Restoration of the Self (Madison, CT: International Universities Press).
Kohut, H. (1984). How Does Analysis Cure? (Chicago: University of Chicago Press).
Kukla, R. and Lance, M. (2009). Yo and Lo! (Cambridge, MA: Harvard University Press).
Lansky, M. (1992). Fathers Who Fail: Shame and Psychopathology in the Family System (Hillsdale,
NJ: Analytic Press).
Lansky, M. (1995). ‘Shame and the Scope of Psychoanalytic Understanding’. American Behavioral
Scientist 38: 1076–90.
Lansky, M. (2003). ‘The “Incompatible Idea” Revisited: The Oft-Invisible Ego-Ideal and Shame
Dynamics’. The American Journal of Psychoanalysis 63: 365–76.
Lansky, M. (2004). ‘Trigger and Screen: Shame Dynamics and the Problem of Instigation in
Freud’s Dreams’. The Journal of the American Academy of Psychoanalysis and Dynamic
Psychiatry 32: 441–68.
Lansky, M. (2007). ‘Unbearable Shame, Splitting, and Forgiveness in the Resolution of
Vengefulness’. Journal of the American Psychoanalytic Association 55: 571–93.
Lewis, H. (1971). Shame and Guilt in Neurosis (New York: International Universities Press).
Lipps, T. (1903). ‘Einfühlung, Innere Nachahmung und Organempfindung’. Archiv für gesamte
Psychologie 1: 465–519.
Litz, B., Stein, N., Delaney, E., Lebowitz, L., Nash, W. et al. (2009). ‘Moral Injury and Moral
Repair in War Veterans: A Preliminary Model and Intervention Strategy’. Clinical Psychology
Review 29: 695–706.
McLaughlin, J. (1991). ‘Clinical and Theoretical Aspects of Enactment’. Journal of the American
Psychoanalytic Association 39: 595–614.
Macnamara, C. (2012). ‘ “Screw You” & “Thank You” ’. Philosophical Studies 130: 893‒914.
Maguen, S. and Brett, L. (2012). ‘Moral Injury at War’. National Center for PTSD <http://www.
ptsd.va.gov/professional/pages/moral_injury_at_war.asp> (accessed 6 March 2012).
Martin, A. M. (2014). How We Hope (Princeton: Princeton University Press).
Morris, H. (1976). On Guilt and Innocence (Berkeley: University of California Press).
Murphy, J. (1999). ‘Shame Creeps Through Guilt and Feels Like Retribution’. Law and
Philosophy: An International Journal for Jurisprudence and Legal Philosophy 18: 327–44.
Nash, W., Krantz, L., Stein, N., Westphal, R., and Litz, B. (2011). ‘Comprehensive Soldier Fitness,
Battlemind, and the Stress Continuum Model: Military Organizational Approaches to
Prevention’. In Caring for Veterans with Deployment-Related Stress Disorders, ed. J. Ruzek,
P. Schnurr, J. Vasterling, and M. Friedman (Washington, DC: American Psychological
Association), 193‒214
Neff, K. (2003). ‘Self-Compassion: An Alternative Conceptualization of a Healthy Attitude
Toward Oneself ’. Self and Identity 2: 85–101.
Nussbaum, M. (2001). Upheavals of Thought: The Intelligence of Emotions (Cambridge: Cambridge
University Press).
Pettit, P. (1995). ‘The Cunning of Trust’. Philosophy and Public Affairs 224: 202–25.
Pigman, G. (1995). ‘Freud and the History of Empathy’. The International Journal of
Psycho-Analysis 76: 237–56.
University Press).
Roberts, R. (2013). ‘Justice as an Emotion Dispositon’. In On Emotions: Philosophical Essays, ed.
J. Deigh (New York: Oxford University Press), 14‒28.
Schechtman, M. (2001). ‘Empathic Access: The Missing Ingredient in Personal Identity’.
Philosophical Explorations: An International Journal for the Philosophy of Mind and Action
4: 95–111.
Shay, J. (1994). Achilles in Vietnam: Combat Trauma and the Undoing of Character
(New York: Touchstone).
Shay, J. (2002). Odysseus in America: Combat Trauma and the Trials of Homecoming
(New York: Scribner).
Sherman, N. (1989). The Fabric of Character: Aristotle’s Theory of Virtue (Oxford: Oxford
University Press).
Sherman, N. (1995a). Making a Necessity of Virtue: Aristotle and Kant on Virtue
(New York: Cambridge University Press).
Sherman, N. (1995b). ‘The Moral Pespective and Psychoanalytic Quest’. Journal of the American
Academy of Psychoanalysis 23: 223–42.
Sherman, N. (1998a). ‘Concrete Kantian Respect’. Social Philosophy and Policy 15: 119–48.
Sherman, N. (1998b). ‘Empathy and Imagination’. Midwest Studies in Philosophy 22: 82–119.
Sherman, N. (1998c). ‘Empathy, Respect, and Humanitarian Intervention’. Ethics and
International Affairs 12: 103–19.
Sherman, N. (2010). The Untold War: Inside the Hearts, Minds, and Souls of our Soldiers
(New York: Oxford University Press).
Sherman, N. (2013). ‘Guilt in War’. In On Emotions: Philosophical Essays, ed. John Deigh
(New York: Oxford University Press), 179‒97.
Sherman, N. (2014a). ‘The Moral Psychic Reality of War’. In Reading Walzer, ed. Y. Benbaji and
N. Sussman (London: Routledge), 302‒27.
Sherman, N. (2014b). ‘The War Within: Moral Injury and Guilt’. In Heroism and the Changing
Character of War, ed. Sibylle Scheipers (Basingstoke: Palgrave Macmillan).
Sherman, N. (2014c). ‘Recovering Lost Goodness: Shame, Guilt, and Self-Empathy’.
Psychoanalytic Psychology.
Sherman, N. (2015). Making Peace with War: Moral Injury and Moral Repair (New York: Oxford
University Press).
198 Nancy Sherman
Smith, A. (1976 [1759]). The Theory of Moral Sentiments (Indianapolis: Liberty Classics).

Sophocles (2007). Ajax, trans. P. Meineck and P. Woodruff. In Four Tragedies (Indianapolis:
Hackett).
Strawson, G. (2004). ‘Against Narrativity’. Ratio: An International Journal of Analytic Philosophy
17: 428–52.
Strawson, P. F. (1962). ‘Freedom and Resentment’. Proceedings of the British Academy 48: 1–25.
Taylor, G. (1985). Pride, Shame, and Guilt (Oxford: Oxford University Press).
Titchener, E. (1909). Lectures on the Experimental Psychology of the Thought-Processes
(New York: Macmillan).
Velleman, J. (2001). ‘The Genesis of Shame’. Philosophy and Public Affairs 30: 27–52.
Velleman, J. (2003). ‘Don’t Worry, Feel Guilty’. Philosophy: The Journal of the Royal Institute of
Philosophy 52 (supp): 235–48.
Walker, M. J. (2006). Moral Repair: Reconstructing Moral Relations after Wrongdoing
Williams, B. (1973). Problems of the Self (Cambridge: Cambridge University Press).
Williams, B. (1993). Shame and Necessity (Berkeley: University of California Press).
Wollheim, R. (1984). The Thread of Life (Cambridge, MA: Harvard University Press).
Woodruff, P. (2011). The Ajax Dilemma: Justice, Fairness, and Rewards (New York: Oxford
University Press).
13
Emotions and the Virtues of
Self-Understanding
Michael Lacewing
‘When beliefs . . . become hostage to desires and wishes, they do so only as the

result of hidden and indirect processes, against which the disciplines of the virtues
of truth are directed’.
Williams (2002: 83)
1. The thought that emotions play a central role in moral epistemology goes back at
least to Aristotle. It is, of course, the centrepiece of various non-cognitivist theories,
but has more recently been defended by cognitivists on the basis of cognitivist theories
of emotion. Scanlon (1998), for example, talks of ‘judgment-sensitive attitudes’, which
can arise from and be embedded in emotional and conative responses to the world.
Such attitudes present the agent with reasons to favour or disfavour their objects, and
the clarification of reasons, and so which attitudes are appropriate, proceeds by criti-
cal reflection upon them (ch. 1, §12). The response-dependence theories of Wiggins
(1987) and McDowell (1985) provide further examples. I shall assume that desires and
emotions (‘passions’ from now on) are, at the very least, an important source of intui-
tions about moral reasons (I shall focus on reasons from now on), whether one gives a
cognitive or non-cognitive analysis of them.
As one might expect from attributing such a central role to passions, such theo-
ries have remarked upon the relation between moral enquiry and developing
self-understanding (e.g. Lenman 2008). But the theme has not been much elaborated
upon. What follows is a contribution, though from a very specific angle.
2. We naturally seek to understand ourselves and the vision of the world presented in
passions. For example, when reflecting on emotions, we look for a ‘sufficient explana-
tion’, which either shows that the emotion is appropriate, timely, proportionate, and so
on, or explains why it is not. This understanding is the ground of the judgement that
the emotion is appropriate, and so presents some reason to react or act, or that it is
200 Michael Lacewing
inappropriate, and so does not. For desires, we consider whether what is desired is in
fact desirable. It is a commonplace that passions are not always appropriate and may
mislead us regarding what reasons we have to feel, want, and do. Hence, there is a need,
first, to recognize when and how we go wrong, and second, to correct ourselves and, if
possible, prevent or mitigate future mistakes.
One traditional model proposes that, in calmer moments, the apparent good pre-
sented by a passion can come to be understood and rationally evaluated through
reflective introspection. Understanding why one reacts as one does secures some con-
trol over the passion. Passions respond to reasons; reasoned accounts of the object and
of the passion should therefore alter the passion if necessary. The model assumes that
passions are ‘transparent’, and in two ways. First, attention to the passion in introspec-
tive reflection is sufficient for understanding it and the vision of the world and the
good that it presents. The content of passions is unproblematically available for accept-
ance or rejection in reflection. Second, these processes are sufficient to change the pas-
sion—the question ‘what do I feel?’ gives way to the question ‘what ought I to feel?’,
such that the answer to the latter determines the answer to the former.1
Helpful though this is, it egregiously oversimplifies the complexity of the challenge.
It presumes a way of identifying and then getting behind or beyond the misleading
passion by sheer reflection, as though there were some intellectual space free from the
influence of emotions. But passions can skew the ‘epistemic landscape’ without our
awareness (Goldie 2008: 159f.). When this occurs, reflection takes place within the
skewed landscape that seems ‘true’. Thus, even if there were an ‘emotion-free’ intel-
lectual space, the grounds for confidence that one currently occupies it are often lack-
ing. Furthermore, the process of reflection on passions and the reasons they present
is itself deeply informed by further passions—sympathy, feelings of approbation, and
other attitudes—that arise as reflection proceeds. This applies as much to reflecting
upon the appropriateness of a passion as to other forms of moral enquiry (Lacewing
2005). The role passions play in moral enquiry extends to include understandings of
and reflections upon the situation and the self. A third difficulty forms the primary
focus in what follows: passions are not always readily understood in reflective intro-
spection alone; self-understanding is more demanding, and requires engagement
with ‘hidden and indirect processes’ (Williams 2002: 83). The assumption of transpar-
ency is mistaken.2
1
The first sense is discussed and rejected by Cottingham (1998) and Goldie (2008); the second is dis-
cussed and defended by Moran (2001).
2
Much discussed at present are results from social psychology that support a ‘dual process’ model of
psychology in general (Chaiken and Trope 1999; Wilson 2002; Dijksterhuis 2010), and of moral thinking
in particular (Haidt 2001, 2007; Doris et al. 2010). This is not my focus here. As Tiberius (2008: ch. 5), Snow
(2010), and Kennett and Fine (2009) make clear, drawing upon a wealth of empirical studies, while the chal-
lenge requires a significant re-evaluation of the nature and relation of introspection to self-knowledge, it
has been somewhat overstated. There are possibilities for correcting misleading passions that arise from the
kinds of non-conscious, automatic processes discussed in this literature, at least over time, if not always in
the moment.
Emotions and the Virtues of Self-understanding 201
Hursthouse’s famous (1991) treatment of the debate around abortion from the per-
spective of virtue ethics provides a suitable example to begin with. She notes that,
quite aside from the question of whether women have a moral right to have an abor-
tion, in deciding to have an abortion on any particular occasion, a woman may be cal-
lous, light-minded, selfish, self-righteous, or disloyal (1991: 235). We can expect that the
woman’s motivations and emotions that would make the act vicious lead her to under-
stand the situation and possible courses of action in a particular, and inaccurate, light,
and will feed into her deliberations about what to do. If she is unaware of these motiva-
tions, she fails to understand her decision for what it is. This is not improbable, since atti-
tudes towards human life, death, parenthood, sex, and family relationships will influence
attitudes towards abortion (1991: 238), and the passions surrounding such fundamental
issues of human life are ambivalent and painful, to say the least. Thus, attitudes towards
and deliberation on (an) abortion will be fed by many and various passions about other
matters, including ambition, loyalty, fear, frustration, love, jealousy, and ambivalence,
perhaps only tangentially related to the issue at hand. This is especially so when reflect-
ing on a particular case that involves oneself, but can apply as much to forming a judge-
ment about another’s situation. To develop Hursthouse’s example in a classically feminist
way, if a man is unaware of his anxieties surrounding female sexuality, or unaware of
their influence on his views about abortion, then his judgements may be distorted.
The line of thought so far is this: passions heavily inform moral enquiry (especially in
particular cases), but because they can be inappropriate and thus misleading, the acquisi-
tion of moral knowledge requires their refinement. This is no easy matter, as they lack
transparency and can influence other passions and thoughts in complex ways of which we
are unaware.
3. It may be thought that the only or best counterpoint to misleading passions is

provided by a communal form of moral enquiry, a process of mutual correction, as
Wiggins and Lenman emphasize. But Williams (2002: 198) notes that while, in commu-
nal enquiry, ‘we can help sustain each other’s sense of reality, stopping wishes becom-
ing beliefs’, it may also be that ‘I may reinforce your fantasies, and we may conspire in
projecting wishes into a deceptive social hologram’. More generally, the structure of
one’s passions affects the extent to which one can both contribute to and make use of
communal enquiry. Thus the virtues with which we shall be concerned, as a solution
to the problem outlined, are not developed and refined merely by joint ethical enquiry.
4. One wish—perhaps best conceived as the central or underpinning wish—involved

in many of the processes that may distort the passions, or indeed moral enquiry directly,
is the wish to avoid psychological pain, such as anxiety, fear, guilt, shame, and envy, par-
ticularly in relation to questions of self-esteem. There are many means by which this is
done; for example, through distortions in understandings and experience of the self, of
passions and thoughts, of others, or of the world. Psychologists call such means, when
they occur unconsciously and unintentionally, ‘defence mechanisms’ (Vaillant 2000;
Cramer 2006). This idea frames and develops aspects of our common-sense under-
standing of ourselves: we can immediately recognize the descriptions it gives from our
experience.3 This is best shown by example.
Returning to Hursthouse’s example: we noted that attitudes to such fundamen-
tal issues as life, death, parenthood, sex, and families are implicated in deliberations
regarding abortion. A more fertile soil for defensive reactions is scarcely conceivable.
Let us develop the idea that the decision is ‘light-minded’. It fails to take proper account
of the value of human life, perhaps dismissing the death of the foetus as simply the
medical extirpation of unnecessary cells, no more morally taxing than the removal of
a benign tumour. Such an emotional response may defend against an unconscious fear
of taking responsibility for life (one’s own or another’s). Or again, the woman’s deci-
sion is disloyal; perhaps it is even vengeful. Thoughts (which may be perfectly correct
in themselves) that the decision is hers because it is her body may mask or crowd out
grievances about the father, such that even if she is aware of the grievances, she fails to
connect them to her decision. This kind of case applies as well when the decision is not
to abort; for example, the woman may wish for ‘more’ from her man, and having the
child is her way of obtaining it symbolically.
Anna Freud (1936) discusses a case of ‘false altruism’. Her example is of a woman, but
it applies as well to men, so let us vary the gender. A man represses his own wishes, and
projects them onto others. He then strongly identifies with other people. He therefore
expresses great concern for them, but not for himself. He believes it is acceptable to
fulfil their desires, and works to do so, but not to fulfil his own. However, he becomes
annoyed if their desires are frustrated, as if wishes should be fulfilled without hin-
drance; and he becomes angry with people who are not similarly altruistic, as though
this were some personal affront to him. Nietzsche (1998 [1886]: §194) also comments
on false altruism, though with a different emphasis:
In helpful and benevolent people one nearly always finds a clumsy cunning that first rearranges
the person who is to be helped so that, for example, he ‘deserves’ their help, needs their help in
particular, and will prove to be deeply grateful, dependent, subservient for all their help.
For both Anna Freud and Nietzsche, false altruism operates as a defence against var-
ious painful thoughts and feelings related to one’s own neediness.
Our concern here is not with how such people act, but with what they know. Having
this complex of passions and defences leads to a distorted understanding of the moral
situation and one’s reasons for acting. The passions—the desire to help, the empa-
thetic response, the altruistic concern—will present themselves as selfless and ‘other’-
directed. But in fact, there is an inappropriate conceptualization of the needs of others;
a lack of recognition of the parity between oneself and others; a misunderstanding of
one’s relations to others who are also in a position to help; little recognition of others’
3
Experimental evidence for the idea of defence mechanisms and the psychological models it deploys is
neatly marshalled and discussed in Cramer (2006).
responsibility for themselves, and with this, no real understanding of when, why, and
how the needs of others furnish reasons to act on their behalf, the forms of response
that are appropriate (which may be other than simply meeting those needs), and what
reasons (for gratitude? reciprocity? resentment?) one’s action supplies the other.
A third example. In a recent paper on lying, Alessandra Lemma (2005: 738) notes that
‘[i]‌n the moment of the lie, the liar creates the illusion that he can control, and therefore
that he “knows”, what the other will believe and think.’ This may serve different func-
tions, which we may connect to different contexts of defence. In the first, ‘the intent is
to attack and triumph over the duped other. The object [i.e. the other] needs to be con-
trolled and humiliated for the self ’s gratification, often to reverse an earlier experience of
humiliation’ (2005: 738). The liar may be unaware of these needs to control and humili-
ate, perhaps as part of a defence against the pain of their own humiliation (if applicable).
Such lack of awareness will affect an appreciation of when a lie is justified and when it
is not, and more generally, the reasons relating to treating others with respect. In quite
a different context, lying may ‘represent an attempt at communication with [someone]
felt to be emotionally unavailable or inscrutable. The lie is used to substitute the “real”
self felt to be unlovable for a “made-up” version of the self felt to guarantee the [person’s]
love’ (2005: 738). Again, if this is not recognized by the liar, the need to please the other
will distort judgements about appropriate forms of relationship, including the justifica-
tion of the lie. Third, the lie may be used as protection against someone who is intrusive
or controlling (2005: 749–50). In some people, lying becomes the primary way of dealing
with such relationships, and even an automatic response to intimacy (which carries a
potential threat of intrusiveness). Here again, an awareness of what one is doing and why
may help one to see individual situations in a clearer light, and to understand how lying
may well be justified, on occasion, when dealing with intrusive and controlling people.
Defence mechanisms occur far more commonly than we might think. They are used
universally in childhood and adolescence, as they are entirely necessary in psycho-
logical development. In forming a healthy self-esteem, there are many small battles in
which children must first impose their wishes on their experience of reality and later
relinquish such influence, maturing in themselves and their understanding as they
go. The task of development involves giving up such distortions in early adulthood,
but few people do so completely.4 (Defences may remain adaptive in certain situations
in adulthood, such as in situations of unbearable conflict or sudden change—a clear
example is their use in the grieving process.) The use of defence mechanisms is par-
ticularly prevalent in conditions of stress and especially when self-esteem is at issue—
both of which commonly apply in morally challenging situations.
4
Vaillant (1993: 132, Table 4) provides evidence that of those in the top 20% on a scale of psychosocial
adjustment at 65 years old, 50% still use less than mature defences, and the percentage for those lower on
the scale is considerably higher. Cramer (2006: 204) notes that neurotic defences are likely to survive into
adulthood, and remarks on the widespread distribution in ‘normal’ samples of characteristics defining psy-
chological disorders, such as depressive tendencies, phobias, pathological aggression, antisocial traits, and
so on, that are associated with the use of defence (2006: 224, 235).
5. When they occur, defence mechanisms can lead to a misapprehension not only of
what is felt and why, and the nature of the situation to which the passion is a response,
but also, and because of this, of the reasons for acting, feeling, and desiring furnished
by the situation. Thus the agent misunderstands both the reasons why he actually feels
and acts as he does, generating a lack of transparency, and the reasons the situation in
fact gives him to act and feel differently, generating error.
Because these distortions occur unintentionally and unconsciously, any simple form
of directed introspection of the passion itself cannot reliably detect whether the passion
is influenced by defences or not. If defences or their results could be identified in this way,
they would not be sufficiently outside awareness to work. As Cramer (2006: 29–30) puts it,
[t]‌he purpose of attributing one’s own anger or envy to others [in projection] is to absolve oneself of
the discomfort of harboring unacceptable thoughts or feelings. To realize that this negative perception
of others is based on the attribution of one’s own negative emotions would be to acknowledge that one
has such unacceptable emotions; such a realization would be a cause for self-reproach and anxiety.
Furthermore, the distortions they impose on the understanding of the self and oth-
ers need not be singular, confined to this or that situation. Rather, the distortion is in
one’s general view of the world:
A patient of mine inhabited a disappointing world. Although she was quite successful at work,
had friends, and so on, there was no success in the social world that would not be interpreted
by her under an aura of disappointment. If she got a raise at work, it was because the boss was
shamed into it—he really wanted to give someone else in the office a raise, but he felt he had to
give her one to appear fair. If she was invited out for a date, the person had already tried to go out
with others and had failed. If someone congratulated her on some accomplishment, they were
just being polite. And so on. From a distance it is clear to us, as it was not clear to her, how active
she was in understanding her world in ways that were bound to disappoint. (Lear 2003: 48–9)
Given this state of affairs, it is hard to see how the woman can begin to construct an alter-
native set of emotional responses; the ones she has are clearly justified by how the world
is (as she experiences it). Her disappointment, what motivates and sustains it, how it acts
as a defence, and against what—all this is not transparent to her. Whatever the answers,
reflecting on her disappointment will not be enough to reveal or transform them.
6. There is little discussion of defence mechanisms in the literature on moral episte-

mology. It would benefit our understanding to be able to classify their effects in epis-
temological terms, and not only the usual psychological ones. In this connection, it is
worth noting that analysis of Roberts and Wood’s (2007) excellent and extensive dis-
cussion of intellectual vices reveals myriad references to psychic defence, sometimes
explicit but more often implicit. I present three examples here.5
5
Similar connections to those drawn below can be made for dogmatism (195), stolid perserverance
(200), comprehensional rigidity (205), scrupulosity (231), vanity (237), arrogance (243‒5), vices related to
aspects of intellectual autonomy regarding one’s relations to others (265‒6, 268), and a lack of intellectual
generosity (287).
The first comes in their discussion of Plantinga’s (1993: 12) discussion of an example

from Locke, providing a nice historical precedent for the case made here. Locke (1690: IV.
XX.11) imagines a professor’s reaction to a bright student who makes an objection that
undermines his life’s work. Understandably, the professor immediately assumes the stu-
dent is wrong. This unwarranted belief is the product of a defence against the shame of
an academic life in error, and prevents the professor from thinking clearly about the
objection (Roberts and Wood 2007: 96). This may be a momentary lapse in response to
sudden change (which defence mechanisms can enable us to cope with), and as such,
the professor may in time be able to see it for what it is. But if the defence against shame
remains in place, operating unconsciously, his conviction that the student is wrong will
feel epistemically justified, and the possible consequences for his self-esteem, if he is
aware of them at all, dismissed as irrelevant to his rejection of the objection.
Under ‘failures of concern to know’, a vice relative to the virtue of the love of knowl-
edge, Roberts and Wood note that people may decline opportunities to test their cher-
ished beliefs or offer arguments they somehow know to be inadequate (2007: 170).
A negative corollary here is an ‘unvirtuous concern not to know’ the truth. Bernard
Williams (2002: 134) points out that because it is difficult to know whether one has
invested enough effort in finding out the truth about some matter, ‘it is easy to con-
vince oneself that one has taken enough pains’ when one has some internal obstacle
to taking more, such as ‘at the most obvious level, laziness, but, more interestingly,
the desires and wishes that are prone to subvert the acquisition of true belief ’. Where
such desires and wishes or their effects on one’s epistemic states are unrecognized, the
lack of concern to reach the truth, or active concern to avoid it, is either the product of
defence or motivates it. The defences protect one against the anxiety that would result
from facing something unwished for or a threat to one’s sense of self or self-esteem.
Third, Roberts and Wood note ‘the ubiquity of fear and the incidence of cowardice’ in
relation to self-knowledge (2007: 222), but also point out that people fear knowledge of
other kinds—criticisms of pet views, of their work, of facts that are painful to them, of
others disagreeing with them, and of looking bad in front of their colleagues (2007: 219).
In all these cases, what Roberts and Wood do not point out is that what people fear is tied
up with issues of self-esteem, thus increasing the risk of activating defence mechanisms,
which may blind them to the truth about such criticisms, facts, and disagreements.
7. These last remarks indicate that courage may decrease the use of defence mecha-
nisms. But before developing this line of thought, it is worth rehearsing the importance
of communal enquiry (§3). Some authors conceive this as a process of mutual correc-
tion; Williams casts his net more broadly to talk of ‘sustaining each other’s sense of
reality’ (2002: 198), which may be done by means other than correction. It is not only
that if people are not defensive at the same time, or over the same issues, then they may
see each other’s defences for what they are. As we will see, it is also that close relation-
ship and dialogue with others is a means to the deconstruction of defences and the
development of the virtues that enable this.
This is important, since it is not merely identifying errors, but (where relevant)
understanding their basis in defence that is needed; and it is not merely identifying,
but deconstructing, defence mechanisms that is the aim here, for their continued acti-
vation will continue the misapprehensions of the past. Given that the passions are an
important source of judgements about reasons to feel, desire, and act in certain ways,
deconstructing defence mechanisms will contribute to moral enquiry.
We noted previously that defence mechanisms lead to both a lack of transpar-
ency and error, not only with regard to one’s reasons, but also in self-understanding.
The deconstruction of defence mechanisms is therefore, at the same time, a gain in
self-understanding.
8. Baehr (2011: 177) defines intellectual courage as ‘a disposition to persist in or with a

state or course of action aimed at an epistemically good end despite the fact that doing
so involves an apparent threat to one’s own well-being’. This is needed to face and expe-
rience the painful mental states against which defences protect. We often associate
courage closely with strength of will and self-control, with being ‘tough’ and ‘impervi-
ous’ (Baehr 2011: 178), but in this context, that association can be misleading. For what
is required is a kind of ‘letting go’. In contrast to directed introspection or reflection,
this involves an openness to one’s passions, allowing them to ‘surface’. The aim of con-
trol is—at least temporarily—relinquished in favour of an approach of understanding
and engagement with whatever is felt, however painful, inappropriate, or irrational it
seems. The challenge is compounded by the knowledge that one’s emotional life is not
under one’s control, even if it can, over time, be transformed.
But the danger is that defence mechanisms kick in unless one actively seeks to toler-
ate and admit into conscious thought whatever it is one feels. Thus, the courageous
exercise of will lies in making and sustaining a commitment to this openness despite
the pain it brings. To put it into Baehr’s terms, opening oneself to feelings that are pain-
ful, and thus pose an apparent threat to one’s well-being, forms part of the deconstruc-
tion of one’s defences, which is a course of action aimed at the epistemically good end
of moral knowledge. Intellectual courage enables the pursuit of the project of moral
enquiry in the face of the pain it brings.
9. We have seen several times that defences are connected especially to self-esteem.
As often as not, mental states defended against are not intrinsically painful, but become
so because the agent evaluates them negatively. They are inappropriate, unacceptable,
shameful, fearful, and so on. In the final analysis, they are deemed to diminish one’s value;
or in more concrete psychological terms, to make one less lovable or perhaps unlovable;
or yet more specifically, unlovable by particular people. Discovering the existence of such
states for a moment—which may soon be succeeded by re-establishing the defence against
them—is far easier than integrating them into one’s sense of self (Freud 1926: 224). The
deconstruction of a defence requires a reconstruction of the self. Where parts of the self are
defended against because they are inappropriate, and so on, what is needed is acceptance.
This idea needs careful understanding. Acceptance does not mean moral approval.
Someone who only accepts in herself what she can approve of falsely conflates her real psy-
chological self with her ideal self. Defences are often responses to, and supported by, a false
and idealized sense of oneself. Again, it can be the inconsistency with one’s self-image that
makes the unacknowledged mental state painful (Lear 2003: 117–19). But understanding
both the passion and the ideal must precede the judgement that it is the ideal that should
stand firm, and the passion should give way. It cannot be assumed, at the outset, that the ide-
als are supported by the balance of reasons if one’s grasp of those reasons is only secure once
self-understanding is achieved. Ideals can themselves be defences against passions (witness
the cases of false altruism and abortion). Both passion and ideal need to be understood and
allowed—for now—to stand as genuine expressions of the self. The tension between them
may diminish as self-transformation follows upon the deconstruction of defences.
It is, in any case, a mistake to attempt to change one’s passion before one has understood
what there is of oneself in it; self-improvement should follow on from, not substitute for,
self-understanding (Wollheim 2003: 35). Moran (2001) is right that, on many occasions
of reflection, the question ‘what do I feel?’ may be answered by considering reasons to feel
this way or that; that is, ‘what should I feel?’ But on many others, what is felt or desired is
given to one, and awareness of the passion is arrived at quite independently of consider-
ing reasons for feeling or wanting. Of course, this way of feeling may be inappropriate,
and the agent may judge it so. Moran (2001: 59–60) proposes that if such a passion is not
corrected by reflection on what to feel, it is ‘alien’ to the agent. This is epistemologically
and psychologically unsatisfactory (Wollheim 2003: 32f.). The theory of defence explains
why rational reflection fails, for it cannot uncover nor undo what is sustaining the pas-
sion. But the passion is not to be disposed of by the thought that it is unworthy and there-
fore ‘alien’ to what one takes to be one’s ‘true’ self, but is in fact one’s ‘idealized’ self.
We may understand acceptance as involving the work of a range of intellectual vir-
tues. First, there is the resistance to the influence of ideals (and desires to be ideal) on
beliefs about the self, deriving from a ‘virtuous concern to know’. Second, the truth is
accorded its appropriate importance, as facts about, and experiences of, the self are
neither dismissed or ignored. In this, of course, courage will play its role. A third virtue
is more usually associated with character than the intellect—a form of love we may call
compassion or loving kindness.
10. The passions against which people defend can only secure a place in their con-
ception of their (imperfect) selves as they come to reject the sense that such pas-
sions make them unlovable. Love, therefore, plays a necessary role. If one’s passions
do not make one unlovable, they do not need to be hidden; honesty, at least with
oneself, becomes possible.6 This is one role that relationships with others can pro-
vide. A model of how such relationships work may be gleaned from psychoanalysis.
6
As defences have a developmental history, and were often initiated for good reason, a person’s sense that
some passion makes them unlovable may be realistic in the context of a specific relationship. For example,
Contemporary psychoanalysts emphasize the nature of the therapeutic relationship as

much as ‘insight’ in the account of how defences are deconstructed (see Lacewing 2014
for detailed discussion and empirical defence of the following sketch). The therapist
retains their interest, curiosity, and care, and remains non-moralizing and non-retalia-
tory in the face of the patient’s revelations and passions. This repeated disconfirmation
of the patient’s expectations challenges their sense of self and relationship, and enables
their defences to be withdrawn.
This process is, at the same time, supportive of the development of self-understanding.
One key expression of this supportive form of relationship is precisely the open-minded,
empathic, non-defensive exploration of the patient’s mind. This truthfulness is itself
part of the caring relationship that corrects the patient’s defensive expectations. Over
time, the patient internalizes both the truthfulness and the care shown by the therapist,
and thus the development of self-understanding can lie in relationship with another
(Eagle 2011: 285). Compassion for oneself—for one’s pain and vulnerability to it, for the
inevitable failure to be perfect—enables the pain against which defence mechanisms
defend to be tolerated without denying reality. Compassion complements courage in its
epistemic function: courage enables the pain of deconstructing defences to be endured,
compassion diminishes or contains the pain as it preserves a sense of self-worth in the
midst of emotions that challenge that sense.
11. It can be that the search for self-understanding itself contributes to self- transfor-
mation. Developing Cramer’s example of projected anger from §5: having projected his
anger onto Chrissy, Chris responds by withdrawing affection, citing ‘her’ anger as his
reason. If he recognizes that he is angry, but projected this onto Chrissy, he may under-
stand his withdrawal as a form of punishment motivated by his anger. If he does not
attempt to foreclose his self-recognition here, Chris may seek to understand his anger
and why he projected it, being open to the anxiety his anger brings him (at least in rela-
tion to Chrissy) and whatever imaginings, memories, or other emotions may follow.
Chris struggles here, so he asks someone with whom he can talk about such things, and
remains open and alert to what arises in him in response. A vague sense of discomfort
at being angry clarifies into a richer, more content-laden fear. As a result, Chris comes to
understand the situation in which he considered withdrawing affection quite differently,
such as as one in which he feared his need for love would not be met. In defence, he had
imagined the need for love was Chrissy’s, and punished her for that neediness by being
angry with her. But this was all still too anxiety-ridden, so he projected the anger onto
someone may defend against her anger, which is repressed or transformed into depression, as previous love
objects, such as parents, did indeed withdraw their love whenever she expressed it. But, assuming she is now
grown up, what she fears and expects is not just that getting angry will disrupt her current relationship with
her parents, but that anger is unacceptable. But the childhood situation does not generalize, and others will
not all act the same way; nor is it right to say, simply, that the fault lies in her anger per se. A process of further
self-development will involve discovering the reasonable place of anger in loving relationships.
her, which ‘justified’ his withdrawal of affection—neatly turning his need for love into a
fantastical act of self-sufficiency.
This kind of recognition can contribute to preventing the same defence from occur-
ring in the future; while it is not the whole story, self-understanding can contribute to
the deconstruction of defence.
12. The deconstruction of defence mechanisms is transformative. It is transforma-

tive of one’s passions, for they no longer threaten one’s sense of self-worth; of one’s
self, as the range of emotional experience in which one can be oneself, is expanded;
and of self-understanding, as the passions’ lack of transparency, at least as it derives
from defence, yields to a sense of their meaning and significance as they come to be
understood in consciousness. With this, the passions themselves may simply die away,
diminish in force, take new and more acceptable forms, or become amenable to rational
reflection; and the same is true of our ideals and judgements of appropriateness.
But should both the passions and the judgements of their inappropriateness remain
unchanged, as frequently happens, then having granted them a place in one’s conscious
psychology, one may at least become aware of such influence as they may have on one’s
sense of oneself and the situations one faces, and correct for it as best one can.
In either case, this transformation of the structure of the passions impacts directly
upon moral enquiry. With defensive distortions diminished, we may re-evaluate
what we have reason to do with greater perspicuity. And from the position of
greater self-understanding, we shall better contribute to and learn from communal
enquiry.
13. We began from the assumption that the passions are an important source of intui-
tions about reasons to act, feel, and desire in certain ways. They are active not only in
responses to situations, but also in processes of reflection and deliberation. But they can
be misleading, and in ways that operate outside unawareness; they lack transparency.
One cause of this is the occurrence of defence mechanisms creating unconscious dis-
tortions in the agent’s understanding of what he feels and why. Such distortions in turn
result in distortions in his understanding of the situations to which he responds and the
reasons which they furnish, as demonstrated in the examples of abortion, false altruism,
and lying, and the discussion of three intellectual vices from Roberts and Wood. Thus,
moral enquiry may be aided by the deconstruction of defence mechanisms. We identi-
fied the importance of close relationship and dialogue with others, together with specific
forms of courage, self-acceptance, and finally compassion, as productive in this regard.
The deconstruction of defence mechanisms and the greater self-understanding
that results is, of course, not sufficient for moral knowledge. But in its absence, I have
argued, we may fail to understand the passions on which our moral judgements rest.
The refinement of our moral sensibilities involves the deconstruction of our defences,
as a step towards gaining both a finer, more nuanced, and realistic grasp of the situa-
tions with which we are confronted, and refined and more appropriate responses to
them, no longer distorted in ways we fail to recognize by the influence of passions that
we cannot tolerate in ourselves.7
References
Baehr, J. (2011). The Inquiring Mind: On Intellectual Virtues and Virtue Epistemology
Chaiken, S. and Trope, Y. (1999). Dual-Process Theories in Social Psychology (New York: Guilford
Press).
Cottingham, J. (1998). Philosophy and the Good Life: Reason and the Passions in Greek, Cartesian
and Psychoanalytic Ethics (Cambridge: Cambridge University Press).
Cramer, P. (2006). Protecting the Self: Defense Mechanisms in Action (New York: Guilford Press).
Dijksterhuis, A. (2010). ‘Automaticity and the Unconscious’. In Handbook of Social Psychology,
ed. S. Fiske, D. Gilbert, and G. Lindzey (New York: McGraw-Hill), 228‒67.
Doris, J. M. and the Moral Psychology Research Group (eds.) (2010). The Moral Psychology
Handbook (Oxford: Oxford University Press).
Eagle, M. N. (2011). From Classical to Contemporary Psychoanalysis: A Critique and Integration
(New York: Taylor & Francis).
Freud, A. (1936). The Ego and the Mechanisms of Defence (London: Hogarth Press).
Freud, S. (1926). ‘Inhibitions, Symptoms and Anxiety’. In The Standard Edition of the Complete
Psychological Works of Sigmund Freud, vol. XX, ed. J. Strachey (London: Hogarth Press),
75‒176.
Goldie, P. (2008). ‘Misleading Emotions’. In Epistemology and Emotions, ed. G. Brun,
U. Doguoglu, and D. Kuenzle (Aldershot: Ashgate Publishing), 149‒66.
Haidt, J. (2001). ‘The Emotional Dog and Its Rational Tail: A Social Intuitionist Approach to
Moral Judgment’. Psychological Review 108: 814–34.
Haidt, J. (2007). ‘The New Synthesis in Moral Psychology’. Science 316: 998–1002.
Hursthouse, R. (1991). ‘Virtue Theory and Abortion’. Philosophy & Public Affairs 20: 223–46.
Kennett, J. and Fine, C. (2009). ‘Will the Real Moral Judgment Please Stand Up?’ Ethical Theory
and Moral Practice 12: 77–96.
Lacewing, M. (2005). ‘Emotional Self-Awareness and Ethical Deliberation’. Ratio 18: 65–81.
Lacewing, M. (2014). ‘Long-Term Psychodynamic Psychotherapy, Insight and Therapeutic
Action’. Clinical Psychology: Science and Practice 21: 156‒73.
Lear, J. (2003). Therapeutic Action: An Earnest Plea for Irony (London: Karnac).
Lemma, A. (2005). ‘The Many Faces of Lying’. International Journal of Psychoanalysis 86: 737–53.
Lenman, J. (2008). ‘Expressivism and Epistemology: What is Moral Inquiry?’ Proceedings of the
Aristotelian Society, Supplementary Volume,81: 63–81.
Locke, J. (1690). An Essay Concerning Human Understanding (London).
McDowell, J. (1985). ‘Values and Secondary Qualities’. In Morality and Objectivity, ed.
T. Honderich (London: Routledge), 110‒29.
7
Thanks to my research assistant Maarten Steenhagen for his invaluable feedback, and to Heythrop
College for providing the funding to enable this. Thanks also to the participants of the conference on ‘Moral
Emotions and Intuitions’ in The Hague in 2011, colleagues at Heythrop, and Sabine Roeser and Cain Todd for
providing many helpful comments on previous drafts.
Moran, R. (2001). Authority and Estrangement: An Essay on Self-Knowledge (Princeton: Princeton

University Press).
Nietzsche, F. W. (1998 [1886]). Beyond Good and Evil, ed. and trans. Marion Faber
Plantinga, A. (1993). Warrant: The Current Debate (Oxford: Oxford University Press).
Roberts, R. C. and Wood, W. J. (2007). Intellectual Virtues: An Essay in Regulative Epistemology
(Oxford: Clarendon Press).
Scanlon, T. M. (1998). What We Owe to Each Other (Cambridge, MA: Belknap Press).
Snow, N. E. (2010). Virtue as Social Intelligence: An Empirically Grounded Theory
(New York: Taylor & Francis).
Tiberius, V. (2008). The Reflective Life: Living Wisely With Our Limits (Oxford: Oxford University
Press).
Vaillant, G. E. (1993). The Wisdom of the Ego (Cambridge, MA: Harvard University Press).
Vaillant, G. E. (2000). ‘Defense Mechanisms’. In Encyclopedia of Psychology, vol. 2, ed. A. Kazdin
(Washington, DC: American Psychological Press), 454‒7.
Wiggins, D. (1987). ‘A Sensible Subjectivism’. In Needs, Values, Truth: Essays in the Philosophy of
Value (Oxford: Clarendon Press), 185‒214.
Williams, B. (2002). Truth and Truthfulness: An Essay in Genealogy (Princeton: Princeton
University Press).
Wilson, T. D. (2002). Strangers to Ourselves: Discovering the Adaptive Unconscious (Cambridge,
MA: Harvard University Press).
Wollheim, R. (2003). ‘On the Freudian Unconscious’. Proceedings and Addresses of the American
Philosophical Association 77: 23–35.
14
Emotion and Agency
Jan Slaby and Philipp Wüschner
William James famously held that we “feel sorry because we cry, angry because we
strike, afraid because we tremble, and not that we cry, strike, or tremble, because we are
sorry, angry, or fearful as the case may be” (1884: 190). We disagree, but not entirely.
We suggest it would be more adequate to say that we are sad insofar as we cry or
otherwise mourn, angry insofar as we strike or otherwise confront, fearful insofar
as we hide or run or otherwise evade, happy insofar as we laugh, jump for joy, or
smile. Emotions unfold in the act—they are not felt responses to prior and independ-
ent reactions or actions nor clearly separate causal entities or events prior to our
actions, but acted-out engagements with the world.1 The specific mode or style of
one’s engagement, the intensity, the emphasis or vivacity we bring to these acts shape
our emotions, determining their ‘what’ and their ‘how’. The active nature of emotion
importantly sheds light on the way emotion relates to value. Emotional engagement
is what lets value manifest and become concrete, in that it opens up a practical sphere
rife and buzzing with what ought to be (or not be), and thus what ought and can
be done—by me, by us, here and now. Value, on our view, is both constituted and
detected by our emotional engagements (cf. Helm 2001)—a view whose paradoxical
initial appearance will be mitigated by our account of emotions as active engage-
ments with the world.
Accordingly, our aim in what follows is to outline a philosophical view of emotion
that puts agency much closer to the heart of what an emotion is. Two interlocking
aspects are central to this perspective: (a) emotions are relational; that is, they are con-
stituted by a dense phenomenal coupling to the agent’s environment; (b) they are dia-
logical in that their acting-out as an engagement with the environment helps to shape
the space of possible further ways of acting them out and thus partly determines how
the emotion will subsequently unfold. Furthermore, conceptualizing emotions as
1
To our knowledge, Solomon (2004) has introduced the concept of “engagement with the world” into the
philosophy of emotion, as a way to make his former ‘judgementalist’ theory of emotion more adequate to the
phenomenology of emotional experience.
Emotion and Agency 213
unfolding in relational and dialogical acts demands seeing them less as mental states
and more as temporally extended episodes involving a person’s entire comportment
in and toward the world. Now, the problem of agency in part amounts to the question
in what ways we can appropriate these agentive episodes as ours. This appropriation is
crucial, because only then, by actively and consciously getting a hold of an emotional
episode that—at the same time—we passively undergo, is the realm of value disclosed
and (re)constituted in the same sequence of unfolding acts.2
Accordingly, what we will outline in the following could be put like this. The expres-
sive qualities of our environment, its manifest value (that is, things that matter to us in
some specific way or other) can draw us in and exert an affective pull on us. This might
be seen as the ‘passive aspect’ of emotions. However, in opposition to many theories of
emotion currently debated, we believe that an emotional episode does not stop there,
but essentially involves a moment of ‘phasing-over’ or transforming the initially pas-
sive experience into an active engagement. This acting-out is not a separate and blind
reaction to the affective pull but depends on the kind of person we are, on our various
abilities and capacities and, not least, on the values that we already uphold. In addition,
the way we act out our emotions not only conditions the further ways we can proceed
in acting them out, it also shapes the phenomenal aspects; that is, the felt dimension
of the emotion in question. Therefore, we argue, it is essential not to strip the phe-
nomenological description of emotions of their active momentum but, on the con-
trary, to emphasize it. This view avoids an apparent gap between emotions as passive
experiences and the actions they warrant. Rather than bridging the gap between pas-
sive emotions and the ensuing actions by some sort of obscure ‘motivational force’, we
suggest to think of emotions in a way that does not open this gap in the first place: the
action is already part of the emotion, it is no less than its processual core.3
In the following we will zoom in from the broad perspective of the socially engaged
person into what one might call her active-affective ‘minimal self ’, whose catastrophic
2
The view that emotional engagement ‘activates’ value is not new; it is a staple in phenomenological
thinking on emotion—articulated, for instance, by Sartre (1994 [1939]), with clear echoes from Heidegger
(1927), and also in slightly different terms by Paul Ricoeur (1966 [1950]). Solomon (1976), once again, tried to
appropriate the view for a contemporary Anglo-American philosophy, speaking of emotions as “constitutive
judgments” that imbue the world with value. With his sometimes exuberant formulations, Solomon tended
to oscillate between cognitivism (emotions as judgments detecting value) and projectivism (emotions as
bringing forth value). We believe that there is an important and correct intuition behind this very oscillation:
neither the tracking of pre-existing value nor its mere subjective projection captures what really goes on—
the point of speaking of an emotional constitution of value is exactly this: to avoid either extreme. Bennett
Helm (2001) comes quite close to a cogent explication of this no-priority view, albeit with somewhat exces-
sively rationalistic overtones. Part of the point of our chapter is to make a fresh start in this direction with a
focus on the aspect of agency in emotion.
3
Sabine Döring (2003, 2007) proposes an account of emotion that likewise closes the ‘motivational gap’
that presumably opens up, in the specific case discussed by her, between moral insight and moral agency
on standard philosophical accounts of moral motivation (what Michael Smith calls ‘the moral problem’, see
Smith 1994). However, Döring’s proposal differs in significant respects from ours as she deems emotions
not as directly agentive but rather as perceptual-cum-motivational states of a sui generis kind (‘affective
perceptions’).
214 Jan Slaby and Philipp Wüschner
erosion in conditions like severe depression lends additional phenomenological credi-

bility to our proposal. In this way we try to show that emotions are so intimately entan-
gled with action that a philosophical account of emotion is well advised to start—and
subsequently stay—right there: where the action is.4
1. From Experiential to Action-Oriented Theories

of Emotion
The discussion of the pros and cons of cognitive theories that has dominated the recent
decades in the philosophy of emotion seems by now rather tired. It is no longer a revo-
lutionary idea, as it might have been in the early time of cognitivism when the views
of Kenny, Bedford, Solomon, and others first emerged, that emotions have intentional
contents both capable and in need of rational evaluation. The motivating idea of cogni-
tivism no longer requires provocative statement, since in some broad sense, most or all
theorists currently working on the topic agree that emotions are rationally evaluable,
that they can be described as somehow ‘apprehending’ import. Accordingly, it makes
sense to assess their accuracy in doing so, for example, by distinguishing their fac-
tual contents from their evaluative contents, and likewise their world-relatedness from
their self-disclosive dimension (see Slaby and Stephan 2008).
Most theorists also agree that the cognitive or intentional characteristics do not
exhaust the emotions’ nature. It is their experiential dimension, their qualitative char-
acter, so the consensus goes, that distinguishes emotions from intentional states of
other kinds: emotions are felt, and essentially so. They are described variously as felt
evaluations (Helm 2001), feelings towards (Goldie 2000), affective perceptions (Döring
2007), affect-imbued concern-based construals (Roberts 2003), or felt evaluative atti-
tudes (Deonna and Teroni 2012). In viewing emotions as ways of affectively experienc-
ing the world, these approaches combine the insights of cognitivist and feeling theories
without sharing the excesses of either.
As we will show in the following, this common orientation towards experience as
the core dimension of emotion is not entirely misguided, but it might lead in the wrong
direction, particularly if ‘experience’ is understood in a narrowly perceptual and more-
over ‘passive’ sense. The danger in focusing on an emotion’s experiential rather than its
active character is that the importance of agency for emotion is lost sight of and with it
what might be an emotion’s most relevant feature. While paying regular lip service to
emotions’ motivational force, and to the fact that there are often characteristic actions
out of emotions (see, e.g., Goldie 2000: 37–49), proponents of philosophical theories
of emotion for the most part fail in providing a cogent connection between an emo-
tion’s feeling component and the action into which the emotion unfolds. Right from
4
Special thanks to Jonas Klein for suggesting several formulations for this introduction by providing a
highly illuminating written comment on an earlier draft of our chapter.
the start, emotions have to be understood in relation to our ability to act and to engage
with the world practically, or so we will argue in the following. Emotions themselves
are in fact best understood as forms of active comportment in and towards the world,
and in many cases their unfolding is not at all distinguishable from intentional action
as such. In these latter cases, it is easy to say that emotions are something we do—a view
that Robert Solomon famously over-generalized and read into Sartre (Solomon 1976).
This general view warrants a careful reappropriation that will also begin to clarify what
exactly our talk of an ‘agentive character of emotion’ implies—and what not.5
2. Phenomenal Coupling or: Who Took Passivity

Out of Passion?
Initially, it certainly seems that not all emotions are correctly described as active
engagements. For the most part, one may argue, the description of emotions as passive
experiences is much more adequate.6 How else could one conceptualize the feeling of
being overcome or swept away by an emotion, or capture emotions’ sometimes para-
lyzing effects? In order to show that even in cases of extreme passivity the temporal
unfolding of an emotion is agentive, it is advisable to start from a point least likely to be
associated with ‘activity’.
Consider cases of art appreciation. Many of the emotions we experience in response
to music, film, theatre, or dance are such that their full phenomenal quality, their
capacity to ‘move’ us, cannot be adequately characterized without recourse to the
expressivity of the artworks themselves—it is these expressive features that engage us
affectively: we are made to feel, acted upon from without, in and through our affectiv-
ity.7 What is certainly not needed, or so it might seem initially, is a reference to anything
we do. Here, as also in most cases of interpersonal emotions, what affects us in the envi-
ronment is itself active and expressive,8 whereas the ‘agent’ is in the grip of whatever is
going on in the world.
5
Somewhat contrary to its neglect in the philosophy of emotion, agency has figured more prominently in
the psychological research literature on emotion. Leading the way is Nico Frijda’s explicitly action-oriented
theory of emotion (see Frijda 1986); another approach that gives pride of place to action in emotion is by
Parkinson (1995), who views emotions as strategies of relationship configuration. Griffiths and Scarantino
(2009) provide a good overview over some of the empirical literature. Also see Lambie and Marcel (2002) for
a thorough account of emotional experience from a psychological perspective.
6
We omit a further complication of this matter by disregarding the plausible possibility that many expe-
riences might themselves be activities rather than processes passively undergone. With our proposal, we
chiefly aim to oppose passivist views of experience that treat experiences primarily as exercises of a person’s
receptivity (as opposed to a person’s spontaneity; see, e.g., McDowell 1994). With our stance, we are in line
with authors in the tradition of enactivism who defend a view of experience as essentially active (see, e.g.,
Noë 2005).
7
We use the term ‘affectivity’ in a generic sense to refer to affective phenomena; that is, emotions, moods,
and intentional feelings such as feelings of being satisfied or dissatisfied, and so on.
8
Art appreciation here serves only as a very intriguing example of the passive experience of emotion. We
cannot enter into the intricate debates about the expressivity of works of art or the actual activities it might
demand from the beholder. One plausible approach, put forth by Levinson in the context of an analysis of
But even here, we argue, the engagement with the expressive and thus active envi-
ronment comprises some sort of activity on part of the appreciator. Using a term that
might strike readers as somewhat unusual, we will call this ‘phenomenal coupling’ (see
Slaby 2014).
Phenomenal coupling is the direct engagement of an agent’s affectivity with an
environmental structure or process that itself has affect-like, expressive qualities—for
example, in the form of an affective atmosphere (Anderson 2009; Schmitz et al. 2011) or
as an expressive quality of a piece of music (Levinson 2009). The most relevant range of
examples for phenomenal coupling is in the social-interactive domain: probably noth-
ing is as emotionally engaging as the dynamic expressivity of fellow humans—indi-
viduals as well as groups regularly draw us into emotional experiences that we would
not be able to experience on our own. But even in these cases of apparently passive
immersion the core element of the emotional experience is a kind of engagement, a
form of agency.
Froese and Fuchs (2012), in part echoing Merleau-Ponty, have provided an account
of how this might be played out in interpersonal interaction: namely, in the form of
a dialogical interplay of actio and passio, expression and impression—with the lived
body as a “felt resonance-board for emotion” (Froese and Fuchs 2012: 212). In these
inter-affective exchanges, the manifested emotional expression (face, gesture, body
posture, and so on) of one agent is apprehended by the other in the form of an affec-
tive bodily comportment. This in turn modifies the second person’s expressivity, which
is again taken up by the other, and thus a dialogical sequence of mutual corporeal
engagement unfolds.9 Importantly, the expressive behavior of each party is not merely
a mirroring of the partner’s expressions; rather, it is dynamically thrusting forward,
thus enabling a genuine dialogue.10
Besides being felt, emotional episodes—even in their immediately felt, phenomenal
character—are forms of an intentional engagement with the world; they involve an
emotional comportment in the world.11 The feelings in questions are feelings-towards
in Peter Goldie’s sense (Goldie 2000) and thus forms of affective world-disclosure. The
music, holds that rhythmic movement and other dynamic features of musical sound constitute patterns
resembling the dynamics of human gesture (understood broadly). Musical expressivity would thus be deriv-
ative from the expressivity of human conduct (see Levinson 2009). As to the alleged ‘active’ expressivity
of other forms of art, such as paintings, we symapthize with enactivist accounts that stress the inherent
dynamic character of artful composition (see, e.g., Noë 2012).
9
Joel Krueger has provided in-depth descriptions of affect-rich embodied interaction, and he helpfully
invokes the concept of a ‘we-space’ as the specific interpersonal realm that is created and then negotiated in
these dialogical embodied exchanges (see Krueger 2011).
10
Enactivists speak of this dynamic interaction as “participatory sense-making”—a joint acting-out that
lets meaning become manifest in a way that is socially shared, situated, and embodied from the outset (see
de Jaeger and di Paolo 2007).
11
We prefer speaking of “comportment in . . .” instead of “comportment toward the world” in order to
avoid the impression of a cleavage between an agent’s engagements and the world that is thereby actively
disclosed. We follow Heidegger (being-in-the-world) and Merleau-Ponty (1962 [1945]) in subscribing to a
thoroughly anti-Cartesian conception of world-disclosure—one that does not start from a presupposed
ontological distinctness of ‘subjects’ and ‘objects’, but instead moves clear of these notorious dichotomies.
active lived body is the partially transparent medium of the emotional engagements, not
the opaque phenomenal endpoint of these experiences, as Jesse Prinz would have it (see
Prinz 2004; Slaby 2008b). Now, our view differs from Goldie’s in that we explicate what he
calls feeling towards as active engagements-with: where there is emotion, there is agency,
and thus an agent. But, at first glance, it is not clear that the action in fact originates, inten-
tionally, from the agent—since, as we have seen, emotions are typically experienced as
something that comes over us, at least initially. We will, however, subsequently take hold of
our emotional engagements, and sometimes we can even deliberately ‘stay in the mood’.
Either way, we actively hold ourselves in the comportment in question; it is now our emo-
tion in the same way that an action is our action. It is here that the descriptive shift from
mental states to temporal episodes is most noticeable. Imagine someone pushes you from
behind on the sidewalk in a rude way to pass by. What do you experience? First, you will
probably feel an unqualified surprise, finding yourself in a state of alarm, maybe even
shock. Your surprise will soon evolve into anger, calm down to annoyance, and might give
way to a general frustration about modern life in the city. But now, let us look at what you
do. Instinctively and almost mechanically your eyes will open wide, your heartbeat rises,
and so on. But soon you will change your facial expression to a frown, you will look for the
approval in the faces of others around you, if you find it, maybe you will exchange a smile
with that person, if not, you might shake your head and decide to move to the countryside.
In our view, there are two potential mistakes one might make in analyzing this
event. The first one would be to break up this emotional episode into different sets of
emotion-and-reaction patterns. The sequence only makes sense as a whole, because it
is unreasonable to believe that either your anger simply stops as soon as you see some-
one smile approvingly or that you now have two completely separate and independ-
ent emotions that do not interrelate with each other. The second mistake would be to
consider what you do simply as reactions to what you feel. The shocked expression,
your frowning, your looking for approval, your shaking your head . . . are the ways in
which your anger unfolds in time; and moreover they are ways in which you actively
integrate this episode into your life. They transform what happened to you into your
active engagement with the world. Some of these actions seem more deliberate than
others, some involve judgments, maybe even reasoning, some do not. So, transform-
ing the ‘passive’ experience into an active engagement might encompass anything
from impulsive physical responses to specific intentional actions until we reach a state
that we deem appropriate to the situation. This might raise the objection that it now
becomes unclear when an emotional episode actually ends. This might indeed be so,
but is it really an objection? Or is it rather a precise description of the fact that we are
constantly trying to obtain a hold over our affections and emotions?
This idea brings us very close to Aristotle’s concept of hexis as this possible ‘hold’
we can have over our emotions.12 Strictly speaking, the term hexis is not limited to
12
Hexis is a notion that has been unjustly neglected in comparison to Aristotle’s much better known
notion of the passions of the soul (pathe). The term comprises, like most Aristotelian terms, a variety of
the human sphere alone; however, its most influential role has been its use in the
Nicomachean Ethics to describe “those things in virtue of which we are in a good or bad
condition with respect to the feelings” (NE 1105b 25–6). It is this narrower meaning
with its strong connection to moral value that is of interest here. With hexis, Aristotle
provides a term lacking in most modern theories of the mind, as it conceptualizes this
precise moment of ‘phasing over’ or transition from pathe to praxis as a genuinely
human embodied capacity. It is crucial to understand that this ‘phasing over’ does not
involve moral reasoning or theoretical reflection. By our hexeis we are already well- or
ill-disposed to act in the light of strong emotions or passions. By virtue of our cour-
age, for example, we are well-disposed to strike when facing a threat; but this does not
mean we will not feel fear; it means that we will turn our fear around into brave action,
and in this action we can, so to speak, own our fear. This should not be misconstrued
as the commonplace ability to act in various ways out of our emotions. What Aristotle
aims for is the ability to mould the emotion itself in accordance with our capacities and
the situation at hand. This ‘moulding’ or shaping the emotion is carried out by acting
appropriately in line with who or what we are in relation to the established values of the
polis.
Hexis, deriving from the Greek verb for ‘having’ (echein) entails a concept of posses-
sion in contrast to being possessed—hence hexeis are more than just dispositions; they
are ways of self-ownership. For the Ancient Greeks with their concept of a perfect kos-
mos, being changed against one’s nature amounts to the ultimate moral failure, since
in doing so you lower yourself to the state of an animal or, even worse, a stone being
kicked around. Viewed from this Aristotelian perspective, it would be right to say that
conceptualizing an affective episode as completely passive will lead to a depersonalized
view of the subject involved.
Change, or movement, is the primordial condition of the world of the Ancient
Greeks, even prior to the distinction of actio and passio; and as a human you cannot but
constantly try to maintain ownership of that change at every level. You do so by praxis;
you execute your human nature only insofar you act. And therefore, it is in that praxis
that value is already contained, as its immanent telos.
Of course, there are aspects of the Aristotelian teleological concept that do not trans-
late as easily into our modern worldview, but the general idea points right to the centre
of an ongoing debate concerning agency and the problems it faces. David Velleman
famously presented the problem of agency as the problem of where to put the agent
in the “explanatory order of the world” (1992: 465). For Velleman, we understand
(human) action only if it can be traced back to an (human) agent as its cause, and
thus we are confronted with Thomas Nagel’s concern that “[e]‌verything I do is part
of something I don’t do, because I am a part of the world” (Velleman 1992: 467; Nagel
meanings reaching from disposition to capacity or even property and quality. Still the most thorough inter-
pretation of hexis can be found in Hutchinson (1986), given its dependency on Aristotle’s Physics VII, the
authenticity of which is not beyond doubt.
1986: 114). At this point, Velleman sets out to look for “events and states to play the role
of the agent” (1992: 475) and finds, naturally—desires, motives, reasons, and other men-
tal states (1992: 477). It is here that Velleman makes a critical difference between those
events that are “functionally identical to the agent” (1992: 475) and those that are not,
coming to the conclusion that personhood is to be found only in the “desire to act in
accordance with reasons” (1992: 478)—making in fact a very Aristotelian point.
We cannot fully enter this discussion about how to appropriate causes as ours in
the right way and whether reason should really be the criterion here, nor can we solve
Nagel’s problem of agency. However, we cautiously suggest two things: first, this appro-
priation is already imposed on us by what we usually call our passions (that is, what
matters to us, what concerns us as individuals) and not only by those desires that are in
accordance with reason or some sort of human nature; and second, this appropriation
itself is already essentially agentive. It is something we do, and we do it so continuously
and at the most basic level that in order to get a true picture of human nature, our
agency is the starting point that explains the rest and not some obscure problem we are
left with at the end of the day. So Nagel’s worry “[e]‌verything I do is part of something
I don’t do, because I am a part of the world,” is unfounded, since “being part of the
world” is essentially something we do.
It should be clear that we by no means want to take passivity out of emotions, nor
do we claim that experience is not a crucial factor in what we call affectivity. In fact, in
our view, it is highly doubtful whether ‘passivity’ and ‘activity’, ‘experience’ and ‘action’
need at all be conceptualized as mutually exclusive. Our concern here is only with the
right order of explication. For once one starts viewing emotions through the lens of
passive experience one will have a hard time bringing activity back into the picture in
a satisfying way.
3. Sartre’s Hodological Space and the Sense of

Ability
Emotions are widely accepted as motivational forces (that might lead to action) and
as ‘ways’ of doing things—like lovingly gazing at someone you like or angrily shout-
ing at someone you do not like. Yet this general acceptance is either overturned by the
neglect that agency suffers in philosophical theorizing about emotions, or confused
by overly complex theories, for example, by those in which desires are understood
as prime sources of human action.13 This neglect (or confusion) leads to an uncom-
fortable explanatory gap in the description of emotional episodes. Even in the most
13
A recent partial exception to the neglect of agency in the philosophy of emotion is the account by
Deonna and Teroni (2012: see esp. ch. 7). Our proposal can be seen as more radical than theirs in chiefly two
respects: first, what Deonna and Teroni call ‘attitudes’ is on our account specified as more active engagement
or comportment, and second, we opt for a more intimate entanglement and co-articulation of agent and
world that is more in line with the existentialist branch of the phenomenological tradition.
careful approaches, emotions tend either to be turned into distinct causal entities that
somehow lead to an ensuing action, or into detached felt feedbacks that show up when
the real action is already over. We have indicated that even in cases of affectivity that
seemed to involve a passive recipient rather than an agent, this involvement itself is
best described as an active engagement with the world, first in the form of an invol-
untary comportment, and then, at least in many cases, as an active appropriation and
continuation of the initial engagement. Now it is time to show how the true agentive
center of affectivity lies exactly at this point: at the junction where someone’s passivity
and activity oddly melt into her being actively involved with the world.14
It might not be difficult to agree that affectivity is something like an embodied and
enactive ‘interface’ between experiential awareness and intentional action. Emotional
episodes mark a kind of switch point or transitional zone at which an evaluative aware-
ness of a situation (something relevant grabbing attention) is phasing over into active
engagement (our doing something about, with, or in relation to it). But we want to
suggest a more direct involvement of agency in emotion: affective responses consist in
more than merely a felt pressure or pull to act in relation to what is grasped as important
in the current situation. In our view, action and engagement themselves make up the
substance of an emotional episode. A situation-directed motivational pull marks the
onset, and often it is then actively taken up and followed through with in the course of
the emotion (anger is paradigmatic here, but also joy, also grief, also fear). Something
in the world demands to be acted upon, or demands us to specifically avoid or evade
it, and this attention-grabbing onset directly leads to a form of engagement that is the
emotion—we are angry insofar as we strike, ashamed insofar as we avert the gaze of
others, afraid insofar as we hide or run, happy insofar as we rejoice. In emotional expe-
rience, there is for the most part no salient difference between the apprehension of the
importance of something and one’s being pulled into engaging the world in accord-
ance with this situational significance. There is no gap: to be emotional is to be engag-
ing the world those relevant ways (see Helm, 2001, 2002 for a related view).
To understand how this commencing engagement shapes the experiential character
of an emotion, imagine the following scenario. You watch a philosophical debate at
home with some friends. One of the discussants makes an appalling comment about
something of importance to you (for example, the role of agency in emotional expe-
rience). At home you will heatedly express your anger, pointing at the screen and
maybe uttering one or two insults. Now, imagine yourself sitting in the audience of
this same talk. Here you cannot point, scream, or insult—instead you might fiercely
roll your eyes, shake your head, fold your arms, or sigh in disbelief. Next, imagine you
are on stage in the discussion facing that imbecile yourself. Now you may lean forward,
14
This process of transformation from passive to active is also such that an emotional episode’s inten-
tionality shifts from merely experiential intentionality (something is revealed as being such and such) into
goal-oriented or teleological intentionality (some effect is to be realized, as for instance in flight, fight, or
communication and the like).
and stare your opponent down while simultaneously controlling your anger so that
it will still support your sharp reply but without making you look like a fool. In all
three cases, some sort of anger was induced that not only revealed the way in which
the situation demanded to be engaged, but also consisted in (at least) the beginning of
the execution of such an engagement with the situation. However, some possibilities
of action were blocked in each situation, whereas others were highlighted and almost
obligatory. Accordingly, depending on the concrete circumstances, your anger was
acted out differently in each case.
It is this phenomenon that Sartre in his Sketch of a Theory of Emotions—a particu-
larly explicit and helpful approach to action in emotion—understood as a hodologi-
cal space.15 The different possibilities of acting on your anger are also reflected in your
emotional experience. Your anger was mixed with joy insofar as you could freely insult
your opponent while safely at home among your reassuring, possibly laughing, friends.
Your anger was worsened and stained with helplessness when you sat in the audience,
unable to express it appropriately. And your anger was bordering on excitement and
aggression when you were on stage under the spotlight, ready to strike with a brilliant
counter-argument. In each scenario, the specific disvalue of the discussant’s comments
was clearly manifested, revealed in light of your emotional evaluation—brought to its
full fervor through your expressive acting-out in the first case; appearing in a fleeting,
unbearable obtrusiveness, reflecting your nervous paralysis in the second case; while
in the third case, the disvalue is already about to give way to a brighter evaluative out-
look, precisely because you are now able to engage the situation effectively, in order to
make it right by critiquing and potentially correcting what had just been said.16
In a related but slightly different vein, Sartre construes emotions as ‘magical transfor-
mations of the world’, often in response to obstacles, problems, or hindrances encoun-
tered by a person in the course of her activities. Sartre is here clearly influenced by the
Gestalt psychologists since, like them, he understands the world as a network of trails
and paths presenting opportunities and obstacles. At each moment and each place
the world displays an ‘index of adversity’, that is, the specific degree of practical dif-
ficulty and thus nuisance that the world currently presents to our attempts to navigate
it. To Sartre, emotions are situated embodied engagements, sometimes employed stra-
tegically (not always consciously), acted out in response to unwanted or unexpected
disruptions of activities, in the face of the specific obstacles or opportunities that the
world offers. An emotion is thus a bit like a play; the emoter is performing a drama
15
The notion ‘hodological space’ was introduced into psychology by the Gestaltist Kurt Lewin (1938); for
a contemporary adoption of the notion within a psychological theory of emotion, see Lambie (2009).
16
This again illustrates the complexity of emotional value disclosure. While anchored in objective fea-
tures of the world, value also crucially depends on characteristics of the person(s) apprehending it—notably
in those that bear on one’s capacities to actively engage the relevant situation in meaningful ways. Thus, a
subjective, capacity-based element is introduced into value-constitution, putting limits on the generaliz-
ability of value-ascriptions. We cannot go into this here, but see Slaby (2008a) and of course Helm (2001) for
more encompassing treatments.
geared to an adverse environment, so that, in line with this acted-out performance, cer-
tain further activities or actively adopted or maintained stances become appropriate.
In an important sense the world is thereby reconstituted in the act, so as to be ‘inhabit-
able’ again—new domains of value are thus opened up. Sartrean emotions are acted-
out plots of social action and interaction, and not passive experiential occurrences.17
Parts of these plots certainly contain moments of passivity—like sitting helplessly in
the audience, frozen in one’s seat, watching without any chance to act as one would like
to. But even these passive episodes are acted out in some sense. You still do something
when you shake your head, fold your arms, or roll your eyes, or even if your whole body
tightens in self-restraint. And surely you thereby engage value, concretize it, actively
dedicate yourself to it, and project potential future actions in accord with it (see also
Ricoeur 1966 [1950]: esp. 72–7). With emphasis on cases like these, Sartre’s account of
emotion begins to seem much less counter-intuitive from the point of view of the stand-
ard understanding of emotions as passions.18
One important issue that comes into focus on the action-centered view is the extent
to which emotions are closely linked to an agent’s sense of ability. Emotional engage-
ments are in part rooted in our sense of what we can do, what we are capable of, and also
what we can cope with or what we can ‘take’ more generally. Emotional engagements
take shape within a dynamic ‘I can’ or ‘I cannot’—schema of relating to the world—a
corporeal sense of ability or its opposite, a corporeal sense of inability or incapacity in
relation to what confronts one that marks the heart of action and emotion. Depending
on this changeable background of capability (unthematically disclosed in experi-
ence and agency), different types of emotions unfold according to these changing
circumstances.19
How I affectively engage a situation I find myself in partly results from my sense of
ability brought to bear on the accessible relevant features of the situation. In this way, it
is plausible to understand emotions as embodying specific manifestations of an active
sense of possibility: emotional engagements actively disclose what a situation affords in
terms of potential doings and potential happenings. Emotional engagement is often
even a matter of being forced into a dynamic space of possibilities. The agent is driven
into acting in accordance with a specific trajectory of opportunities and obstacles, in a
certain style or mode of engagement that can be quite hard to snap out of. Again, and in
17
Robert Solomon has followed closely in Sartre’s footsteps, insofar as he spelled out and defended a
theory of emotions as essentially active, deliberate strategies or strategic choices (see Solomon 1976). More
recently, Paul Griffiths has advocated an approach of emotional content as “action-oriented representations”
and of emotions in general as ‘Machiavellian’; that is, evolved strategic responses to socially significant situ-
ations. As an evolutionary naturalist, Griffiths works within a rather different theoretical framework than
Solomon (see Griffiths 2004; Griffiths and Scarantino 2009).
18
Obviously, as is well known, Sartre is not willing to let the emoter off the hook by excusing him for alleg-
edly ‘involuntary’ emotional reactions. Sartre construes even the passive emotions as exercises of a person’s
freedom (see, e.g., Sartre 1994 [1939]: 44).
19
These and related ideas are expressed in a slightly more detailed manner in Slaby (2012) and Slaby et al.
(2013).
light of this, it still makes sense, even on the activist view developed here, to view emo-
tions as ‘passions’—as something that the agent might in some part passively undergo,
and not actively initiate herself. But one has to note that many emotions indeed become
thoroughly active after their comparatively passive onset, often even in the full-blown
sense of intentional action deliberately executed to fulfill specific aims (see Griffiths
and Scarantino 2009 for a related view).
In all cases of emotion, there are two distinguishable aspects—one situational,
the other one agentive—which come together to form a unified mode of relational
and dialogical engagement with the world. Together they make up a performative or
enacted affective awareness of situation, where ‘awareness’ must not be construed as
narrowly mental and passively experiential but instead as something inextricable from
our practical dealings with and comportments in the world.
With this, the corporeal nature of emotional engagement comes into sharper
focus. The central ‘vehicle’ or performing medium of emotion is the acting and acted-
upon lived body (Merleau-Ponty 1962 [1945]: see esp. Part I, §19). The lived body is
the dynamic framework of a person’s active, corporeal situatedness in the world.20
The environment, the hodological surround, becomes manifest, concretizes in and
through one’s corporeal sense of ability, in and through one’s felt capacity to act or
to cope with what affects one. It is here where self- and world-disclosive aspects of
emotional engagements come in view as always already unitary, because my sense of
agency and capability is from the outset a sense of both myself and the world: as the
space of my possible acting or being acted upon—and these aspects are inseparable.
4. The Active-Affective Minimal-Self

Let us narrow our focus somewhat. So far we have seen how agency in one impor-
tant way defines affectivity even when it is crucially shaped by the environment (2).
We then tried to explain the engagement of the agent with this environment in terms
of inter-action and by that shed light on a more basic understanding of what agency
might entail (3). With this, our account also begins to explicitly link affectivity, agency,
and the self. What is the relationship between affectivity, understood actively, and that
which might lay claim to be called ‘the self ’ (in its most basic form)? This is what we
turn to in the present section. With some due caution, one might speak of the self-
disclosive aspect of emotional engagement as a kind of affective self-construal. As out-
lined previously (at the end of section 3), this is not a separate experiential structure,
but comes in view primarily as a modifier of the process of active engagement, as an
accompanying sense of ability and possibility, and is thus inseparable from the actions
and activities of the emoting agent. An affective self-construal does not comprise a
20
For more detailed phenomenological explorations of the corporeal nature of emotional engagement,
see Colombetti and Thompson (2008); Ratcliffe (2008a); Schmitz et al. (2011); and Slaby (2008b).
separate structure of self-directed mental contents. It is caught up in the act—it, too, is

enacted (see Slaby 2012).
Affective self-construal constitutes a kind of ‘minimal self ’. It is crucial to see that
this basic structure is a matter of agency from the outset. Not through reflection, but
in an immediately affective way, our being is disclosed to us in relation to what we
are currently concerned with. This affective-agentive sense of possibility comprises a
sense of facticity—what the current situation manifestly presents; and second, a sense
of ‘what is next’, ‘what needs to be done’, including, most importantly: ‘can it be done?’;
that is, specific contentful ways of projecting ahead of what is currently manifest.21
Touch is a good case to illustrate the interrelatedness of activity, self-disclosure,
and world-disclosure on the most basic level. In touch, I obviously have a correlation
between feeling myself and, in the very same experience, a feeling of something in the
world as that which is touched. But quite importantly, only the fact that I am active,
that I do the touching, lets both self and world take shape in contrastive correlation
(see Ratcliffe 2008a: ch. 3 and 2008b). This point generalizes: without the moment of
activity, no experience could take place, because no basis for a self/non-self distinction
would be provided for. The ‘I do’ is as deep as it gets in the constitution of experience.
This moment of activity is not an abstract principle (like the Kantian ‘I think’), but
a modifiable, qualitative dimension: it can be emphatic or tentative, strong or weak,
resolute or shaky. This partly determines the particular kinds of the encounters with
the world that the agent might enter into, so different styles of engagement reflect back
on the agent and let him or herself take a particular shape, and this is what is ‘disclosed’
in affective self-disclosure or ‘self-feeling’.22
By now it should seem plausible that a basic agentive affectivity is among the funda-
mental sources of what constitutes a self-conscious subject; more specifically, it is also
what enables a person to assume the specific characteristics that are definitive of her
as an individual—it is thus a fundamental source of the ‘self ’ (if we can speak in that
objectifying way). Affectivity constitutes the very dimension in which things can pos-
sibly concern us or be an issue for us. Modify or take away a person’s sense of possibil-
ity, and there is not much left in the dimension of selfhood—no agency, no valuing, no
motivation, just a colorless plain condition.
This is a condition that in some (hard) cases might be approached by the unfortu-
nate sufferers of severe depression. Depression, in its extremes, seems to wipe out the
21
This is deliberately reminiscent of Heidegger’s famous account of the care-structure of the being of
dasein: thrown projection, discursively articulated—or, in Heidegger’s more vivid explication: dasein’s essen-
tial ‘being-ahead-of-itself’ in ‘being-already-in’ as ‘being-alongside’ that serves to also articulate the primor-
dial temporality of the being of dasein (see Heidegger 1962 [1927]: esp. §§41 and 64; and Haugeland 2013 for
illumination). Besides much else, what we especially take from Heidegger is the conviction that no mean-
ingful distinctions can be drawn between selfhood and a basic form of self-consciousness (Heidegger, of
course, would not use the latter term), and neither between selfhood and dasein’s care-structure as such, so
that selfhood is an always already affective thrust toward significance and value.
22
On the outdated phenomenological notion of ‘self feeling’, see Frank (2002); see also Slaby (2012) for an
adoption of that notion to the current discussion of emotion.
self via an annihilation of agency and affectivity. Depression seems to directly affect the
deepest point at which talk of ‘selfhood’ is appropriate, if one lends credibility to patient
self-reports (see Slaby et al. 2013; Ratcliffe 2012). In profound melancholic depression,
the dimension of affective self construal—the dynamic, agentive core of a person’s
perspective on the world—seems to be eroded, and ultimately eradicated entirely.
Goal-directed activity, active self-stabilizing, adopting or continuing a stance, resist-
ing impulses or oppositions—all that is made harder and harder before it might even-
tually become extinguished. A person’s active world relatedness in the mode of I can
deteriorates into an all encompassing I cannot, so that even the most routine of activi-
ties become practically impossible. As a consequence, the environment is increasingly
apprehended as oppressive, threatening, not manageable, while one’s existence amidst
others and amidst the entities of the world is disclosed as fragile, endangered, vulner-
able, and at the mercy of alien forces. In this way, depression seems to present a kind of
negative mirror image of an undisturbed affectivity (again, see Slaby et al. 2013). This is
the utter horrifying strangeness of severe depression. Thus afflicted, you cannot simply
point somewhere and say “this is what I have, this is my ailment . . .”—but it is the very
dimension that is you, your agency, the very instance that might lay claim on being
called ‘self ’ that is affected.
A direct consequence of this is that the universe of value collapses. Without the
active capacity, the engagement of the self that is wiped out in depression, there is no
basis to the ‘holding’ oneself in a realm of significance, in the thrust and dedication
that is needed to let purpose and meaning become manifest. The result is, as so often
observed in depressed patients, endless spirals of futile reflection and the anxiety char-
acteristic of the uprooted, degraded self. An agent’s engagement loses its traction on an
inhabitable, meaningful environment. Conversely, the world loses its practical shape,
its character as something that grips and enthralls, its status as a realm buzzing with
significance so that our action and engagement is evidently called for. Instead, to the
depressive, nothing is of value—nothing matters any more. In our view, this profound
desertedness, the nihilism of the depressive’s world, is the phenomenal flipside of that
deep pre-personal inability and incapacity to act and to ‘hold oneself ’ that comes with
the catastrophic erosion of the affective minimal self.
5. Conclusion
We have argued for a shift of emphasis in the philosophical study of emotion—a shift
from viewing emotions exclusively as a certain class of (mostly passive) experiences
to viewing them as much closer to and more intimately tied up with our agency: as
engagements with the world. If the emotions are ‘passions’, then they are the pas-
sive modifications of our active nature—the various ways our activities are shaped
by dynamic goings-on in our surroundings. And we have pointed out that even the
basic dimension of selfhood—basic affective self-awareness or self-feeling—has to
be understood as an essentially active structure, so that we can conclude that making

sense of our affectivity is making sense of ourselves as agents.23
References
Anderson, B. (2009). “Affective Atmospheres.” Emotions, Space and Society 2: 77–81.
Aristotle (2002). Nicomachean Ethics, trans. S. Broadie and C. Rowe (Oxford: Oxford University
Press).
Colombetti, G. and Thompson, E. (2008). “The Feeling Body: Toward an Enactive Approach to
Emotion.” In Body in Mind, Mind in Body: Developmental Perspectives on Embodiment and
Consciousness, ed. W. F. Overton, U. Mueller, and J. Newman (New York: Lawrence Erlbaum),
45‒68.
de Jaegher, H. and di Paolo, E. (2007). “Participatory Sense-Making: An Enactive Approach to
Social Cognition.” Phenomenology and the Cognitive Sciences 6: 485–507.
Deonna, J. and Teroni, F. (2012). The Emotions: A Philosophical Introduction (New York:
Routledge).
Döring, S. (2003). “Explaining Action by Emotion.” Philosophical Quarterly 211: 14–30.
Döring, S. (2007). “Seeing What To Do: Affective Perception and Rational Motivation.”
Dialectica 61: 363–94.
Frank, M. (2002). Selbstgefühl [Self-Feeling] (Frankfurt: Suhrkamp).
Frijda, N. H. (1986). The Emotions (Cambridge: Cambridge University Press).
Froese, T. and Fuchs, T. (2012). “The Extended Body: A Case Study in the Phenomenology of
Social Interaction.” Phenomenology and the Cognitive Sciences 11: 205–35.
Goldie, P. (2000). The Emotions: A Philosophical Exploration (Oxford: Clarendon Press).
Griffiths, P. E. (2004). “Towards a ‘Machiavellian’ Theory of Emotional Appraisal.” In Emotion,
Evolution and Rationality, ed. P. Cruse and D. Evans (Oxford: Oxford University Press),
89‒105.
Griffiths, P. E. and Scarantino, A. (2009). “Emotions in the Wild: The Situated Perspective on
Emotion.” In The Cambridge Handbook of Situated Cognition, ed. P. Robbins and M. Aydede
(Cambridge: Cambridge University Press), 437‒53.
Haugeland, J. (2013). Dasein Disclosed: John Haugeland’s Heidegger, ed. J. Rouse (Cambridge,
MA: Harvard University Press).
Heidegger, M. (1927). Sein und Zeit (Tübingen: Niemeyer Verlag). Trans. J. Macquarrie and
E. Robinson, 1962, as Being and Time (Oxford: Blackwell).
Helm, B. W. (2001). Emotional Reason: Deliberation, Motivation, and the Nature of Value
Helm, B. W. (2002). “Felt Evaluations. A Theory of Pleasures and Pains.” American Philosophical
Hutchinson, D. S. (1986). The Virtues of Aristotle (London: Routledge & Kegan Paul).
James, W. (1884). “What is an Emotion?” Mind 9: 188–205.
23
We are grateful to the editors of this volume and, moreover, to Jonas Klein, Jean Moritz Müller, and
Nina Trcka for their detailed and highly helpful comments on earlier drafts of this chapter.
Krueger, J. (2011). “Extended Cognition and the Space of Social Interaction.” Consciousness and
Cognition 20: 643–57.
Lambie, J. A. (2009). “Emotional Experience, Rational Action, and Self-Knowledge.” Emotion
Review 1: 272–80.
Lambie, J. A. and Marcel, A. J. (2002). “Consciousness and the Varieties of Emotion
Experience: A Theoretical Framework.” Psychological Review 109: 219–59.
Levinson, J. (2009). “The Aesthetic Appreciation of Music.” British Journal of Aesthetics
49: 415–25.
Lewin, K. (1938). “The Conceptual Representation and the Measurement of Psychological
Forces.” Contributions to Psychological Theory (Vol. 1, No. 4) (Durham, NC: Duke University
Press).
McDowell, J. (1994). Mind and World (Cambridge, MA: Harvard University Press).
Merleau-Ponty, M. (1962 [1945]). Phenomenology of Perception (London: Routledge & Kegan
Paul).
Nagel, T. (1986). The View from Nowhere (Oxford and New York: Oxford University Press).
Noë, A. (2005). Action in Perception (Cambridge, MA: MIT Press).
Noë, A. (2012). Varities of Presence (Cambridge, MA: Harvard University Press).
Parkinson, B. (1995). Ideas and Realities of Emotion (London and New York: Routledge).
Prinz, J. J. (2004). Gut Reactions: A Perceptual Theory of Emotion (New York: Oxford University
Press).
Ratcliffe, M. J. (2008a). Feelings of Being: Phenomenology, Psychiatry, and the Sense of Reality
Ratcliffe, M. J. (2008b). “Touch and Situatedness.” International Journal of Philosophical Studies
16: 299–322.
Ratcliffe, M. J. (2012). “Varieties of Temporal Experience in Depression.” Journal of Medicine and
Philosophy 37: 114–38.
Ricoeur, P. (1966 [1950]). Freedom and Nature: The Voluntary and the Involuntary, trans. E. V.
Kohák (Evanston: Northwestern University Press).
Roberts, R. C. (2003). Emotions: An Essay in Aid of Moral Psychology (Cambridge: Cambridge
University Press).
Sartre, J.-P. (1994 [1939]). Sketch for a Theory of the Emotions (London: Routledge).
Schmitz, H., Müllan, R. O., and Slaby, J. (2011). “Emotions Outside the Box: The New
Phenomenology of Feelings and Corporeality.” Phenomenology and the Cognitive Sciences
10: 241–59.
Slaby, J. (2008a). Gefühl und Weltbezug. Die menschliche Affektivität im Kontext einer
neo-existentialistischen Konzeption von Personalität (Paderborn: Mentis).
Slaby, J. (2008b). “Affective Intentionality and the Feeling Body.” Phenomenology and the
Cognitive Sciences 7: 429–44.
Slaby, J. (2012). “Affective Self-Construal and the Sense of Ability.” Emotion Review 4: 151–6.
Slaby, J. (2014). “Emotions and the Extended Mind.” In Collective Emotions, ed. M. Salmela and
C. von Scheve (Oxford: Oxford University Press).
Slaby, J., Paskaleva, A., and Stephan A. (2013). “Enactive Emotion and Impaired Agency in
Depression.” Journal of Consciousness Studies 20: 33‒55.
Slaby, J. and Stephan, A. (2008). “Affective Intentionality and Self-Consciousness.” Consciousness
and Cognition 17: 506–13.
Smith, M. (1994). The Moral Problem (Oxford: Blackwell).

Solomon, R. C. (1976). The Passions: Emotions and the Meaning of Life (New York: Doubleday).
Solomon, R. C. (2004). “Emotions, Thoughts, and Feelings: Emotions as Engagements with the
World.” In Thinking about Feeling: Contemporary Philosophers on Emotion, ed. R. C. Solomon
(Oxford and New York: Oxford University Press), 76‒88.
Velleman, D. J. (1992). “What Happens When Someone Acts?” Mind 101: 461–81.
15
Evaluating Existential Despair
Matthew Ratcliffe
1. Introduction
In this chapter I begin by describing an emotional state that is sometimes referred to as
‘existential despair’, and then address the question of whether or not it involves an accurate
appraisal of the human predicament. I take my lead from a posthumously published paper
by Peter Goldie, which explores the relationship between emotional feeling and evaluative
belief by reflecting on those unsettling occasions when our intellectual life ‘goes cold on
us’. He describes the experience as follows:
We know that this new book or paper on just the topic that we are ourselves researching will be
interesting and challenging, and we know that we ought to read it. But we have no curiosity, no
wonder at what we read, no feeling of hope that we will find some answers, and no courage to
keep on reading. We no longer feel the keenness to write new material, to face up to challenges
to our views, merely going through the motions, driven by mere habit or by the requirements of
one’s job, churning out more and more variations on the same old stuff. Our intellectual life has
gone cold on us. (Goldie 2012: 124)
In these circumstances there is a deadening of what Goldie calls “intellectual emotions,”

which he defines as emotions that play a role in intellectual activity, rather than types
of emotion that are exclusive to it. Examples include “delight, wonder, awe, fascination,
courage, surprise, worry, doubt, curiosity, concern, tenacity, and hope” (2012: 122). Such
emotions play epistemic roles in regulating inquiry and nudging us towards certain con-
clusions, as well as motivating us and partly constituting the sense that our activities are
worthwhile.1 Adopting a distinction made by D’Arms and Jacobson (2010) between loss
of emotional dispositions and blockage of their effects due to obscuring factors, Goldie
considers the difficult and important question of how, when our intellectual life goes cold,
we might distinguish blockage by a transient mood from loss, and thus determine the
1
See also Hookway (2002), Thagard (2002), and Morton (2010) for good discussions of the various roles
that emotions play in inquiry.
230 Matthew Ratcliffe
appropriate course of action.2 Like D’Arms and Jacobson, he maintains that introspection
is not a reliable guide here. So, he observes, we may have to embark upon an uncertain,
painful, and dangerous but potentially constructive journey, the result of which is some-
times but not always “intellectual conversion.”
The distinction between blockage and loss of emotional dispositions and associated
values is not a tidy one. It is not specific to intellectual matters, and applies equally
to the influence of emotional feelings upon other life projects, commitments, and
activities. For example, in evaluating a difficult marriage, a person might be unsure
whether the love she felt for her spouse is temporarily eclipsed by a disruptive mood
or altogether gone. It is also unlikely that loss of intellectual vitality is ever exclusively
‘intellectual’, as an inability to engage in intellectual projects that were central to one’s
life will impact upon other aspects of life and vice versa. Furthermore, many differ-
ent kinds of state may be able to block emotional dispositions, perhaps not just other
emotional states. Some blockages will no doubt be longer-term or more recalcitrant to
change than others, and some may be associated with partial loss, where weakening of
a disposition then renders it more vulnerable to interference. There is also the issue of
what exactly is lost in any given case, given that projects are sometimes revised rather
than altogether discarded. Even projects that are abandoned for many years are some-
times rekindled, making it unclear whether and when the relevant dispositions should
be declared ‘gone’.
Despite such concerns, a rough distinction between loss of affective dispositions
and blockage of their effects is, I think, informative, where an ‘affective disposition’
is understood as a disposition to experience feeling x in context y, my focus being
on those cases where the presence or absence of x has some effect upon our values.
Whole systems of projects, values, and commitments can be held hostage to unruly
emotional states such as moods, the obscuring influence of which is not always trans-
parent to first-person reflection.3 One is not always aware that one’s evaluations are
under the influence of mood p and, even when one is aware, one may be unable to
distinguish mood p’s interfering with the effects of affective disposition q from one’s
having altogether lost q. We might go so far as to say that certain moods and feelings
present us with evaluative ‘truths’ about our lives, in ways that are sometimes reliable
and sometimes not. That position suggests a conception of emotion and, more spe-
cifically, ‘feeling’ according to which feeling is inseparable from the attitude and con-
tent of belief, at least when it comes to systems of evaluative beliefs.4 Indeed, Goldie
2
As D’Arms and Jacobson (2010: 598) put it, there are “various obscuring factors, which generate or sup-
press emotional responses in ways that do not reveal the agent’s underlying sensibility.”
3
See Ratcliffe (2010) for a detailed discussion of how certain moods have a profound effect upon one’s
experience and yet remain phenomenologically ‘inconspicuous’.
4
For current purposes, I regard ‘the emotions’ as a complicated and possibly heterogeneous collection of
phenomena, and ‘feeling’ as an important constituent of emotion. I focus here upon ‘feeling’ rather than on
other alleged aspects of emotion and mood, which include ‘belief ’, ‘evaluation’, and ‘judgment’. I thus avoid
slipping into trivial claims to the effect that a mental state that incorporates evaluative belief contributes to
evaluative belief. Where I refer to moods, my emphasis is upon the ‘felt’ character of mood (Ratcliffe 2010).
Evaluating Existential Despair 231
(e.g. 2000, 2009, 2012) maintains that many of the feelings implicated in emotional
experience are not merely “bodily feelings.” They also have world-directed intention-
ality; they are “feelings towards,” through which we encounter aspects of our worldly
situation. These feelings are not mere “add-ons” to our thoughts. Instead, they partly
comprise our evaluative beliefs (and perhaps other kinds of belief too). An “evaluative
outlook,” according to Goldie (2012: 124), is a “structure of caring” that incorporates
dispositions towards “affective” responses. I am sympathetic to that view, and I find it
especially compelling when interpreting the subject matter of this chapter: ‘existential
despair’—something that is frequently described in ways that suggest the insepara-
bility of feeling, content, and conviction.5 However, even those who deny that feeling
plays a constitutive role in evaluation can accept much of what I will say here, so long as
they at least concede that some emotional states causally influence some evaluations,
to the extent that the person would not form a given evaluation if affective disposition
x were obfuscated or lost.
The topic of existential despair is hardly a novel one. However, I will adopt a dis-
tinctive and hopefully illuminating approach towards it, by applying the distinction
between blockage and loss of affective dispositions. I focus upon an especially evoca-
tive description of existential despair offered by Tolstoy in his memoir A Confession.
I begin by offering some account of its content, but add that this content may not be
fully graspable in isolation from an associated mood or feeling. Unlike the more cir-
cumscribed crisis of value addressed by Goldie, I note that existential despair itself
looks like a kind of intellectual position—one that has a feeling of certainty about it.
I ask whether or not this ‘position’ is most plausibly construed as a misguided affec-
tive evaluation that arises due to obfuscation of more reliable affective dispositions.
In order to make a convincing case for this, an account is needed of why blockage is
experienced as precisely the opposite, as a sweeping away of what previously prevented
one from appreciating an unbearable and inescapable truth. I offer such an account,
but also consider some objections to it. I go on to argue that the allure of existential
despair is at least partly attributable to blockage of affective dispositions, and conclude
by tentatively suggesting an additional line of response, which emphasizes the role of
interpersonal concern.
2. Tolstoy’s Well

In A Confession, Tolstoy recounts an experience of suicidal despair that would nowa-
days be diagnosed as ‘major’ or ‘severe’ depression. At one point, he conveys it in terms
of an “Eastern fable, told long ago.” A traveler runs from a beast and seeks refuge in
a well. At the bottom of the well is a dragon, and so the traveler is unable to climb
out or climb down. He clings to a twig growing from the side of the well, which two
5
See, for example, Ratcliffe (2005, 2008) for the view that so-called ‘bodily feelings’ need not and gener-
ally do not have the body or part of the body as their exclusive or even principal object.
mice—one black and the other white—nibble at in turn. As the traveler awaits his inev-
itable fate, he consoles himself by licking drops of honey off leaves that grow from the
twig, the taste of which distracts him from his plight. Tolstoy’s problem was that the
honey stopped tasting sweet:
So I too clung to the twig of life, knowing that the dragon of death was inevitably awaiting me,
ready to tear me to pieces; and I could not understand why I had fallen into such torment. I tried
to lick the honey which formerly consoled me; but the honey no longer gave me pleasure, and
the white and black mice of day and night gnawed at the branch by which I hung. I saw the
dragon clearly, and the honey no longer tasted sweet. I only saw the unescapable dragon and the
mice, and I could not tear my gaze from them. And this is not a fable, but the real unanswerable
truth intelligible to all. (Tolstoy 2005 [1882]: 17‒18)
This description makes salient the sense of revelation and certainty that character-
izes Tolstoy’s despair. It presents itself as the “real unanswerable truth,” and as some-
thing that was always lurking in the background but formerly eclipsed by distractions.
Once those distractions are swept away, he can no longer hide from a way of being that
offers only futility and then extinction.6 So something is ‘blocked’ here: a capacity to
immerse oneself in various pursuits. But the blockage yields apparent revelation: those
pursuits, however we might once have regarded them, now appear as mere distrac-
tions from something universal, dreadful, and inescapable. There is an analogy here
with more mundane experiences of losing ourselves in something to take our minds
off something else. One might spend an evening with friends and feel temporary relief
from the pain of bereavement, or go to the cinema to forget about an impending job
interview. But what Tolstoy describes is more profound—one flees not from some con-
tingent circumstance but from the structure of human life.7
It is important to distinguish the kind of despair Tolstoy describes from other kinds
of predicament that are sometimes associated with depression diagnoses, which
might also be labeled as ‘hopelessness’ or ‘despair’. Compare the following first-person
accounts:
“When depressed I feel I have no future and lose any hope in things improving in my life. I feel
generally hopeless.”
“I feel hopeless, as though there is nothing I can do that will ever truly improve my life. I often
feel like I’m in a rut, like I’m stuck.”
“Whilst depressed, I feel an impending sense of doom. I feel hopeless and useless, and my
self-confidence drops so low that sometimes I cannot even leave the house to buy food as I don’t
feel worthy to be taking up any space and time.”
6
We find similar themes in some of Heidegger’s works, where it is claimed that certain mood changes
involve the ‘awakening’ of a mood that was already there, rather than the replacement of one mood by
another: “ ‘Whatever is sleeping’ is in a peculiar way absent and yet there. When we awaken an attunement
[Stimmung], this means that it is already there. At the same time, it expresses the fact that in a certain way it
is not there” (Heidegger 1995: 60).
7
A comparison can be drawn here with Kierkegaard’s claim that most of us are in a state of unknowing
despair (e.g. Kierkegaard 1989).
“The world looks very different when depressed as I find my life becomes valueless. The world
seems very bleak and there seems to be no point in anything. All actions and tasks become
pointless and irritating. Daily tasks become a chore and social contact becomes a real hassle. The
problem with depression is you lose hope and then you get very self-destructive. I also find that
the world becomes a dark and dangerous place and I become unable to find any joy or happiness
in it.”8
We cannot make confident phenomenological claims on the basis of cursory

remarks. Even so, if these descriptions are taken at face value, several different forms
of ‘hopelessness’ or ‘despair’ can be discerned. Feeling that “I have no future” differs
from the conviction that no human being has a meaningful future, and “I am in a rut”
suggests something temporary and less profound still, where there remains an appre-
ciation that one’s own life could improve. The “impending sense of doom” in the third
response is not necessarily incompatible with a meaningful life but surely interferes
with it. Here, it is associated with lack of confidence in the efficacy of one’s actions, but
‘I am likely to fail at p’ is different from ‘p is of no worth’. The fourth response is closest
to what Tolstoy conveys. It does not apply to a specific project, a wider system of pro-
jects, parts of one’s life, or even the whole of one’s life. Instead, there is a dread-imbued
realization that all human life is bereft of value. However, Tolstoy’s revelation can be
put more strongly still: it is not just that one takes all human life to be without value;
one cannot even contemplate the possibility of its being otherwise; the experience has
a feeling of irrevocable certainty to it.9 That feeling is associated with or perhaps partly
constituted by an experienced absence of certain affective dispositions. Importantly,
it involves loss or blockage of types of affective disposition rather than loss of however
many tokens of those types. The situation is not merely that one can no longer gain
pleasure from p or immerse oneself in purposive activity q; the capacity to feel pleasure
in anything or immerse oneself in purposive activity is altogether absent. As a result,
there is no source of distraction from the well.10
However, this does not suffice to fully characterize the experience. Why does all
human activity appear but a futile distraction? A heightened and/or altered awareness
of mortality seems to be largely responsible, and Tolstoy couches existential despair in
terms of a negative response to the question “Is there any meaning in my life that the
8
These are representative responses to a questionnaire study on the experience of depression, which
I conducted with colleagues as part of the AHRC- and DFG-funded project ‘Emotional Experience in
Depression: a Philosophical Study’. Most of those who completed the questionnaire had diagnoses of major
depression, and approximately two in three reported being depressed at the time of writing. Unless other-
wise attributed, quotations that appear later in this chapter are also taken from questionnaire responses.
9
Here and elsewhere in the chapter, I use the term ‘experience’ in a fairly non-committal way. I do
not want to insist that the content of existential despair is wholly or partly integral to sensory perception.
However, I do maintain that it is not principally a propositional attitude of the form ‘I believe with confi-
dence that all human life is without value’. Such formulations originate in a felt evaluation that one is ‘struck
by’ something that has a kind of immediacy to it and is bound up with how the world ‘appears’.
10
See also Garrett (1994), Steinbock (2007), Webb (2007), and Ratcliffe (2015) for recent attempts to dis-
tinguish this type of despair from less profound experiences that might also be labeled as ‘hopelessness’ or
‘despair’.
inevitable death awaiting me does not destroy?” (2005 [1882]: 21). There is more to it
than that though. In addition to the poignant awareness of life as finite and meaning-
less, it involves (in Tolstoy’s case, at least) an unpleasant feeling of urgency, a kind of
teleological drive. There is a felt need to act for the sake of some end, which is rendered
insatiable by the ever-present sense of mortality:
Had I been like a man living in a wood from which he knows there is no exit, I could have lived;
but I was like one lost in a wood who, horrified at having lost his way, rushes about, wishing to
find the road. (Tolstoy 2005 [1882]: 19)
So the experience is something like this: a heightened sense of mortality comes to

light when the capacity for effortless, pleasurable immersion in activity is blocked, and
this renders worthwhile activity unintelligible. An agitated need to achieve something
lingers on, with no possible outlet. Why, though, should an appreciation of mortality
be incompatible with purposive activity? The answer, it seems, is that a sense of any
activity’s being worthwhile tacitly depends upon the possibility of its infinite teleologi-
cal development. This is incompatible with the extinction of every human accomplish-
ment, something one accepts as inevitable in properly grasping the nature of mortality;
there is the realization that everything we do will ultimately leave no trace upon the
universe. The association between mortality and futility can be better understood if
we also emphasize the theme of evil, something that is more prominent in some of
William James’ works, especially Varieties of Religious Experience (in a well-known
chapter entitled “The Sick Soul” where James quotes from and discusses Tolstoy’s A
Confession at length). For James, an intense awareness of human mortality and the
inevitability of suffering becomes an all-enveloping sense of the world as fundamen-
tally evil, a place in which we can never be safe, at home. Our projects crumble, given
that they are mortgaged upon a hope or faith in the goodness of life that is revealed as
unfounded:
The fact that we can die, that we can be ill at all, is what perplexes us; the fact that we now for a
moment live and are well is irrelevant to that perplexity. We need a life not correlated with death,
a health not liable to illness, a kind of good that will not perish, a good in fact that flies beyond
the Goods of nature. (James 1902: 140)
For those who James calls “sick souls,” the feeling of evil is ever-present: “the evil
aspects of our life are of its very essence, and that the world’s meaning most comes
home to us when we lay them most to heart” (James 1902: 127‒31). The theme is pre-
sent in Tolstoy’s account too when, for instance, he recalls witnessing an execution in
Paris some years earlier and feeling that the horror of the guillotine could never be
reconciled with a fundamentally good world that accommodates worthwhile human
activity.
How might one respond to such an experience? Tolstoy’s existential journey ends in
religious conversion. He comes to recognize that what first struck him as a truth about
all human life was actually more parochial, and concerned the privileged, parasitic
social elite to which he belonged. He distances himself from that way of life to discover
the faith of the peasants. Tolstoy is clear that there is no purely intellectual solution
to be found, as existential despair sweeps away the ground upon which all intellec-
tual endeavors rest. What he finds is a new way of living, more so than a new way
of thinking. However, his ‘solution’ is unsatisfying for various reasons. The view of
Tolstoy’s pre-conversion life that he presents in A Confession is one-sided and unchari-
table. Furthermore, after his conversion he was inconsistent and conflicted in many
respects, and was consumed until his death by an exceptionally unhappy and destruc-
tive marriage. As one biographer remarks, “once one is alerted to the danger signals, A
Confession, precisely because of its artless sincerity, is revealed as a transparent piece of
self-deception: transparent, that is, to everyone except the author” (Wilson 2001: 312).
Even setting aside such concerns, Tolstoy’s solution is historically specific and does
not provide clear guidance to us now. Who, for us, are analogous to his “simple labor-
ing folk” (Tolstoy 2005 [1882]: 62)? With so many cultures and attitudes to sample, it
is not at all clear where to look for practical wisdom. In any case, my question here is
somewhat different. For current purposes at least, I am concerned with identifying the
appropriate intellectual response, rather than with coping strategies.11 I do not seek to
address whether and how we ought to react to a first-person experience of existential
despair, or how we might console others who are afflicted by it. Instead, I want to ask
what grounds, if any, those of us who are not currently in existential despair have for
regarding it as misguided or, at least, as a less appropriate evaluation of the human pre-
dicament than our own. To be more specific, how can we determine whether Tolstoy’s
well contains a distorted view of human life, attributable to blockage of affective dis-
positions, or a revelation facilitated by loss of affective interference? In addressing that
question, I set aside the possibility of a religious response (which is not to dismiss such
responses). My aim is instead to explore how those with temperaments unsusceptible
to religious conversion (however much they might wish for it) and those who firmly
reject a religious solution ought to respond. If it can be shown that existential despair
involves a kind of ‘evaluative illusion’, attributable to affective disturbance, then we do
not need any further intellectual defenses against it, religious or otherwise.
3. Existential Despair as Psychopathology

A case for evaluative illusion could proceed as follows: (1) existential despair arises
due to depression; (2) depression is a pathological condition; (3) therefore, existential
despair is pathological; (4) therefore, existential despair is an unreliable guide to the
worth of human activities. It is debatable whether all cases of depression are ‘patholog-
ical’, but let us suppose they are. Two broad conceptions of ‘depression’ need to be dis-
tinguished (Radden 2009: 79‒80). On one account, depression is an as yet unidentified
11
However, I do not rule out the possibility of integrating an intellectual response into a therapeutic
response.
pathological process, which causes the various symptoms we associate with it. On
another, depression just is a ‘syndrome’ or cluster of symptoms. If a causal account
of depression and its symptoms is adopted, then it is unclear why existential despair,
construed as a symptom, should itself be pathological: if y causally depends upon x,
and x is pathological, it does not follow that y is pathological. By analogy, being ill with
influenza might give one time to sit back and reflect upon one’s pastimes in a way that
calls some of them into question. Any resultant insights should not be dismissed on
the basis that they would not have occurred without infection by the influenza virus.
However, if depression is identical with its symptoms and existential despair is a symp-
tom (one that does not feature in all cases of depression though), matters seem simpler.
Depression is its symptoms and depression is pathological. Therefore, those symp-
toms are pathological. But, without recourse to an aetiological account of depression
as a disease process, it is unclear what conception of pathology is at play here. And,
even if existential despair is branded ‘pathological’ in some biological sense, this need
not render it misguided. It has been pointed out many times that certain false beliefs
(including evaluative beliefs) could be biologically advantageous, while access to cer-
tain truths could put one at a biological disadvantage. An oft-used example is ‘I am
invincible in battle’, which conceivably enhances fighting ability in a way that increases
the likelihood of survival. So it can be conceded that existential despair is undesirable
in a biological sense without giving up on the view that it is revelatory. A simple appeal
to ‘pathology’ therefore fails to arbitrate.
Another approach is to look at how depression actually affects the capacity for eval-
uative judgment: does it render the relevant cognitive processes more or less reliable?
According to so-called ‘depressive realism’, it fosters more accurate evaluations, at least
in relation to certain themes, such as one’s social status, abilities, and degree of cul-
pability for undesirable outcomes (Alloy and Abramson 1988). One might say that it
blocks the influence of affective dispositions that more usually distort evaluation. The
general idea is nicely expressed in an earlier essay by Freud:
If [ . . .] he describes himself as a petty, egoistic, insincere and dependent person, who has only
ever striven to conceal the weakness of his nature, he may as far as we know have come quite
close to self-knowledge, and we can only wonder why one must become ill in order to have
access to such truth. (Freud 2005 [1917]: 206)
Hence it could be argued that depressive realism supports the case for despair, but
I doubt that this can be made to work. Proponents of depressive realism concede that
depression not only corrects certain biases; it also makes one more susceptible to oth-
ers (Alloy and Abramson 1988: 243). Furthermore, the case for depressive realism is
questionable in several respects. It is not always clear that there is an objective stand-
ard for comparison to support claims about the appropriateness or otherwise of an
evaluation; the design of some studies has been called into question; the experimen-
tal results are amenable to several interpretations; and almost as many findings are
inconsistent with it as are consistent with it (Ackermann and De Rubeis 1991). At best,
depressive realism seems to be a fragile phenomenon that shows up only under certain
conditions. Crucially, much of the empirical support for it involves subjects who are
not severely depressed, and the effect diminishes and disappears as severity increases
(Ghaemi 2007). So, if we make the plausible assumption that what Tolstoy describes
is generally associated with more severe forms of depression, the depressive realism
findings are inapplicable.
Does empirical evidence instead support the opposing view that severe depres-
sion is associated with unreliable evaluative judgments? It seems plausible to main-
tain that all stages of inquiry are motivated and guided by emotions of various kinds,
including curiosity, doubt, wonder, surprise, and satisfaction (e.g. Hookway 2002;
Thagard 2002; Morton 2010). Insofar as depression diminishes or even extinguishes
the capacity for some such emotions, it surely interferes with belief-forming processes,
especially where value judgments are concerned. Elliott (1999: 93‒7) raises the con-
cern that, even though a person’s reasoning might seem intact when she is depressed,
decision-making could be seriously impaired, as her access to cares and concerns that
would more usually shape it is impeded by an inability to feel:
To put the matter simply, if a person is depressed, he may be aware that a protocol carries risks,
but simply not care about those risks. [ . . .] When a person is caught in the grip of depression,
his values, beliefs, desires and dispositions are dramatically different from when he is healthy. In
some cases, they are so different that we might ask whether his decisions are truly his.
It is thus arguable that existential despair, which arises from an inability to feel and
consequent blockage of access to one’s values, is a deceptive, impoverished evaluation
of human life. However, that view is also problematic. Our epistemic capacities are
surely heterogeneous to some degree—a point that may well apply more specifically
to evaluative tendencies. The capacities needed to form an affective appreciation of
the irrevocable futility of all human life could be quite different from those needed to
make various other types of value judgment. Furthermore, an overarching and accu-
rate evaluation of all human action as futile could detrimentally affect the reliability of
other evaluations. First-person accounts of severe depression frequently convey a rad-
ical change to the experiential world—everything seems somehow different, dimin-
ished and dreadful, and the person feels irrevocably isolated from others:
Most of all I was terribly alone, lost, in a harsh and far-away place, a horrible terrain reserved for
me alone. There was nowhere to go, nothing to see, no panorama. Though this landscape sur-
rounded me, vast and amorphous, I couldn’t escape the awful confines of my leaden body and
downcast eye. (Shaw 1997: 40)
I awoke into a different world. It was as though all had changed while I slept: that I awoke not
into normal consciousness but into a nightmare. (Patient quoted by Rowe 1978: 269)
Surely such an experience would be unsettling enough to interfere with a wide

range of epistemic abilities, even if it did incorporate a profound revelation about the
structure of human life. Hence it is not enough to make a case for impaired evalua-
tive ability in some other context and then appeal to guilt by association. A disturbing
and accurate evaluation of human life could be precisely what impedes one’s ability to
evaluate in that other context.
4. Illusions of Truth

Instead of challenging existential despair on the basis that it is the product of unreli-
able cognitive processes, we could take issue with its content. A simple objection to
existential despair, considered as an intellectual position, is that awareness of mortality
just does not need to be associated with existential catastrophe. I can be well aware that
I and everyone else will suffer and die, but without all of my projects becoming unin-
telligible or the universe taking on an all-pervasive air of evil in the process. However,
it is arguable that both the content of the position and the attitude of acceptance are
partly constituted by feeling, and that neither can be fully appreciated without having
the relevant feeling. As Wynn (2005: 9) puts it, some feelings may “offer our only mode
of access to certain values.” So, when the association between mortality and futility is
casually dismissed, this could be due to confusion between the content of existential
despair and some other content that is superficially similar but also importantly dif-
ferent.12 Thus, even if our own thoughts about death and the worth of human action
do not add up to existential despair, we can still ask whether those who do suffer from
it might have stumbled upon a truth that we have the good fortune of being unable to
access.
One response is to reject the view that feeling contributes to the attitude or content
of belief, in this particular case or more generally. However, it is possible to supply a
phenomenological account of exactly how feeling does so in experiences of depres-
sion, which renders that objection unconvincing. In brief, depression involves a
diminution of certain kinds of feeling that also manifests itself as a draining of sig-
nificance from the experiential world. More usually, our experiences of and thoughts
about the world involve an appreciation that things could differ from how they
presently are in ways that matter. We have an awareness of there being salient pos-
sibilities of various kinds, which can be actualized by our own activities, the activi-
ties of others or by impersonal events. But the world of depression is bereft of the
possibility of significant change; experience and thought no longer incorporate the
sense that the future could differ from the present in ways that matter (good ways, at
least). Consequently, the world of depression is experienced as curiously timeless,
inescapable, and certain. This rough sketch of the phenomenology is consistent with
numerous first-person testimonies, all of which emphasize the loss of possibility. For
example:
12
This is consistent with Tolstoy’s well-known contrast in his short novel The Death of Ivan Ilych (Tolstoy
2004) between two different ways of believing that one will die. The protagonist comes to understand that he
will die, in a felt way that differs from conceding propositionally that all people die, that he is a person and
that he will die.
I have absolutely no faith, in fact, in anything. In a muddy way, I see that depression manifests
itself as a crisis of faith. Not religious faith, but the almost born instinct that things are fluid,
that they unfold and change, that new kinds of moment are eventually possible, that the future
will arrive. I am in a time-locked place, where the moment I am in will stretch on, agonizingly,
for ever. There is no possibility of redemption or hope. It is a final giving up on everything. It is
death. (Lott 1996: 246‒7)
The sense of inescapability is a product of the inability to entertain alternatives,

which renders one unable to grasp the contingency and transience of ‘how things cur-
rently seem’. An evaluation of life shaped by severe depression might feel certain, but
precisely because it is impoverished. Many first-person accounts of depression explic-
itly describe the experience in terms of lack of access to possibilities:
“I remember a time when I was very young—6 or less years old. The world seemed so large
and full of possibilities. It seemed brighter and prettier. Now I feel that the world is small. That
I could go anywhere and do anything and nothing for me would change.”
“It is impossible to feel that things will ever be different (even though I know I have been
depressed before and come out of it). This feeling means I don’t care about anything. I feel like
nothing is worth anything.”
“The world holds no possibilities for me when I’m depressed. Every avenue I consider explor-
ing seems shut off.”
“When I’m not depressed, other possibilities exist. Maybe I won’t fail, maybe life isn’t com-
pletely pointless, maybe they do care about me, maybe I do have some good qualities. When
depressed, these possibilities simply do not exist.”
This suggests a more promising response to existential despair, which runs as fol-
lows: in order to competently evaluate a state of affairs as p rather than q, one must
be able to first comprehend the possibility of q and then rule it out. If one’s ability to
even entertain the possibility of q were blocked due to lack of feeling, then one’s com-
mitment to p would reflect incapacity, rather than p’s relative plausibility. Many auto-
biographical accounts of depression include the observation that, while a person is
depressed, recovery from depression seems not only unlikely but inconceivable. As
one cannot contemplate the possibility of things ever differing from the present in a
positive way, depression itself feels inescapable. But the feeling of certainty is decep-
tive, arising due to the inability to contemplate something that is not only possible but
probable.13 We can understand the combination of feeling, content, and unwavering
conviction that characterizes existential despair in much the same way: progressive
blockage of an ability to contemplate alternatives leads to a pared-down evaluation of
the world that presents itself as comprehensive and irrevocable. If access to alternatives
were restored, it would again reveal itself as a contingent evaluation, and not a very
tempting one either. Consider the following account:
13
See Ratcliffe (2015) for a more detailed phenomenological analysis of depression, which includes and
elaborates upon some of the material in this chapter.
“When I am not depressed my feelings/emotions are totally different, because I can think clearly.
I can see a future for myself. I can feel happiness. I can see the joys in life. I can socialize. I can
be loving and friendly. When I am depressed, I am unable to think clearly. I feel sorrow, anger,
frustration, sadness, lonely, worthless, despair and mainly I feel like my life is not worth living
and I would rather be dead!”
Depression is associated with inability here; loss of possibilities interferes with for-
mation of value judgments. Now, not all instances of depression involve existential
despair of the kind discussed here. Even so, when it is present, we can maintain that it
arises in much the same way: what looks like revelation is actually limitation. Consider
the analogy with dreaming, which likewise involves an epistemic asymmetry: we
might not be aware that we are dreaming while we are dreaming, but we can usually
make the distinction with confidence once awake, when the limitations of the dream-
world become readily apparent. Existential despair is akin to the dream-world insofar
as it is oblivious to its shortcomings.
Unfortunately, matters are not so clear-cut. Canalization of belief through feeling is
perhaps widespread. When beliefs amount to mere intellectual play, commitment to p
rather than q might not demand a feeling of certainty. However, those convictions that
matter to us most, that regulate our activities and our aspirations, are not a matter of
putting ticks next to propositions; confidence comes as one ceases to feel the pull of sig-
nificant alternatives. Something like this arguably applies to certain religious beliefs,
as well as to our most cherished intellectual commitments. As in the case of existential
despair, the sense of certainty can involve an inability to entertain alternative possibili-
ties, or at least take them seriously. One not only believes that p; one becomes increas-
ingly unable to appreciate how anyone could possibly believe otherwise, as nothing
else feels at all salient. So, we can ask, why is existential despair singled out as intellec-
tually dubious, and not all evaluative beliefs that are held with strong conviction? One
difference is that these other cases involve inability to contemplate token possibilities
rather than types of possibility. One retains access to the various ways in which a state
of affairs could matter; it just so happens that q is not experienced as mattering in some
or all of these ways. The ability to evaluate is not deficient in any respect; one does not
form a belief that q does not matter because one is incapable of taking anything to mat-
ter. Existential despair is therefore a special case—one that is attributable to blockage
of access to certain kinds of value.
However, let us briefly turn to what Heidegger says in Being and Time about the
phenomenological role of anxiety (Heidegger 1962). In summary, the claim is that a
‘mood’ of anxiety amounts to a total loss of practical significance from the experienced
world. It is not that one no longer finds p, q, or r practically significant. Rather, one’s
ability to find anything practically significant is absent, temporarily at least. Even so,
Heidegger regards this as potentially revelatory. Ordinarily, he claims, we lose our-
selves in the everyday, public world in ways that eclipse the underlying structure of
human existence. By sweeping away the capacity to find things practically significant,
a capacity upon which the disposition to misinterpret ourselves depends, anxiety gives
us phenomenological access to something that would otherwise be obscured. If some-

thing like this is at all plausible (in this or any other case we might manage to cook
up), then it can be maintained that certain types of affective disposition systematically
mislead us in certain respects. In order to access what they obscure, loss or blockage
of a type of affective response is required. Hence lack of access to types of mattering in
depression need not cultivate illusion—it could instead free us from something that
hides the truth. Furthermore, we might argue that it is by no means unusual in this
respect, and that many different kinds of pervasive ‘mood’ or ‘feeling’ close down and
open up types of mattering in a variety of ways, rather than just determining what is
taken to matter in what way (Ratcliffe 2008).14
Another serious objection to the blockage view is that recovery from depression
does not always bring relief from existential despair. Some people describe their recov-
ery in terms of regaining something they were previously deprived of. For others,
however, existential despair becomes less salient with the return of certain affective
capacities but continues to lurk in the background like a preying monster, with a feel-
ing of irrevocable truth still attached to it:
“I do not have that ‘switch’, that ‘normal’ function, and those like me (other people that are
affected by mental illness) are able to see past the ‘programmed’ normality that the majority of
humans have and realize that there is no point to the world, there is nothing to look forward to,
humans simply exist to perpetuate themselves. [ . . .] This ‘explanation’ of how the world works
does not go completely when I come out of a depression, the thoughts are still there, they are just
lighter and further away . . .”
Even if depression is often or always the route by which one arrives at existential
despair, the sense of irrevocable futility can outlast the depression. The psychiatrist
Nassir Ghaemi (2007: 126) therefore raises the concern that certain treatments may
tackle the depression but leave the person in “existential despair.”15 We must conclude,
then, that despair cannot be attributed solely to blockage of an ability to contemplate
14
In contrast to existential despair, Heidegger maintains that existential anxiety is not incompatible with
purposive striving. Instead, it offers the potential for an ‘authentic’ form of engagement with one’s life, a way
of inhabiting time that reconciles the possibility of purposive activity with acknowledgment of finitude.
Does Heidegger thus offer a potential solution to the problem of existential despair? I cannot do justice to
the complexities of his view here, but I am unconvinced. So far as I can see, there is no reason to rule out this
alternative conception of temporal experience, offered by the phenomenological psychopathologist Eugene
Minkowski: “In life we march toward the future and we march toward death; and these two marches, while
seeming to be congruent, are in reality completely different from each other. The one is composed of that
which is great, infinite, and positive in the future, the other of that which is excluded, limited, and negative
in it” (1970: 137). Existential despair is consistent with what Minkowski describes. A teleological orientation
towards the future is disrupted, revealing something that it served to obscure, something that is in tension
with the kinds of purposive activity in which we are more usually immersed. And no unitary, ‘authentic’
sense of time is to be found.
15
A number of clinicians have attempted to distinguish clinical depression from despair or to associ-
ate what I call existential despair with a particular subtype of depression. For instance, Kissane and Clarke
(2001) and Clarke and Kissane (2002) contrast depression with what they call “demoralization syndrome,”
while Abramson et al. (1989) postulate a subtype of “hopelessness depression,” where hopelessness is what
causes the depression, with some cases involving “generalized hopelessness.”
alternatives, as it can persist in some form without blockage and still ‘feel like truth’.
What grounds, then, are there for rejecting the view that it amounts in such cases to a
legitimate or even uniquely appropriate evaluation of the human predicament?
5. A Partial Response

Even if those of us who are not in despair cannot rule out its being an accurate appraisal
of human life, we can also insist that we have no reason to be intellectually troubled by
it. As James (1902: 74) remarks, “in the metaphysical and religious sphere, articulate
reasons are cogent for us only when our inarticulate feelings of reality have already
been impressed in favor of the same conclusion.” One could argue, along such lines,
that what we have here are different ways of feeling, which crystallize into opposing
intellectual positions, where both the attitude and content of belief remain bound
up with feeling. There is no neutral ground from which to arbitrate between them.
Just as one cannot fully appreciate the content and pull of existential despair without
experiencing the requisite feelings, so too those in existential despair arguably lack
full experiential access to non-despairing forms of experience, even when they are not
obviously depressed. Neither party can be, or even ought to be, intellectually swayed
by the other, and we end up with a stalemate of conflicting feelings.
Having established a stalemate, one could then argue on pragmatic grounds for the
superiority of a non-despairing stance: choosing despair over hope is a lose–lose bet
(Garrett 1994). However, it is debatable whether and to what extent there is a choice
over despair; it has an affective allure that plausibly cannot be overridden by any
amount of cold calculation. A similar concern applies to Cooper’s (2002) view that
despair can be alleviated by nurturing a sense of the world as fundamentally mysteri-
ous. When stuck in Tolstoy’s well, one is unable to contemplate the possibility of the
world’s being mysterious in a fundamentally good or even indeterminate way. Any
sense of mystery that one is able to cultivate involves a sense of inchoate evil. So the
stalemate persists. Even if affective dispositions could be retrained, the worry remains
that success here would involve steering someone away from a sound evaluation of
human life to a more bearable illusion. One cannot simply entertain existential despair
and decide, on pragmatic or other grounds, to reject it, as the predicament incorpo-
rates its own allure. One would instead have to try to escape it, to forget it, to trick one-
self. There is also the concern that existential despair eventually gets the upper hand
anyway. James (1902: 140) points out how we at least glimpse something like Tolstoy’s
well when we are injured, fatigued, or sick: “so with most of us: a little cooling down of
animal excitability and instinct, a little loss of animal toughness, a little irritable weak-
ness and descent of the pain-threshold, will bring the worm at the core of our usual
springs of delight into full view, and turn us into melancholy metaphysicians.” So it is
arguably something that we are only untroubled by if we have not yet had the kinds of
experience that serve to reveal it. To be free of the pull of despair is to be ignorant of
something, and only for a time. If a view along such lines could be made convincing, it
would leave us with an intellectual case for despair, rather than a stalemate.
I will conclude by considering a different approach: what if, instead of trying to chal-
lenge existential despair, we accept it but attempt to mitigate it? Upon recovery from
depression, certain emotional dispositions are unblocked, the blockage of which is
debilitating regardless of any relationship it might have with despair. In the absence of
depression, the person is at least capable of immersing herself in activities again, and
of enjoying herself. Of course, it could be maintained that this just amounts to a revi-
talized capacity for distraction. However, it is not at all clear why all activities should
be incompatible with a heightened appreciation of mortality and finitude. Consider
activities such as building a sandcastle with one’s children or spending the day gar-
dening. One knows, from the outset, that the sandcastle will soon be washed away
without a trace, that many of the plants one handles with such care will die as winter
approaches, that all such activities are temporary and short-lived. It is doubtful that the
majority of sandcastle builders and gardeners fall prey to the illusion that matters are
otherwise, but their projects do not strike them as futile, incoherent, or unintelligible.
The point applies equally to all those intellectual and practical activities that are driven
by curiosity, fascination, aesthetic feelings, and perhaps a wide range of other affec-
tive tendencies. In short, much of what we preoccupy ourselves with does not tacitly
depend upon a conception of life teleology inconsistent with our mortality.
In response, it might be objected that a constant, felt appreciation of one’s unavoid-
able demise disrupts these activities too. But there is a distinction between interfer-
ence ascribable to the realization that one will die and interference caused by repeated
occurrences of beliefs that happen to have the content ‘I will die’. If I experienced inces-
sant occurrent beliefs with the content ‘Durham Cathedral is bigger than York Minster’,
they would no doubt disrupt my concentration too. But the content is incidental; my
life when I am not thinking ‘Durham Cathedral is bigger than York Minster’ is not in
conflict with the belief ’s content. Hence the disruptive effect of a psychological state
need not be wholly attributable to its content, and this applies to ‘death’ beliefs too.
What is disruptive is the intrusiveness of the occurrent thought that one will die, not
acceptance of the fact that one will die, even if we acknowledge that full recognition
of mortality can indeed be distressing. It is therefore arguable that a substantial pro-
portion of our activities are untarnished by the acceptance of existential despair, once
it is extricated from blockage of dispositions that are symptomatic of depression and
not just despair. Only certain kinds of project, with a distinctive kind of motivational
structure, present themselves as incompatible with mortality. To complicate matters,
it is difficult to determine exactly what kinds of project are vulnerable, if the evalua-
tive framework that threatens them incorporates feelings and cannot be fully accessed
without those feelings.
Hence we have at least a partial response to the problem of existential despair. This,
I suggest, can be supplemented by what may turn out to be an even more promis-
ing line of argument. Implicit in many accounts of existential despair is a curiously
individualistic way of construing life projects: what is of worth in my life; can the worth
of my projects withstand my mortality? It is notable that an all-enveloping sense of
alienation from other people is absolutely central to most experiences of depression:
“I feel disconnected from the rest of the world, like a spectator. I only see I was depressed when it
stops. It’s like dust, you don’t notice it until you wipe it off and see the difference.”
“I feel like I am watching the world around me and have no way of participating.”
People who look back on their depression experiences often emphasize the extent
to which they were lonely, self-absorbed, cut off from others. For example, here is how
one author describes her experience of watching the events of 9/11 unfold on television
while she was depressed:
It was that lack of moral outrage and absence of any feeling that, more than anything else, con-
vinced me that I had to do something to ease the terrible grip depression had on me, I was so
lost in my own world that I had ceased to have compassion or feeling for any other. If the sight
of bodies dropping from a burning building did not horrify me, that absence of feeling did.
(Brampton 2008: 176)
Perhaps many of those who have ‘recovered’ from depression but not from despair
remain in a state of loneliness and social isolation, something that might well involve
a heightened preoccupation with one’s own life and what it could possibly encompass.
However, it is by no means clear why all interpersonal cares, concerns, and commit-
ments should be rendered futile in the light of one’s mortality, everyone’s mortality, or
even by a conviction that the world is fundamentally evil. The point is nicely illustrated
by the 2011 film Melancholia, directed by Lars von Trier. Two sisters, Claire and Justine,
are confronted with the prospect of Earth’s imminent and unavoidable destruction by
the approaching planet Melancholia. What kinds of activity are appropriate or even
meaningful while the annihilation of humanity and anything it might have accom-
plished fast approaches? Justine, who is suffering from depression, tells her sister that
life on Earth is evil and that she somehow knows there is no life anywhere else. Her own
experiential world, which is much like Tolstoy’s well, is made concrete for the viewer
in the guise of the planet’s approach. Yet, as the end nears and her depression lifts, she
chooses to be with her nephew and to comfort him. Nothing has rendered that kind
of concern unintelligible to her, even though she dismisses as absurd her sister’s sug-
gestion that they await the end of the world with a glass of wine on the patio. More
generally, the content of existential despair does not preclude interpersonal concern,
other than when it is accompanied by losses of interpersonal feeling that are clearly
attributable to blockage and privation. Of course, this is not yet an intellectual solu-
tion. In other words, it does not enable those of us who are not in existential despair
to maintain, with confidence, that it amounts to an entirely erroneous evaluation of
human life. But it at least points towards the conclusion that existential despair, of the
kind that I have described here, is something that only certain kinds of project and cer-
tain kinds of human life are vulnerable to, those that are shaped by a pervasive sense of
disconnection from other people and an associated self-absorption.
Acknowledgments
The research for this chapter was carried out as part of the project ‘Emotional
Experience in Depression: A Philosophical Study’. I would like to thank the AHRC
and DFG for funding the project. I am also very grateful to Sabine Roeser, Cain Todd,
my project colleagues in the UK and Germany, and an audience at the University of
Osnabrück for very helpful comments.
References
Abramson, L. Y., Metalsky, G. I., and Alloy, L. B. (1989). “Hopelessness Depression: A Theory-
Based Subtype of Depression.” Psychological Review 96: 358–72.
Ackermann, R. and De Rubeis, R. J. (1991). “Is Depressive Realism Real?” Clinical Psychology
Review 11: 565–84.
Alloy, L. B. and Abramson, L. Y. (1988). “Depressive Realism: Four Theoretical Perspectives.” In
Cognitive Processes in Depression, ed. L. B. Alloy (New York: Guilford Press), 223–65.
Brampton, S. (2008). Shoot the Damn Dog: A Memoir of Depression (London: Bloomsbury).
Clarke, D. M. and Kissane, D. W. (2002). “Demoralization: Its Phenomenology and Importance.”
Australian and New Zealand Journal of Psychiatry 36: 733–42.
Cooper, D. E. (2002). The Measure of Things: Humanism, Humility and Mystery (Oxford:
Clarendon Press).
D’Arms, J. and Jacobson, D. (2010). “Demystifying Sensibilities: Sentimental Values and
the Instability of Affect.” In The Oxford Handbook of Philosophy of Emotion, ed. P. Goldie
(Oxford: Oxford University Press), 585–613.
Elliott, C. (1999). A Philosophical Disease: Bioethics, Culture and Identity (London: Routledge).
Freud, S. (2005 [1917]). “Mourning and Melancholia.” In On Murder, Mourning and Melancholia,
trans. S. Whiteside (London: Penguin), 201–18.
Garrett, R. (1994). “The Problem of Despair.” In Philosophical Psychopathology, ed. G. Graham
and G. L. Stephens (Cambridge, MA: MIT Press), 73–89.
Ghaemi, S. N. (2007) “Feeling and Time: The Phenomenology of Mood Disorders, Depressive
Realism, and Existential Psychotherapy.” Schizophrenia Bulletin 33: 122–30.
Goldie, P. (2000). The Emotions: A Philosophical Exploration (Oxford: Oxford University Press).
Goldie, P. (2009). “Getting Feeling into Emotional Experience in the Right Way.” Emotion
Review 1: 232–9.
Goldie, P. (2012). “Loss of Affect in Intellectual Activity.” Emotion Review 4: 122–6.
Heidegger, M. (1962). Being and Time, trans. J. Macquarrie and E. Robinson (Oxford: Blackwell).
Heidegger, M. (1995). The Fundamental Concepts of Metaphysics, trans. W. McNeill and
N. Walker (Bloomington: Indiana University Press).
Hookway, C. (2002). “Emotions and Epistemic Evaluations.” In The Cognitive Basis of Science,
ed. P. Carruthers, S. Stich, and M. Siegal (Cambridge: Cambridge University Press), 251–62.
James, W. (1902). The Varieties of Religious Experience (New York: Longmans, Green and Co).
Kierkegaard, S. (1989). The Sickness unto Death, trans. A. Hannay (London: Penguin).
Kissane, D. W. and Clarke, D. M. (2001). “Demoralization Syndrome: A Relevant Psychiatric
Diagnosis for Palliative Care.” Journal of Palliative Care 17: 12–21.
Lott, T. (1996). The Scent of Dried Roses (London: Viking).

Minkowski, E. (1970). Lived Time: Phenomenological and Psychopathological Studies, trans.
N. Metzel (Evanston: Northwestern University Press).
Morton, A. (2010). “Epistemic Emotions.” In The Oxford Handbook of Philosophy of Emotion, ed.
P. Goldie (Oxford: Oxford University Press), 385–99.
Radden, J. (2009). Moody Minds Distempered: Essays on Melancholy and Depression
Ratcliffe, M. (2005). “The Feeling of Being.” Journal of Consciousness Studies 12: 43–60.
Ratcliffe, M. (2008). Feelings of Being: Phenomenology, Psychiatry and the Sense of Reality
Ratcliffe, M. (2010). “The Phenomenology of Mood and the Meaning of Life.” In The Oxford
Handbook of Philosophy of Emotion, ed. P. Goldie (Oxford: Oxford University Press), 349–71.
Ratcliffe, M. (2015). Experiences of Depression: A Study in Phenomenology (Oxford: Oxford
University Press).
Rowe, D. (1978). The Experience of Depression (Chichester: John Wiley).
Shaw, F. (1997). Out of Me: The Story of a Postnatal Breakdown (London: Penguin).
Steinbock, A. (2007). “The Phenomenology of Despair.” International Journal of Philosophical
Studies 15: 435–51.
Thagard, P. (2002). “The Passionate Scientist: Emotion in Scientific Cognition.” In The Cognitive
Basis of Science, ed. P. Carruthers, S. Stich, and M. Siegal (Cambridge: Cambridge University
Press), 235–50.
Tolstoy, L. (2004). The Death of Ivan Ilych and Other Stories (Ware: Wordsworth Classics).
Tolstoy. L. (2005 [1882]). A Confession, trans. A. Maude (Mineola, NY: Dover Publications).
Webb, D. (2007). “Modes of Hoping.” History of the Human Sciences 20: 65–83.
Wilson, A. N. (2001). Tolstoy: A Biography (New York: W. W. Norton).
Wynn, M. R. (2005). Emotional Experience and Religious Understanding: Integrating Perception,
Conception and Feeling (Cambridge: Cambridge University Press).
Index
abortion 201, 202, 207, 209 performative or enacted awareness of

Abramson, K. 187 n. situation 223
Abramson, L. Y. 236, 241 n. specificity of the object of 65
acceptance 9, 200, 206–7, 238, 243 strong normative constraints bearing on
accessibilism 111 23 n.
Ackermann, R. 236 affectivity 215, 216, 219, 223, 226
action-oriented theory 215, 222 n. agentive 220, 224
active-affective minimal self 9, 213–14, 223–5 undisturbed 225
aesthetic value 7, 109 agency:
aesthetics 1, 3, 4, 32 n., 66, 91, 92, 113, 155, 243 emotion and 9, 212–28
psychology of 188 limits of 194
affect 86, 160–1 moral 184, 191, 213 n.
appropriate 78, 85 shame as damaging to 8–9
‘catching’ 188, 193 Ajax (Sophocles) 186–7
cognitive relation thoroughly infused Alloy, L. B. 236
with 192 n. Almerding, A. 142
explicit attributions to extraneous Alston, W. 110, 115, 171
sources 176 n. Anderson, A. K. 53 n.
negative 5, 46, 47–8, 59, 64, 69 Anderson, B. 216
positive 5, 46, 57, 60, 63–4, 68, 69 anxiety 57, 59
affect heuristic 114–15 extreme 68 n.
affection 72–3, 75, 101 free-floating 33
benevolent 73–4 general 108
feelings and 76, 77, 79, 82, 190, 193 individual differences in 60
negative 74, 76 may lead to conceptual fragmentation 60,
positive 74, 190 64, 69
withdrawing 208, 209 phenomenological role of 240
affective access 192, 193 primarily elicited 58
affective atmosphere 216 apparently objective values 6, 90–104
affective dispositions 241 Appiah, K. A. 176 n.
blockage of 231, 235, 236 approbation 74, 79, 80, 82, 184, 200
experienced absence of 233 and disapprobation 72–3
loss of 230, 231 appropriate emotions 8, 154, 159, 163 n.
affective perceptions 213 n., 214 reasons not to have 175
affective pull 213 see also inappropriate emotions
affective responses 22 appropriateness 19, 82, 117, 148–9, 150, 152, 161,
appropriate or fitting 6 164, 176, 180
emotion defined as 78 assessment of 151 n., 155
loss or blockage of a type of 241 comparison to support claims about 236
structure of caring that incorporates determined 178
dispositions towards 231 justifying 175, 179
successful imagining seems to depend reflecting upon 200
on 156 specification of 151
affective self-construal 223–4, 225 see also fittingness
affective states/aspects 1, 75, 78–9, 83, 87, 95, appropriateness conditions:
155 n., 156, 173, 174, 177 n., 183, 190, 218, 222, representation-dependent 156
237, 242 states with different kinds of 158, 159
feelings 9, 77 Aristotle 38, 39 n., 187, 190, 193, 199, 217–18, 219
judgement undermined by the absence of 76 Arpaly, N. 129 n.
248 Index
attention 48, 80, 137, 139, 142, 176 n., 191, 200 benevolent regard 193
captured 28, 52, 128 Bengson, J. 83 n.
changing patterns of 100–1 Bennett, J. 129 n.
connection between emotion and 5, 52 Bergmann, M. 111, 116 n.
consumption of 52 Berlin, H. A. 174 n.
directed 81 Bermudez, J. L. 20 n.
grabbing 220 Blackburn, S. 103 n., 165 n.
increased 138 Block, Ned 32, 44 n.
visual 57–9, 60 Brady, M. 4, 22 n., 52 n., 103 n., 107, 116–17,
well-known emotional effect on 53 118–19, 120, 125 n., 126, 128, 139, 189
attentional focus: Brampton, S. 244
broadened 52, 56, 59, 61, 64, 67 Brandt, J. 59
consequential 62, 63, 64 Branigan, C. 57–9, 60 n., 63 n.
constitutive 62, 63 Brentano, Franz 38 n.
narrowed 59, 62, 64, 67 Brett, L. 185 n.
negative and positive emotions’ effects on Brewer, B. 23 n.
breadth of 60 Brewer, T. 112, 113
persistence of 54 Brison, S. 190 n.
valence and 61, 63 Brothers Grimm 164
attitudes 22, 82, 83, 169, 200, 201, 202, Buras, T. 114 n.
219 n., 235 Butler, L. 155 n.
acceptance 238
appropriate 199 Calhoun, C. 193
assent 134 Cartesian ‘think’ 38 n.
authorial, distance or disapproval 164 Caston, V. 39 n.
bodily 5, 15, 26, 28, 29 categorization 2, 33, 35, 37 n., 57, 60
connection between values and 94 anxious subjects and 64, 68–9
desired 19–20 different forms of 61
epistemic 110 influence of positive affect on 63
evaluative 28, 29 Chaiken, S. 200 n.
judgment-like 19, 20 n., 25–6, 199 Chandrasekaran, R. 186 n.
objective truth unachievable in absence chaos 67
of 77 Chused, J. 189
perception-like 25–6 Cicero 191
positive 163 Claparède, E. 27
pro or con 112 Clarke, D. M. 241 n.
propositional 18 n., 131, 133, 233 n. Clore, G. 57, 59, 107, 119, 176 n.
trust 170, 171 CMEs (cognitive moral emotions) 74–81, 83
unfaked 140 Coalition Provisional Agency (US et al) 185
see also content (attitudes and); FA analysis; cognitive capacities:
reactive attitudes complex 22
attitudinal theory 5, 15, 25–9 limited 20
Auerbach, E. 163 n. cognitive engagement 193
awareness-of-awareness 38–41, 43–5, 47 cognitive phenomenology 34, 39 n., 40, 46, 50
advocates of 36
Baehr, J. 206 very primitive 37
Baghdad 184, 185 Colombetti, G. 23, 223 n.
Basso, M. R. 57, 59 common sense 26, 69, 98, 138, 202
Baudelaire, Charles 163–4 First Principle of 179
Bayne, T. 34 n., 37 n. supposed 99 n.
Bealer, G. 83 n. compassion 73 n., 179, 186, 190, 192, 193, 194, 207,
Bedford, E. 214 209, 244
being-in-the-world state 216 n. acceptance and 9
beliefs, see emotion-based beliefs; evaluative acknowledgement through signalling
beliefs; false beliefs; true beliefs of 189
benevolence 9, 179, 189, 202 attempts to show disrupted by bureaucracy 8
see also affection see also self-compassion
Index 249
Conee, E. 111 facing 141

conscientiousness: grief at 99, 118
emotional 180 inevitable 234
epistemic 8, 172, 180 life and 143–5, 201, 202, 234
consciousness 27, 33–5, 37, 128, 169, 209, 237 march toward 241 n.
access 32 sadness about 38, 42, 43, 46, 47, 50, 96, 97, 98,
givenness of objects and states of affairs 101, 149
in 32 shame for not having prevented 9
sentience and 66 sorrow at 62
see also self-consciousness death certificates 185–6
content 5, 29, 33 n., 52, 67, 79, 84, 155, 208 death of friends 38, 42, 50
alien 183 disvalue of 43, 46–7
assertion of 20 n. defence mechanisms 6, 201–2, 203–10
attitude and 18, 19, 21, 26, 27, 28, 230, 238, deconstruction of 9, 209
242 Deigh, J. 18 n., 19, 184 n.
cognitive 173 n., 177, 178, 189 n. Deonna, J. 1 n., 3 n., 4–5, 20–2 nn., 25–7 nn., 96
cognitive conflict in 133, 135 n., 98 n., 103 n., 121, 165 n., 184 n., 214,
conceptual 20 n., 25 219 n.
descriptive 114 depression 33, 59, 65, 108, 203 n.
emotional 222 n. anger repressed or transformed into 208 n.
evaluative 15, 19, 26, 64, 65, 94, 214 clinical 241 n.
existential despair 233 n., 238, 242, 244 melancholic, profound 225
external 36, 38 severe 9, 214, 224, 237
factual 214 Derryberry, D. 59, 60, 61–2, 63, 64, 65 n., 68 n.
feeling, conviction and 231, 239 desire(s) 18, 19, 24, 73, 98, 100, 113
genuinely and fully repudiated 134 acceptable to fulfil 202
imagining 148, 156, 157 n., 161, 163, 164 acting in accordance with reasons 219
incidental 243 beliefs and 5, 7, 159, 169, 171, 173
mental, self-directed 224 changing 170
non-conceptual 22, 23, 25 n. emotions and 78, 159–61
passion 200 evaluative outlook conception of 112
phenomenological 35, 36, 38, 39, 44 goal-directed 131
sensory 36 Humean 79
story 137, 147 natural 171, 172
theory of 35–8, 41 right 78
truth-assessable 20 n. Sartrian 139
see also propositional content; satisfiable 172
representational content understood as prime sources of human
contrastivity 142–3, 144 action 219
Cooper, D. E. 242 wishes and 199, 205
Cottingham, J. 200 n. see also passions
Cramer, P. 202, 203 n., 204, 208 despair:
Crane, T. 20 n., 21 n., 125 n., 130–1 attempts to distinguish clinical depression
Creese, A. 20 n. from 241 n.
Currie, G. 6, 7–8, 148 n., 155 n., 156 n., depressive realism supports the case for 236
158 n. suicidal 231
unknowing 232 n.
Damasio, A. 3, 27, 158 n., 174 n., 176–7 nn. see also existential despair
Dancy, J. 4, 5–6, 83 n., 85 n. Dijksterhuis, A. 200 n.
D’Arms, J. 2 n., 6, 17 n., 85–6, 90 n., 91, 92, 94–7, disapprobation 72–3, 110, 184
99, 101, 102, 184, 229, 230 Dolan, R. J. 53 n.
Daubman, K. A. 60 Döring, S. A. 3, 6, 7, 21 n., 22 n., 108, 129–30 nn.,
Davidson, R. 1 n., 3 n. 132–4 nn., 213 n., 214
De Rubeis, R. J. 236 Doris, J. M. 116, 200 n.
death 48, 146, 232, 235, 238, 239, 243 Dretske, F. 23 n., 36 n.
causing by accident 183 Driscoll, D. M. 174 n.
enemy-inflicted combat 187 Driver, J. 142
250 Index
Eagle, M. N. 208 evaluative properties 17, 28, 32, 54, 91, 95,
Eisenberg, N. 188 n. 121, 189
Elgin, C. Z. 108 emotions and 5, 16, 21, 26, 33, 34, 40, 49,
emotion-based beliefs 116, 117, 118 98, 134
justification of 7, 114, 121 evaluative phenomenology and 34, 36,
emotional phenomenology 4, 5, 19, 29, 43, 46, 38, 50
212 n., 213 non-evaluative properties and 120
apparent objectivity and 97–102, 103 perception-like experiences of 15
appeal to 96 perceptual relation with 24
bodily nature of 27 real 99
inaccurate conception of 6 response-dependent accounts of 102
understanding 9 thick 109, 112
emotional responding: evaluative property attributions 35
appropriateness in 161 evaluative phenomenology and 41–8
norms of 149–54 evaluative representations 26, 33, 35, 49
emotional self-trust 169–82 experiential 5
empathy 148, 202, 208 Evans, S. 121 n.
and imagining 188, 193 evidence 36, 53, 62, 81, 111, 112–13, 114, 140, 155,
see also self-empathy 163 n., 164, 174, 179, 203 n.
empirical evidence 5, 52, 54, 55, 56, 69, 237 anecdotal 115
fixated 64 appropriate level of 159 n.
good 119 best 161
relevance of 64–5 conclusive 63
three types of 57–61 corroborating 119
Engel, P. 165 n. defeating 121
Epicurean teaching 191 experimental 64, 65, 67, 68, 202 n.
epistemic justification 6, 107, 110, 111–12, 113, external 107
117, 205 failing to respect norms of 158
emotional justification depends on 118 good 152, 153
sources of 7, 108, 119, 121 internal 107
sustained and systematic defence of 7 perceptual 120
ultimate, circularity of 171 piling up of 154
epistemic ridicule 179 rule of 152, 153, 154
epistemic self-trust 8 scant 117
basic 169 sensory 107
rational inescapability of 171–2 stronger 117
reflective 172 suggestive 16
epistemic value 52, 55, 56, 63, 64, 65 supporting 120
positive 69 target-detection 63
epistemology 3, 6, 77, 90, 121, 140, testimonial 118, 152–3
142, 207 throwing out 141–2
justification 108, 110, 111 weighing of 146, 153
moral 199, 204 see also empirical evidence
essential contestability 90–4, 95, 103 existential despair 8
evaluative beliefs 230, 231, 236, 240 content of 233 n., 238, 242, 244
emotional feeling and 229 evaluating 9, 229–46
justified 111–14, 118, 119, 120, 121 phenomenology of 10
thick 110, 121 externalism 33 n., 111, 116
evaluative emotions 145
evaluative judgments 15, 18–21, 22, 25, 110 FA (fitting attitude) analysis 6, 17 n., 90–4, 95,
depression affects capacity for 236 98, 102
unreliable 237 factive emotion 138
see also value judgments Fallujah 184
evaluative knowledge: false altruism 202, 207, 209
acquisition of 1 false beliefs biologically advantageous 236
grounding 4, 6 potentially 98
justifying 3 rational beliefs are sometimes 149
Index 251
risk of acquiring 155 defences often initiated for 207 n.

feedback 62, 220 moral 176 n.
negative 59, 61 Goodman, N. 140, 152 n.
Feldman, R. 111 Gopnik, A. 138
fiction 4, 7–8, 32 n., 82, 146–66 Gordon, R. 3 n., 138
fictive emotions 7, 138, 150 n., 157, 165 Grandin, T. 113–14, 119
representations associated with 156 Greene, J. 2
rule of representational correspondence Greenson, R. 189
constrains 151 Greenspan, P. 3 n., 20 n., 33 n., 78–9, 111, 117,
Fine, C. 200 n. 125, 129 n.
fittingness 4, 8, 88, 91, 94, 131, 169 Griffiths, P. E. 215 n., 222 n., 223
assessments of 96, 97 Griswold, C. 184 n., 193
illuminating notion of 95 guilt 18, 62, 92, 189, 192, 201
judgements of 90, 96 accident 187
justifying 177 anguish of 191
objectivity of 6, 96, 98, 99, 101, 102, 103 by association 237
plausible account of 95 confused judgement of 153
relativistic notion of 6, 93, 95, 96, emotional subject’s focus framed by 189
98, 103 epistemically irrational 187
way to formulate 92 n. help to cope with feelings of 8
Foley, R. 171 horrific and unabated 187
Frankfurt, H. 164 n. important philosophical work on 184 n.
Franklin, Benjamin 178 overwrought 187
Frederickson, B. L. 57–9, 60 n., 63, 67–8 persistence of 54
free jazz 67 punishing 184
Freud, Anna 202 shame covering up/masking by 183, 184
Freud, Sigmund 188 nn., 189 n., 206, 236 subjective 183
Friend, S. 150, 158 typically experiencing 194
friendship 16, 130, 139, 140, 141, 144, 204, Gunther, Y. H. 26 n., 99 n.
220, 232 Gurwitsch, A. 38 n.
best 190, 193
see also death of friends Hague, The 103 n., 121 n., 183
Frijda, N. H. 107, 215 n. Haidt, Jonathan 2, 175–6, 200 n.
Froese, T. 216 Hall, Major Jeffrey 184–6, 192
Fuchs, T. 216 Harman, G. 36 n., 116
Harmon-Jones, E. 61 n.
Gable, P. 61 n. Hegel, G. W. F. 66 n.
Gallegos, F. 195 n. Heidegger, Martin 213 n., 216, 224 n.,
Ganzfeld effect 138 232 n., 240–1
Garrett, R. 233 n., 242 Helm, B. 3 n., 7, 22 n., 23 n., 125–6, 129 n.,
Gaspar, K. 57, 59, 107, 119 135, 213 n., 214, 220, 221 n.
Gendler, T. 148 n., 158 n. Helm objection 124, 125, 126, 130, 134
generalization problem 52, 56–61, 64 Herzberg, L. A. 26 n.
Geneva 165 n., 183 n. Hitchcock, C. 143
see also University of Geneva Hjort, M. 163 n.
Gestalt 27, 109, 132, 221 hodological space 221, 223
Ghaemi, N. 237, 241 Holocaust 113
Giardina, A. 142 Hookway, C. 229 n., 237
Gibbard, A. 84, 160 Horgan, T. 37 n.
Gilmore, J. 149 n., 151 n., 165 n. Hume, David 2 n., 32, 79, 124, 125, 188, 193
Glazer, T. 194 n. Hurley, E. 191
Goldie, P. 1 n., 3 n., 5–6, 9–10, 86, 87, 88, 97 n., Hursthouse, R. 201, 202
103 n., 107, 131, 165 n., 183, 184 n., 187 n., Husserl, Edmund 38 n.
188–9, 191–2, 193 n., 194, 200, 214, 216–17, Hutchinson, D. S. 218 n.
229, 230–1
good reasons 17, 55, 91, 112, 115, 117–18, 120 Ichino, A. 155 n., 156 n., 158 n.
belief can be based on 110 imaginative engagement 148, 193
252 Index
imaginative resistance 148, 154, 155 wrongness 83

imagining 24, 62, 91, 97, 100, 134, 138, 143, 151 n., see also evaluative judgments; moral
157–8, 162, 169, 173, 174, 190–1, 192, 208, 217, judgments
220–1 justification 3, 10, 20, 25 n., 96, 129, 130, 149, 159,
believing and 159, 160–1 179, 189, 200, 204, 209
constraints on 146–9 circular 175, 178
contents of 163, 164 defeasible 115
empathy and 188, 193 emotional 7, 109–10, 111, 117–21, 169, 181
successful 156 emotions are opposed to 2
vivid 98 externalist theories of 115, 116
inappropriate emotions 18, 160 foundationalist and coherentist theories
intellectual emotions 229 of 111
intended emotions 163 internalist and evidentialist theories of 115
intentional content 38, 39, 44, 108 non-circular 177, 180, 181
claim that all content is 36 non-testimonial (typically perceptual) 117
emotions have 130, 214 partial 180
evaluative 4 pragmatic or moral 110
phenomenal character and 3, 4 prima facie 115, 116
sense perceptions have 130 reflective 171, 172
intentional objects 65 n., 79, 97, 98, 172 reliability of 8
appeal to 96 sense perceptions can confer 108
characteristic 94–5 sources of 114, 116 n., 117, 119–20
intentionality 5, 34, 109 n., 131 ultima facie 115, 116
evaluative 17 see also epistemic justification; justificatory
experiential 41, 44, 45, 220 n thesis
goal-oriented or teleological 220 n. justificatory thesis 7, 108, 110
world-directed 3, 231 argument in favor of 111
internalism 115 common reason for rejecting 116
awareness 111 modified to fit coherentist theories 111
Iraq, see Baghdad; Coalition Provisional objections to 115, 121
Agency; Fallujah; Hall; Saddam remains unconvincing to many 114
irrationality 7, 10, 20, 22, 127, 135, undermining 116, 120
137, 160, 187
intuition that recalcitrant emotions Kadlac, A. 121 n.
involve 125, 128 Kahneman, D. 2
Isen, A. M. 60, 63, 68 Kamtekar, R. 116
Kant, Immanuel 2 n., 224
Jackson, F. 114 Kass, L. 176 n.
Jacobson, D. 2 n., 17 n., 85–6, 90 n., 91, 92 n., 184, Kedem, P. 60
229, 230 Kennedy assassination, see JFK
Jaegher, H. de 216 n. Kennett, J. 200 n.
James, William 27, 212, 234, 242 Kenny, A. 18 n., 134 n., 214
JFK (movie 1991) 150, 155, 156, 158 Kierkegaard, S. 232 n.
Johnson, D. 121 n. Kimchi, R. 57, 58
Jolley, K. D. 121 Kissane, D. W. 241 n.
Jones, K. 129 n., 133, 184 Klein, J. 214 n., 226 n.
judgment 29, 109, 114–15, 117, 169, 172, 230 n. knowledge 9, 76, 115, 119, 126, 132, 134
acceptance of logical consequences fear of 205
of 23 justification is necessary for 110–11
constitutive 213 n. love of 205
correct 173 moral 201, 206, 209
fully fledged 75 nonevaluative 92 n.
justifying 175 sensory 3
legal 176 see also evaluative knowledge;
perceptual 23 n. self-knowledge
reflective 180 Kohut, Heinz 188 n.
sympathy the necessary effect of 73, 74 Kovakovich, K. 158 n.
Index 253
Kriegel, U. 38 n., 39 n. negative 59, 64, 65, 69

Krueger, J. 216 n. objectless 109 n.
Kukla, R. 183 n. pervasive 241
Kvanvig, J. 110, 120 n. positive 59, 60, 64, 65, 68, 69
transient 229
Lacewing, M. 9, 200, 208 Moore, G. E. 132
Lamb, R. 109 n. moral approbation 73, 74
Lambie, J. A. 215 n., 221 n. moral emotions 5, 7, 72, 85
Lance, M. 183 n. see also CMEs
Lang, G. 90 n. moral judgments 72, 73, 74, 82, 179, 209
Lansky, M. 184 n., 187 accompanied by affections and
Lavers, S. 163 n. feelings 76
Lear, J. 204, 207 emotions in 2, 4, 77, 78, 79, 176 n., 180
LeDoux, J. 55 n., 86 justified 2
Lee, Win-Chiat 121 n. reason for 84
Lemma, A. 203 moral repair 8, 184, 193
Lenman, J. 199, 201 promise of redemption through 187
Levinson, J. 164 n., 215–16 nn., 216 moral value 75, 76, 77, 218
Lewin, K. 221 n. Moran, R. 200 n., 207
Lewis, C. S. 87 n. Morris, H. 184 n.
Lewis, D. 152 n. Morton, A. 7, 165 n., 229 n., 237
Lewis, H. 184 n. motivational aspects 53, 62, 63, 72, 86, 97, 129,
Lipps, Theodor 188 139–40, 201, 204, 208, 237
Little, M. 77–8, 79 changes 83, 84, 95
Litz, B. 185 n. distinctive kind of structure 243
Livingston, P. 163, 164 mechanisms that impervious to 23
Locke, John 38 n., 205 moral judgement involves 78
Lott, T. 239 presentation 85
lying 152, 203, 209 motivational force 213
Lyons, W. 3 n., 134 n. emotions and 128, 214, 219
motivational pull 3
Mackie, J. L. 32 situation-directed 220
Macnamara, C. 183 n., 184 n., 191, 194 n. Müller, J. M. 226 n.
Maguen, S. 185 n. Müller-Lyer illusion 83–4, 125, 126,
Marcel, A. J. 215 n. 127, 132
Martin, A. M. 184 n. Mulligan, K. 16 n., 21 n., 103 n.
Mary the Scientist case, see Jackson Murphy, J. 184 n.
Massin, O. 103 n. music 67, 113, 215, 216
Mayek, Joseph 187
McDowell, J. 23 n., 79, 199, 215 n. Nagel, T. 218–19
McKaughan, D. 114 n., 121 n. narratable conception of self 191–2, 194
McLaughlin, J. 189 Nash, W. 185 n.
Mele, A. R. 163, 164 Neff, K. 191
Meltzoff, A. 138 negative emotions 5, 52, 65, 68 n.,
Merleau-Ponty, M. 216, 223 115, 204
Mesquita, B. 107 attention 60, 61, 62, 65
metaethics 1, 2, 4 categorization 69
Metzger, W. 138 distinct 57–8
Mikulincer, M. 60 local bias 63
Miller, C. 116, 119 n., 121 n. personality traits associated with 59
Minkowski, E. 241 n. target detection 60–1, 62, 63
Montague, M. 4, 5, 37 n., 97 n., 99 n. way of reducing 69
Montaigne, Michel de 124, 125 neuropsychology 3
moods 33, 108, 158, 172, 215 n., 217, 231, Nichols, S. 160–1
232 n., 240 Nida-Rümelin, M. 103 n.
disruptive 230 Nietzsche, Friedrich 202
happy 67, 68 Noë, A. 215 n., 216 n.
254 Index
Nozick, Robert 5, 65–7, 68–9 more important 55

Nussbaum, M. 3 n., 18 n., 34, 107, 173 n., 176, processes that may distort 201
184 n. reflecting upon appropriateness of 200
revelations and 208
objection 4, 7, 52, 129, 177–8, 205, 217, 231, 241 role in moral enquiry 200, 201
common source 116–21 transformation of the structure of 209
limited 189 Paz, D. 60
unconvincing 238 Peano axioms 43
unreliability 114–16 Pelser, A. C. 4, 6–7, 24 n., 109
see also Helm objection perception 15, 25–6, 114, 130
objectivity 4, 10 affective 213 n., 214
apparent 6, 97–102, 103 see also sense perceptions
appearance of 99 perceptual experiences 37 n., 39, 44, 107,
emotions are opposed to 2 112, 120
objects 35, 53, 103, 108, 120, 149, 216 n. characterized 23
abstract 24, 150 n. emotions grounded in 24
apprehended as losses 17 justificatory role for 111
appropriate 150 n. perceptual theory/model 4, 15, 21–5, 26, 29, 54,
attitudes towards 26 68, 114, 121, 125
belief in the reality of 157 defence of 7, 110, 124
bodily awareness towards 27 philosophers sympathetic to 3
categorized 60, 63, 64, 68–9 Pettigrove, G. 114 n.
constitutive attention to 62, 64 Pettit, P. 184
construal of 109, 112 Phelps, E. A. 53 n.
emotional 61 phenomenal coupling 212, 215–19
evaluations of 32, 33, 34, 50 phenomenology 8, 108, 157, 214, 219 n., 223 n.,
fitting 8, 169, 170, 173, 177, 180 224 n., 230, 233, 238, 239 n., 240, 241
formal 17, 28, 29 agentive 34 n.
givenness of 32 colour 34, 40, 48
inconsistent representations of 158 essential (constitutive) feature of 38
intrinsically linked 145 evaluative 5, 32–51
negative or positive value of 46 existential despair 10
particular 131 metaphysical link between intentionality of
perceived 20 n. experience and 44
perceptual 24 objective 103
potentially significant/important 54, 55 salient 21, 25 n., 27
properties of 16, 24–5, 36, 41, 42, 45, 50 sensory or bodily 5, 33–4, 36, 39, 43–4, 49, 50;
reasons to favour or disfavour 199 see also cognitive phenomenology
right 116 shape 34, 40, 48
situational 117 sui generis kind of 5, 33, 34, 36–7, 40, 97 n.
value-laden 109 taxonomy of 35
values exemplified by/instantiated in visual-shape 40
16, 112 see also emotional phenomenology
see also intentional objects; physical philosophy of mind 1, 34, 36 n.
objects physical objects 36, 41
Oddie, G. 112 objective property of 46 n.
paradigm cases of properties of 42
Palmer, S. 57, 58 Pigman, G. 188 n.
Paolo, E. di 216 n. Plantinga, A. 205
Parkinson, B. 215 n. positive emotions 5, 52, 57, 115
passions 199, 202, 204, 206, 210, 215–19 attention 60, 61, 63, 64, 65
apparent good presented by 200 categorization 68, 69
awareness of 207 global bias 63
content of 200 help in moral recovery 184
emotions as 222, 223, 225 seeing things as good can reflect and
ideals can be defences against 207 generate 69
inappropriate, thus misleading 200 target detection 61, 62, 63
Index 255
traits associated with 59 self-trust 8, 169

possibility 22 n., 28, 29, 33, 44 n., 50, 62, 65, 66, surprise and 7, 142, 143
70, 101, 133, 142, 147, 153 n., 172, 173, 175, 200 ultimate norm of 172
n., 233, 235 see also irrationality
active sense of 222, 223 reactive attitudes 193
any activity’s being worthwhile tacitly negative 184
depends upon 234 positive 191, 194
belief and supposition relate to 28 Strawsonian models of 183 n.
closed off/blocked off 183, 221 reality 67, 201, 205, 208
condition of 41 children's experience of 203
imaginative exploration of 146 fiction and 152, 161
inability to contemplate 242 inarticulate feelings of 242
logical 132 misrepresenting 152
loss of 238–9, 242 representations of 7, 152, 165
modified or taken away 224 rule of 151, 152, 153, 155–6, 160
plausible 215 n. reality test 149–50
psychological 93 n. reason 2, 7, 10, 82, 116
ruled out 141 moral 84
token 240 practical 84–5, 190
Prinz, J. 3, 21 n., 27 n., 33 n., 80–1, 83, 86, 96 n., see also WKR problem
107, 110, 217 reason-responders/reason-trackers 133, 134–5
Prior, Capt. John 187–8, 192 reasoned pursuit of goals 125, 135
Problem of Parts/Problem of Plenty recalcitrant emotions interfering with 128, 129
80–1, 86 reasoning 67, 153, 155, 217
propositional content 20 n., 24 n., 28, 108, careful 170
131, 133 moral 218
asserting 19 reasons 21, 29 n., 39 n., 48, 65, 76, 88, 157, 159, 171,
beliefs that share 109–10 200, 202–4, 208
conflicts with 115, 132 appeal to 94
evaluative 110 articulate 242
proprioception 33, 36 bad 176 n.
Proust, Marcel 87 balance of 207
psychoanalysis 188 n., 207–8 clarification of 199
uncovering masked shame in compelling 131
treatment 184 n. desire to act in accordance with 219
psychology 1, 8, 25, 28, 52, 69, 85, 92, 93 n., 97, emotions and 9, 117, 119–20, 175, 178, 179
100, 102, 107, 114–15, 192, 201–2, 203, 204, epistemic 117, 119 n.
207, 209, 215 n., 243 evaluative features can be cited as 130
aesthetics 188 evidential 117
dual process model of 200 n. intuitions about 209
moral 2, 4, 116 inventing 55 n.
social 2, 200 n. judgements about 206
see also Gestalt justification via 129, 131
psychopathology 235–8 justifying 107, 117, 121
PTS (post-traumatic stress) 185 memorial 116
Pyrrhonian skeptics 172 non-circular 90, 93, 173
non-emotional 116, 117
R-unitariness 82–3, 85, 88 normativity of 83 n.
Rabinowicz, W. 90 n., 91 ordinary sense perceptual 116
Radden, J. 235 overriding 156
Radford, C. 160 n. prudential 92
Rat-unitariness 83 relevant 95, 97 n.
Ratcliffe, M. 9–10, 223 n., 224, 225, 230 nn., right kind of 90, 93
231 n., 239 n., 241 search for 170, 177
rationality 6, 20, 83, 126, 127, 129, 130 source of 54 n.
emotions are opposed to 2 special 125
practical 2, 4, 10 standard representationalist 36
256 Index
reasons (Cont.) Sacks, O. 113–14

sufficient 117, 119–20 Saddam Hussein 185
testimonial 116 Sander, D. 1 n., 3 n.
theoretical 41 Sartre, Jean-Paul 139, 213 n., 219–23
value 66 n. Scanlon, T. M. 199
weighing up 19, 93 Scarantino, A. 215 n., 222 n., 223
see also good reasons Schechtman, M. 192 n.
recalcitrant emotions 7, 124–36 Scherer, K. 1 n., 3 n.
Reed, M. 59 Schier, F. 151 n.
Reid, T. 55, 72–5, 76, 79, 82, 179 Schmitz, H. 216, 223 n.
Reisner, A. 90 n. Schroeder, T. 129 n.
relatively fitting emotions 6, 90–104 Schwarz, N. 176 n.
representational content 35, 99, 133 SEC (Subjective Evaluative Conditions) 97–100,
unconscious 36 101, 102
see also intentional content self-compassion 191, 208
representational correspondence: self-consciousness 66
rule of 151, 152, 153 selfhood and basic form of 224 n.
test of 150 self-empathy 8–9, 188–98
representationalists 36, 44 self-forgiveness 184 n., 193, 194
representations 20 n., 22, 84, 189 self-knowledge 200 n., 205, 236
action-oriented 222 n. self-love 190
alternative 146 self-trust, see emotional self-trust; epistemic
confusing 7, 165 self-trust
correct 17, 150 self-understanding 4, 9, 10, 199–211
disvalue 43, 46 selfhood 224, 225–6
emotions directed at 156–9 sense of ability 222, 223
enhanced 54, 55, 56, 61, 64 sense perceptions 100, 111, 112, 115,
experiential 46 117, 120
fear of 149, 150, 151 emotions and 6, 107, 108, 109, 121, 130
fictional 7–8, 160, 165 examples biased towards non-plasticity
incorrect 17 of 127
logically prior to emotion 25 paradigmatic 109
non-deceptively false 161 unreliable 116
resemblance theory of 48 sensory experiences 48, 109, 120
somatic 148 sentimentalism 2
value 34, 38 constructive 110
see also evaluative representations; rule of emotion-centred 103
representation rational 6, 94–7, 102, 103
resemblance 45, 47–8 Shakespeare, William:
purported 6 Antony and Cleopatra 154
Ricoeur, Paul 213 n., 215, 222 Henry IV/I) 154
Riggs, W. D. 111 Henry V 152
Roberts, R. 3, 16 n., 19–20, 22 n., 33 n., 97 n., Measure for Measure 154
100–1, 107, 108 n., 109, 110 n., 116, 121 n., 132, Othello (Desdemona in) 159–60, 161
173 n., 189 n., 193, 204, 205, 209, 214 shame 9, 16, 23, 27, 29, 92, 93, 94, 95, 96, 97 n.,
Roeser, S. 29 n., 50 n., 70 n., 74, 75, 76–9, 81, 83, 186, 201, 204, 206, 220
103 n., 108, 165 n., 194 n., 210 n. anguish of 191
Rönnow-Rasmussen, T. 90 n., 91 covering up/masking of 183, 184
Rorty, A. 3 n. defence against 205
Rosch, E. 60 emotional subject’s focus framed by 189
Rosenthal, D. 38 n. experiencing 192, 194
Rowe, D. 237 help to cope with feelings of 8
rule of representation 152, 156, 160 important philosophical work on 184 n.
conformed to 153 persistence of 54
dynamic aspects of emotional response substitutes for 187
sensitive to 154 wounds of 184–6
Rwandan genocide (1990s) 112 Shaw, F. 237
Index 257
Shay, J. 185 n. Tienson, J. 37 n.

Sherlock Holmes (Doyle) 150–51, 157 Titchener, E. 188
Sherman, N. 8–9, 165 n., 183 n., 188 n., 190 nn. Todd, C. 6, 29 n., 50 n., 70 n., 165 n., 210 n.
Slaby, J. 9, 214, 216, 217, 221–4 nn., 225 Tolstoy, Leo 9, 231–5, 237, 238 n., 242
Slovic, P. 61, 114–15 Anna Karenina 146–7, 148, 160, 162, 164
Smith, Adam 188, 193 Transfer Rule 160
Smith, D. 38 n. transparency 200
Smith, M. 213 n. lack of 201, 204, 206, 209
Smithies, D. 36 n. transparency thesis 39
Snow, N. E. 200 n. Trcka, N. 226 n.
Soldati, G. 103 n. Trier, Lars von 244
Solomon, R. 2–3, 18 n., 33, 34, 173 n., 212 n., Trope, Y. 200 n.
213 n., 214, 215, 222 n. true beliefs 114, 115, 121, 172
Sosa, E. 108 n. desires and wishes prone to subverting
Sousa, R. de 2 n., 3 n., 21 n., 33 n., 54 n., acquisition of 205
107, 134 n. tendency of epistemic faculties to
Spelke, E. 138 produce 177
Spence, K. 60 trustworthiness 9, 170, 172, 184
states of affairs 35, 101, 131, 172, 204 aim of getting truth in advance of proof
emotional reactions/responses to 43, 102, 157 of 171
evaluations of 3, 32, 33, 34, 50, 97, 239 see also self-trust; trustworthy emotions
fictional, essential representedness of 151 n. trustworthy emotions 118, 169, 177, 179,
givenness of 32 180, 181
negative or positive value of 46 widespread skepticism about 120
objective properties of 36 see also untrustworthy emotions
properties attributed to 41–2, 43, 47, 50 truth 4, 7, 35, 68, 83, 131, 150, 155, 156, 165, 169,
value-laden 109, 118, 120 199, 208, 234, 242
Steinbock, A. 233 n. access to 236, 238
Stephan A. 214 accorded its appropriate importance 207
Steup, M. 115 appearance of 20 n., 25 n., 133, 135
Stoics 18 n. assessable 20
Strawson, G. 192 n., 194 belief and supposition relate to 28
Strawson, P. F. 183 n. chance of 76
Strayer, J. 188 n. committing oneself to 19, 21
Sturgeon, S. 21 n. evaluative 230
subjectivity 10 evidence for 159 n.
finite 66 n. general 153 n.
Sullivan-Bissett, E. 159 n. illusions of 238–42
surprise 137–45, 229, 239 irrevocable 241
phenomenology of 7 moral 75
unqualified 217 natural desire for 171–2
Sweet Tooth (McEwan) 154 objective 77, 96, 99 n., 100, 102
Sylvan’s Box (Priest) 156 self-conscious attempt to make beliefs
‘syndrome’ term 86 fit 180
success at getting 171, 172, 177
Talbot, M. 138 unbearable and inescapable 231
Tanaka, S. 114 n. unreliability of emotions in tracking 115
Tappolet, C. 3 n., 16 n., 19, 21 n., 22 n., 126–7, 128, unvirtuous concern not to know 205
129–30, 134 n. Tucker, D. M. 59, 60, 61–2, 63, 64, 65 n., 68 n.
Taylor, G. 184 n. Tye, M. 36 n., 39–40 nn.
Teroni, F. 1 n., 3 n., 4–5, 21–2 nn., 25–7 nn., 96 n.,
98 n., 103 n., 121 unitary states 5, 72–89
Teroni, F. 184 n., 214, 219 n. see also R-unitariness; Rat-unitariness
Thagard, P. 229 n., 237 University of Fribourg 103 n.
Thomasson, A. 150 n. University of Geneva 70 n., 103 n.
Thompson, E. 223 n. unsurprising events 140–42, 145
Tiberius, V. 200 n. Untold War, The (Sherman) 183, 187
258 Index
untrustworthy emotions/untrustworthiness Webb, D. 233 n.

114, 170 Wedgewood, R. 108
unwarrantedness 10 Weinberg, J. 159 n.
Weltman, G. 59
Vaillant, G. E. 201, 203 n. Wharton, E. 154
valence 3, 23, 95, 192 Whiting, D. 109 n.
effects on attention 5, 52, 57–9, 60–1, 62, 63 Wiggins, D. 94 n., 199, 201
target detection influenced by 59 Williams, B. 184 n., 188 n., 199, 200, 201, 205
tone and 193 Wilson, A. N. 235
value and 65–9 Wilson, T. D. 200 n.
value judgments 77, 237, 240 Wittgenstein, Ludwig 132 n.
value relativism 90–4 WKR (wrong kind of reason) problem 90, 91–2,
Velleman, J. D. 184 n., 218–19 95, 103
Vischer, Robert 188 Wollheim, R. 192 n., 207
visual experience 17, 35, 38, 39–40, 43–4, Wood, W. J. 116, 204, 205, 209
47, 112 Woodruff, P. 186 n.
Vuilleumier, P. 53 n. Wüschner, P. 9
Wynn, M. R. 238
Walker, M. J. 184 n., 191
Walton, K. 146–8 nn., 157–8 Zagzebski, L. 6, 8, 9, 98 n., 107, 108, 109, 173 n.,
Watkins, M. 121 180–1 nn.
Weatherson, B. 148 n., 156 n. Zahavi, D. 38 n.

Emotion and Value (2015)

Uploaded by

Document Information

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

Emotion and Value (2015)

Uploaded by

Copyright:

Available Formats

Emotion and Value

Emotion and Value

1. Emotion and Value: Introduction 1

14. Emotion and Agency 212

Sabine Roeser and Cain Todd

Part I: Emotion and the Nature of Value

a particular reason-relation that is articulated in terms of the appropriateness of each

Part II: Emotion, Evaluation, and Justification

Part III: Emotion, Value, and the Self

I. Emotion as Access to Value

II. The Evaluative Judgment Theory

on the intelligibility and appropriateness of the evaluative judgments at issue. As is

III. The Perceptual Model

IV. The Attitudinal Theory

lends itself to an appealing development in terms of an aspect of emotions commonly

Crane, T. (1992). “The Nonconceptual Content of Experience.” In The Contents of Experience:

rough taxonomy of kinds of phenomenology, distinguishing evaluative phenomenol-

2. Theory of Content

these thoughts. Finally, ‘evaluative phenomenology’ is—so I propose—a further sui

4. Evaluative Phenomenology and Evaluative

of numbers. We take the natural numbers to have wholly objective, mind-independent

(2) It is in virtue of the subject’s awareness of the sensory phenomenology of the

(3) The evaluative phenomenology that is integral to an emotional experience suf-

5. Fine-grained Evaluative Phenomenology

positive evaluative phenomenology. But are negative evaluative phenomenology and

de Sousa, R. (1987). The Rationality of Emotion (Cambridge, MA: MIT Press).

1. Emotion, Attention, and Evaluation

2. A Generalization Problem

(i) Valence Affects Breadth of Visual Attention

(ii) Valence Influences Target Detection

(iii) Valence Influences Attention More Generally

There is, correspondingly, a significant amount of evidence that positive emotion

3. Constitutive and Consequential Attentional

Derryberry and Tucker (1994: 180).

4. Valence and Value

Derryberry and Tucker (1994: 184).

mer, but not the latter, have intentional objects.

Similarly, we value ecological systems in virtue of their ‘intricate relationships, equi-

My main aim in this chapter is to consider whether it is possible to think of an emotion

But the feeling is not a ‘mere’ feeling:

So Reid offers us a tripartite account of approbation as a combination of judge-

Essay 5.7; p.673.1.

make it more consistent with her original intentions.

Prinz (2004: 241, 242).

understood as a presentation of a supposed fact, where presentations are distinguished

Their appeal to the notion of a ‘syndrome’ worries me. Etymologically, a ‘syndrome’

D’Arms and Jacobson (2003: 138).

Goldie (2011: 131, fn.38).

much improved this chapter.

1. Essential Contestability and Value Relativism

FA inherits the attractions of buck-passing accounts of value, promising to avoid

independently of the evaluative concepts that FA employ them to analyse, although

2. Rational Sentimentalism

3. Emotional Phenomenology and Apparent

1. The Perceptual Character of Emotion and the

2. Emotion as a Source of Justified

Sacks explains that Grandin exhibits a “poverty of emotional or aesthetic response

3. The Unreliability Objection

unduly influenced by whether they experienced positive or negative emotion (affect)

Alternatively, according to certain externalist theories of justification, an emotion’s

4. The Common Source Objection

4.1 Emotional Justification

4.2 Are Emotions Just Reasons to Look for Reasons?

4.3 Why Do We Not Cite Our Emotions As Our Justification?